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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1364 - (show annotations)
Sat Oct 5 15:45:11 2013 UTC (6 years, 1 month ago) by ph10
File size: 252391 byte(s)
Error occurred while calculating annotation data.
Add VT to the set of characters recognized as white space.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.34 xx-xxxx-201x
5 --------------------------
6
7 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8 executable memory. Patch inspired by Carsten Klein.
9
10 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11 this macro is never tested and has no effect, because the work to support
12 coverage involves only compiling and linking options and special targets in
13 the Makefile. The comment in config.h implied that defining the macro would
14 enable coverage support, which is totally false. There was also support for
15 setting this macro in the CMake files (my fault, I just copied it from
16 configure). SUPPORT_GCOV has now been removed.
17
18 3. Make a small performance improvement in strlen16() and strlen32() in
19 pcretest.
20
21 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23 these cases). There was also one in pcretest.c.
24
25 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
26
27 6. In UTF mode, the code for checking whether a group could match an empty
28 string (which is used for indefinitely repeated groups to allow for
29 breaking an infinite loop) was broken when the group contained a repeated
30 negated single-character class with a character that occupied more than one
31 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
32 UTF-8 mode). The effect was undefined: the group might or might not be
33 deemed as matching an empty string, or the program might have crashed.
34
35 7. The code for checking whether a group could match an empty string was not
36 recognizing that \h, \H, \v, \V, and \R must match a character.
37
38 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
39 an empty string. If it can, pcretest shows this in its information output.
40
41 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
42 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
43 matched by pcre_exec() without using JIT:
44
45 (a) If the rest of the pattern did not match after a maximal run of
46 grapheme clusters, the code for backing up to try with fewer of them
47 did not always back up over a full grapheme when characters that do not
48 have the modifier quality were involved, e.g. Hangul syllables.
49
50 (b) If the match point in a subject started with modifier character, and
51 there was no match, the code could incorrectly back up beyond the match
52 point, and potentially beyond the first character in the subject,
53 leading to a segfault or an incorrect match result.
54
55 10. A conditional group with an assertion condition could lead to PCRE
56 recording an incorrect first data item for a match if no other first data
57 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58 first data item, and therefore matched "ca" after "c" instead of at the
59 start.
60
61 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
62 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
63
64 12. The source of pcregrep now includes z/OS-specific code so that it can be
65 compiled for z/OS as part of the special z/OS distribution.
66
67 13. Added the -T and -TM options to pcretest.
68
69 14. The code in pcre_compile.c for creating the table of named capturing groups
70 has been refactored. Instead of creating the table dynamically during the
71 actual compiling pass, the information is remembered during the pre-compile
72 pass (on the stack unless there are more than 20 named groups, in which
73 case malloc() is used) and the whole table is created before the actual
74 compile happens. This has simplified the code (it is now nearly 150 lines
75 shorter) and prepared the way for better handling of references to groups
76 with duplicate names.
77
78 15. A back reference to a named subpattern when there is more than one of the
79 same name now checks them in the order in which they appear in the pattern.
80 The first one that is set is used for the reference. Previously only the
81 first one was inspected. This change makes PCRE more compatible with Perl.
82
83 16. Unicode character properties were updated from Unicode 6.3.0.
84
85 17. The compile-time code for auto-possessification has been refactored, based
86 on a patch by Zoltan Herczeg. It now happens after instead of during
87 compilation. The code is cleaner, and more cases are handled. The option
88 PCRE_NO_AUTO_POSSESSIFY is added for testing purposes, and the -O and /O
89 options in pcretest are provided to set it.
90
91 18. The character VT has been added to the set of characters that match \s and
92 are generally treated as white space, following this same change in Perl
93 5.18. There is now no difference between "Perl space" and "POSIX space".
94
95
96 Version 8.33 28-May-2013
97 ------------------------
98
99 1. Added 'U' to some constants that are compared to unsigned integers, to
100 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
101 variables that are added to signed variables, to ensure the result is
102 signed and can be negated.
103
104 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
105 CMake config files.
106
107 3. Revise the creation of config.h.generic so that all boolean macros are
108 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
109 overriding via -D on the command line possible.
110
111 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
112 to unsigned int is reported to make a quite noticeable speed difference in
113 a specific Windows environment. Testing on Linux did also appear to show
114 some benefit (and it is clearly not harmful). Also fixed the definition of
115 Xop which should be unsigned.
116
117 5. Related to (4), changing the definition of the intermediate variable cc
118 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
119 improvements.
120
121 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
122 unnecessary spaces.
123
124 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
125 and later.
126
127 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
128
129 9. Optimizing fast_forward_start_bits in JIT.
130
131 10. Adding support for callouts in JIT, and fixing some issues revealed
132 during this work. Namely:
133
134 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
135
136 (b) Minimum length was not checked before the matching is started.
137
138 11. The value of capture_last that is passed to callouts was incorrect in some
139 cases when there was a capture on one path that was subsequently abandoned
140 after a backtrack. Also, the capture_last value is now reset after a
141 recursion, since all captures are also reset in this case.
142
143 12. The interpreter no longer returns the "too many substrings" error in the
144 case when an overflowing capture is in a branch that is subsequently
145 abandoned after a backtrack.
146
147 13. In the pathological case when an offset vector of size 2 is used, pcretest
148 now prints out the matched string after a yield of 0 or 1.
149
150 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
151 Only supported by the JIT compiler at the moment.
152
153 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
154
155 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
156 offset of the starting point of the matching process, provided the offsets
157 vector is large enough.
158
159 17. The \A escape now records a lookbehind value of 1, though its execution
160 does not actually inspect the previous character. This is to ensure that,
161 in partial multi-segment matching, at least one character from the old
162 segment is retained when a new segment is processed. Otherwise, if there
163 are no lookbehinds in the pattern, \A might match incorrectly at the start
164 of a new segment.
165
166 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
167
168 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
169 gives some modest performance improvement in 8-bit mode.
170
171 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
172 be expressed in certain programming languages using Universal Character
173 Names.
174
175 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
176 which points out that "non characters" are not "characters that may not
177 appear in Unicode strings" but rather "characters that are reserved for
178 internal use and have only local meaning".
179
180 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
181 there was a conditional group that depended on an assertion, if the
182 assertion was false, the callout that immediately followed the alternation
183 in the condition was skipped when pcre_exec() was used for matching.
184
185 23. Allow an explicit callout to be inserted before an assertion that is the
186 condition for a conditional group, for compatibility with automatic
187 callouts, which always insert a callout at this point.
188
189 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
190 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
191
192 25. (*PRUNE) is now supported by the JIT compiler.
193
194 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
195
196 27. Fix the case where there are two or more SKIPs with arguments that may be
197 ignored.
198
199 28. (*SKIP) is now supported by the JIT compiler.
200
201 29. (*THEN) is now supported by the JIT compiler.
202
203 30. Update RunTest with additional test selector options.
204
205 31. The way PCRE handles backtracking verbs has been changed in two ways.
206
207 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
208 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
209 backtracking. In some cases this makes it more Perl-compatible, but Perl's
210 rather obscure rules do not always do the same thing.
211
212 (2) Previously, backtracking verbs were confined within assertions. This is
213 no longer the case for positive assertions, except for (*ACCEPT). Again,
214 this sometimes improves Perl compatibility, and sometimes does not.
215
216 32. A number of tests that were in test 2 because Perl did things differently
217 have been moved to test 1, because either Perl or PCRE has changed, and
218 these tests are now compatible.
219
220 32. Backtracking control verbs are now handled in the same way in JIT and
221 interpreter.
222
223 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
224 contained a forward subroutine reference caused a compile error.
225
226 34. Auto-detect and optimize limited repetitions in JIT.
227
228 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
229 blocking (*UTF) etc.
230
231 36. In the interpreter, maximizing pattern repetitions for characters and
232 character types now use tail recursion, which reduces stack usage.
233
234 37. The value of the max lookbehind was not correctly preserved if a compiled
235 and saved regex was reloaded on a host of different endianness.
236
237 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
238 of the compiled pattern block, expand the flags field from 16 to 32 bits
239 because it was almost full.
240
241 39. Try madvise first before posix_madvise.
242
243 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
244 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
245 original change locked out all matches of empty strings. This has been
246 changed so that one match of an empty string per line is recognized.
247 Subsequent searches on the same line (for colouring or for --only-matching,
248 for example) do not recognize empty strings.
249
250 41. Applied a user patch to fix a number of spelling mistakes in comments.
251
252 42. Data lines longer than 65536 caused pcretest to crash.
253
254 43. Clarified the data type for length and startoffset arguments for pcre_exec
255 and pcre_dfa_exec in the function-specific man pages, where they were
256 explicitly stated to be in bytes, never having been updated. I also added
257 some clarification to the pcreapi man page.
258
259 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
260 a segmentation fault.
261
262
263 Version 8.32 30-November-2012
264 -----------------------------
265
266 1. Improved JIT compiler optimizations for first character search and single
267 character iterators.
268
269 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
270 Patch by Daniel Richard G.
271
272 3. Single character iterator optimizations in the JIT compiler.
273
274 4. Improved JIT compiler optimizations for character ranges.
275
276 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
277 Reported by Giuseppe D'Angelo.
278
279 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
280 of a line, was being set incorrectly in cases where .* appeared inside
281 atomic brackets at the start of a pattern, or where there was a subsequent
282 *PRUNE or *SKIP.
283
284 7. Improved instruction cache flush for POWER/PowerPC.
285 Patch by Daniel Richard G.
286
287 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
288 grep:
289
290 (a) There is now no limit to the number of patterns to be matched.
291
292 (b) An error is given if a pattern is too long.
293
294 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
295 are now supported.
296
297 (d) --exclude-from and --include-from (multiple use) have been added.
298
299 (e) Exclusions and inclusions now apply to all files and directories, not
300 just to those obtained from scanning a directory recursively.
301
302 (f) Multiple uses of -f and --file-list are now supported.
303
304 (g) In a Windows environment, the default for -d has been changed from
305 "read" (the GNU grep default) to "skip", because otherwise the presence
306 of a directory in the file list provokes an error.
307
308 (h) The documentation has been revised and clarified in places.
309
310 9. Improve the matching speed of capturing brackets.
311
312 10. Changed the meaning of \X so that it now matches a Unicode extended
313 grapheme cluster.
314
315 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
316 out POSIX threads when JIT support is configured.
317
318 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
319
320 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
321 structure could go wrong in environments where size_t is not the same size
322 as int.
323
324 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
325 set.
326
327 15. The EBCDIC support had decayed; later updates to the code had included
328 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
329 general tidy up of EBCDIC-related issues, and the documentation was also
330 not quite right. There is now a test that can be run on ASCII systems to
331 check some of the EBCDIC-related things (but is it not a full test).
332
333 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
334 in a small tidy to the code.
335
336 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
337
338 18. If the --only-matching (-o) option in pcregrep is specified multiple
339 times, each one causes appropriate output. For example, -o1 -o2 outputs the
340 substrings matched by the 1st and 2nd capturing parentheses. A separating
341 string can be specified by --om-separator (default empty).
342
343 19. Improving the first n character searches.
344
345 20. Turn case lists for horizontal and vertical white space into macros so that
346 they are defined only once.
347
348 21. This set of changes together give more compatible Unicode case-folding
349 behaviour for characters that have more than one other case when UCP
350 support is available.
351
352 (a) The Unicode property table now has offsets into a new table of sets of
353 three or more characters that are case-equivalent. The MultiStage2.py
354 script that generates these tables (the pcre_ucd.c file) now scans
355 CaseFolding.txt instead of UnicodeData.txt for character case
356 information.
357
358 (b) The code for adding characters or ranges of characters to a character
359 class has been abstracted into a generalized function that also handles
360 case-independence. In UTF-mode with UCP support, this uses the new data
361 to handle characters with more than one other case.
362
363 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
364 whose other case is greater than 256 are now correctly matched
365 caselessly. Previously, the high codepoint matched the low one, but not
366 vice versa.
367
368 (d) The processing of \h, \H, \v, and \ in character classes now makes use
369 of the new class addition function, using character lists defined as
370 macros alongside the case definitions of 20 above.
371
372 (e) Caseless back references now work with characters that have more than
373 one other case.
374
375 (f) General caseless matching of characters with more than one other case
376 is supported.
377
378 22. Unicode character properties were updated from Unicode 6.2.0
379
380 23. Improved CMake support under Windows. Patch by Daniel Richard G.
381
382 24. Add support for 32-bit character strings, and UTF-32
383
384 25. Major JIT compiler update (code refactoring and bugfixing).
385 Experimental Sparc 32 support is added.
386
387 26. Applied a modified version of Daniel Richard G's patch to create
388 pcre.h.generic and config.h.generic by "make" instead of in the
389 PrepareRelease script.
390
391 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
392 pcre_compile.c when checking for a zero character.
393
394 28. Introducing a native interface for JIT. Through this interface, the compiled
395 machine code can be directly executed. The purpose of this interface is to
396 provide fast pattern matching, so several sanity checks are not performed.
397 However, feature tests are still performed. The new interface provides
398 1.4x speedup compared to the old one.
399
400 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
401 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
402 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
403
404 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
405 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
406 the "old" RFC 2279). Instead, it ended up passing a negative length to
407 pcre_exec().
408
409 31. Add support for GCC's visibility feature to hide internal functions.
410
411 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
412 "unknown -C option" after outputting 0 or 1.
413
414 33. There is now support for generating a code coverage report for the test
415 suite in environments where gcc is the compiler and lcov is installed. This
416 is mainly for the benefit of the developers.
417
418 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
419 unaddressable using valgrind annotations, allowing valgrind to detect
420 invalid memory accesses. This is mainly for the benefit of the developers.
421
422 25. (*UTF) can now be used to start a pattern in any of the three libraries.
423
424 26. Give configure error if --enable-cpp but no C++ compiler found.
425
426
427 Version 8.31 06-July-2012
428 -------------------------
429
430 1. Fixing a wrong JIT test case and some compiler warnings.
431
432 2. Removed a bashism from the RunTest script.
433
434 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
435 to unsigned type, result still unsigned" that was given by an MS compiler
436 on encountering the code "-sizeof(xxx)".
437
438 4. Partial matching support is added to the JIT compiler.
439
440 5. Fixed several bugs concerned with partial matching of items that consist
441 of more than one character:
442
443 (a) /^(..)\1/ did not partially match "aba" because checking references was
444 done on an "all or nothing" basis. This also applied to repeated
445 references.
446
447 (b) \R did not give a hard partial match if \r was found at the end of the
448 subject.
449
450 (c) \X did not give a hard partial match after matching one or more
451 characters at the end of the subject.
452
453 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
454 a partial match for the string "\r".
455
456 (e) When newline was set to CRLF, the metacharacter "." did not recognize
457 a partial match for a CR character at the end of the subject string.
458
459 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
460 running pcretest, the text "(JIT)" added to the output whenever JIT is
461 actually used to run the match.
462
463 7. Individual JIT compile options can be set in pcretest by following -s+[+]
464 or /S+[+] with a digit between 1 and 7.
465
466 8. OP_NOT now supports any UTF character not just single-byte ones.
467
468 9. (*MARK) control verb is now supported by the JIT compiler.
469
470 10. The command "./RunTest list" lists the available tests without actually
471 running any of them. (Because I keep forgetting what they all are.)
472
473 11. Add PCRE_INFO_MAXLOOKBEHIND.
474
475 12. Applied a (slightly modified) user-supplied patch that improves performance
476 when the heap is used for recursion (compiled with --disable-stack-for-
477 recursion). Instead of malloc and free for each heap frame each time a
478 logical recursion happens, frames are retained on a chain and re-used where
479 possible. This sometimes gives as much as 30% improvement.
480
481 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
482 call.
483
484 14. As documented, (*COMMIT) is now confined to within a positive assertion.
485
486 15. It is now possible to link pcretest with libedit as an alternative to
487 libreadline.
488
489 16. (*COMMIT) control verb is now supported by the JIT compiler.
490
491 17. The Unicode data tables have been updated to Unicode 6.1.0.
492
493 18. Added --file-list option to pcregrep.
494
495 19. Added binary file support to pcregrep, including the -a, --binary-files,
496 -I, and --text options.
497
498 20. The madvise function is renamed for posix_madvise for QNX compatibility
499 reasons. Fixed by Giuseppe D'Angelo.
500
501 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
502
503 22. Changed the option for creating symbolic links for 16-bit man pages from
504 -s to -sf so that re-installing does not cause issues.
505
506 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
507
508 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
509 matches in certain environments (the workspace was not being correctly
510 retained). Also added to pcre_dfa_exec() a simple plausibility check on
511 some of the workspace data at the beginning of a restart.
512
513 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
514 was not doing so when it should - probably a typo introduced by SVN 528
515 (change 8.10/14).
516
517 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
518 \w+ when the character tables indicated that \x{c4} was a word character.
519 There were several related cases, all because the tests for doing a table
520 lookup were testing for characters less than 127 instead of 255.
521
522 27. If a pattern contains capturing parentheses that are not used in a match,
523 their slots in the ovector are set to -1. For those that are higher than
524 any matched groups, this happens at the end of processing. In the case when
525 there were back references that the ovector was too small to contain
526 (causing temporary malloc'd memory to be used during matching), and the
527 highest capturing number was not used, memory off the end of the ovector
528 was incorrectly being set to -1. (It was using the size of the temporary
529 memory instead of the true size.)
530
531 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
532 ovector size, it uses memory at the end of the block that it has got.
533
534 29. Check for an overlong MARK name and give an error at compile time. The
535 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
536
537 30. JIT compiler update.
538
539 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
540 Rill for the patch.
541
542 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
543
544 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
545
546 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
547 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
548 was enabled.
549
550 35. Improve JIT code generation for greedy plus quantifier.
551
552 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
553 1 to "aa" instead of to an empty string. The bug affected repeated groups
554 that could potentially match an empty string.
555
556 37. Optimizing single character iterators in JIT.
557
558 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
559 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
560 codepoints that are too big for the mode are faulted, and in a UTF mode,
561 disallowed codepoints are also faulted.
562
563 39. If PCRE was compiled with UTF support, in three places in the DFA
564 matcher there was code that should only have been obeyed in UTF mode, but
565 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
566 processing when bytes with values greater than 127 were present. In 16-bit
567 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
568 both cases the values are those that cannot be the first data item in a UTF
569 character. The three items that might have provoked this were recursions,
570 possessively repeated groups, and atomic groups.
571
572 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
573 and pcregrep, because some OS require shared objects to be explicitly
574 passed to ld, causing the link step to fail if they are not.
575
576 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
577 mode, patterns that started with \h* or \R* might be incorrectly matched.
578
579
580 Version 8.30 04-February-2012
581 -----------------------------
582
583 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
584 name is defined in ctype.h.
585
586 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
587 only in quite long subpatterns.
588
589 3. Removed the function pcre_info(), which has been obsolete and deprecated
590 since it was replaced by pcre_fullinfo() in February 2000.
591
592 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
593 match a (*MARK), and the match failed at the start of the subject, a
594 reference to memory before the start of the subject could occur. This bug
595 was introduced by fix 17 of release 8.21.
596
597 5. A reference to an unset group with zero minimum repetition was giving
598 totally wrong answers (in non-JavaScript-compatibility mode). For example,
599 /(another)?(\1?)test/ matched against "hello world test". This bug was
600 introduced in release 8.13.
601
602 6. Add support for 16-bit character strings (a large amount of work involving
603 many changes and refactorings).
604
605 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
606 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
607 from a file.
608
609 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
610 rounding is not applied in this particular case).
611
612 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
613 if they appear, or are escaped, in patterns.
614
615 10. Get rid of a number of -Wunused-but-set-variable warnings.
616
617 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
618 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
619 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
620 also returns the mark "x". This bug applied to capturing parentheses,
621 non-capturing parentheses, and atomic parentheses. It also applied to some
622 assertions.
623
624 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
625 information out of configure.ac instead of relying on pcre.h.generic, which
626 is not stored in the repository.
627
628 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
629 -lreadline.
630
631 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
632
633 15. Applied Graycode's patch to put the top-level frame on the stack rather
634 than the heap when not using the stack for recursion. This gives a
635 performance improvement in many cases when recursion is not deep.
636
637 16. Experimental code added to "pcretest -C" to output the stack frame size.
638
639
640 Version 8.21 12-Dec-2011
641 ------------------------
642
643 1. Updating the JIT compiler.
644
645 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
646 are added as well.
647
648 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
649 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
650 calling _pcre_jit_exec. Some extra comments are added.
651
652 4. (*MARK) settings inside atomic groups that do not contain any capturing
653 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
654 was introduced by change 18 for 8.20.
655
656 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
657 ECMA-262 standard.
658
659 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
660 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
661 This bug was probably introduced by change 9 of 8.13.
662
663 7. While fixing 6 above, I noticed that a number of other items were being
664 incorrectly rejected as "not fixed length". This arose partly because newer
665 opcodes had not been added to the fixed-length checking code. I have (a)
666 corrected the bug and added tests for these items, and (b) arranged for an
667 error to occur if an unknown opcode is encountered while checking for fixed
668 length instead of just assuming "not fixed length". The items that were
669 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
670 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
671 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
672
673 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
674 being incorrectly compiled and would have given unpredicatble results.
675
676 9. A possessively repeated subpattern with minimum repeat count greater than
677 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
678 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
679 the first (A) could occur when it should not.
680
681 10. Add a cast and remove a redundant test from the code.
682
683 11. JIT should use pcre_malloc/pcre_free for allocation.
684
685 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
686 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
687 is anything other than /usr, -L is still shown).
688
689 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
690
691 14. Perl does not support \N without a following name in a [] class; PCRE now
692 also gives an error.
693
694 15. If a forward reference was repeated with an upper limit of around 2000,
695 it caused the error "internal error: overran compiling workspace". The
696 maximum number of forward references (including repeats) was limited by the
697 internal workspace, and dependent on the LINK_SIZE. The code has been
698 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
699 the default depends on LINK_SIZE. There is a new upper limit (for safety)
700 of around 200,000 forward references. While doing this, I also speeded up
701 the filling in of repeated forward references.
702
703 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
704 incorrectly expecting the subject to contain another "a" after the start.
705
706 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
707 in the match, the SKIP should be ignored. This was not happening; instead
708 the SKIP was being treated as NOMATCH. For patterns such as
709 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
710 tested.
711
712 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
713 now much more compatible with Perl, in particular in cases where the result
714 is a non-match for a non-anchored pattern. For example, if
715 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
716 "m", where previously it did not return a name. A side effect of this
717 change is that for partial matches, the last encountered mark name is
718 returned, as for non matches. A number of tests that were previously not
719 Perl-compatible have been moved into the Perl-compatible test files. The
720 refactoring has had the pleasing side effect of removing one argument from
721 the match() function, thus reducing its stack requirements.
722
723 19. If the /S+ option was used in pcretest to study a pattern using JIT,
724 subsequent uses of /S (without +) incorrectly behaved like /S+.
725
726 21. Retrieve executable code size support for the JIT compiler and fixing
727 some warnings.
728
729 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
730 not work when the shorter character appeared right at the end of the
731 subject string.
732
733 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
734 systems.
735
736 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
737 output it when the /M option is used in pcretest.
738
739 25. The CheckMan script was not being included in the distribution. Also, added
740 an explicit "perl" to run Perl scripts from the PrepareRelease script
741 because this is reportedly needed in Windows.
742
743 26. If study data was being save in a file and studying had not found a set of
744 "starts with" bytes for the pattern, the data written to the file (though
745 never used) was taken from uninitialized memory and so caused valgrind to
746 complain.
747
748 27. Updated RunTest.bat as provided by Sheri Pierce.
749
750 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
751
752 29. Computation of memory usage for the table of capturing group names was
753 giving an unnecessarily large value.
754
755
756 Version 8.20 21-Oct-2011
757 ------------------------
758
759 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
760 a POSIX class. After further experiments with Perl, which convinced me that
761 Perl has bugs and confusions, a closing square bracket is no longer allowed
762 in a POSIX name. This bug also affected patterns with classes that started
763 with full stops.
764
765 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
766 captured substring, but while checking the failing first alternative,
767 substring 1 is temporarily captured. If the output vector supplied to
768 pcre_exec() was not big enough for this capture, the yield of the function
769 was still zero ("insufficient space for captured substrings"). This cannot
770 be totally fixed without adding another stack variable, which seems a lot
771 of expense for a edge case. However, I have improved the situation in cases
772 such as /(a)(b)x|abc/ matched against "abc", where the return code
773 indicates that fewer than the maximum number of slots in the ovector have
774 been set.
775
776 3. Related to (2) above: when there are more back references in a pattern than
777 slots in the output vector, pcre_exec() uses temporary memory during
778 matching, and copies in the captures as far as possible afterwards. It was
779 using the entire output vector, but this conflicts with the specification
780 that only 2/3 is used for passing back captured substrings. Now it uses
781 only the first 2/3, for compatibility. This is, of course, another edge
782 case.
783
784 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
785 main code base, and can be used by building with --enable-jit. When this is
786 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
787 runtime --no-jit option is given.
788
789 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
790 ovector, the return from the function was zero, implying that there were
791 other matches that did not fit. The correct "exactly full" value is now
792 returned.
793
794 6. If a subpattern that was called recursively or as a subroutine contained
795 (*PRUNE) or any other control that caused it to give a non-standard return,
796 invalid errors such as "Error -26 (nested recursion at the same subject
797 position)" or even infinite loops could occur.
798
799 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
800 computing the minimum length on reaching *ACCEPT, and so ended up with the
801 wrong value of 1 rather than 0. Further investigation indicates that
802 computing a minimum subject length in the presence of *ACCEPT is difficult
803 (think back references, subroutine calls), and so I have changed the code
804 so that no minimum is registered for a pattern that contains *ACCEPT.
805
806 8. If (*THEN) was present in the first (true) branch of a conditional group,
807 it was not handled as intended. [But see 16 below.]
808
809 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
810 Sheri Pierce.
811
812 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
813 the first byte in a match must be "a".
814
815 11. Change 17 for 8.13 increased the recursion depth for patterns like
816 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
817 pattern contains any instances of (*THEN). If it does not, the old
818 optimizations are restored. It would be nice to do this on a per-group
819 basis, but at the moment that is not feasible.
820
821 12. In some environments, the output of pcretest -C is CRLF terminated. This
822 broke RunTest's code that checks for the link size. A single white space
823 character after the value is now allowed for.
824
825 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
826 For "fr", it uses the Windows-specific input and output files.
827
828 14. If (*THEN) appeared in a group that was called recursively or as a
829 subroutine, it did not work as intended. [But see next item.]
830
831 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
832 pattern fragments (but not containing any | characters). If A and B are
833 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
834 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
835 D. In other words, Perl considers parentheses that do not contain any |
836 characters to be part of a surrounding alternative, whereas PCRE was
837 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
838 differently. PCRE now behaves in the same way as Perl, except in the case
839 of subroutine/recursion calls such as (?1) which have in any case always
840 been different (but PCRE had them first :-).
841
842 16. Related to 15 above: Perl does not treat the | in a conditional group as
843 creating alternatives. Such a group is treated in the same way as an
844 ordinary group without any | characters when processing (*THEN). PCRE has
845 been changed to match Perl's behaviour.
846
847 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
848 RunGrepTest script failed.
849
850 18. Change 22 for version 13 caused atomic groups to use more stack. This is
851 inevitable for groups that contain captures, but it can lead to a lot of
852 stack use in large patterns. The old behaviour has been restored for atomic
853 groups that do not contain any capturing parentheses.
854
855 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
856 suppress the check for a minimum subject length at run time. (If it was
857 given to pcre_exec() or pcre_dfa_exec() it did work.)
858
859 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
860 fail to work when decoding hex characters in data strings in EBCDIC
861 environments.
862
863 21. It appears that in at least one Mac OS environment, the isxdigit() function
864 is implemented as a macro that evaluates to its argument more than once,
865 contravening the C 90 Standard (I haven't checked a later standard). There
866 was an instance in pcretest which caused it to go wrong when processing
867 \x{...} escapes in subject strings. The has been rewritten to avoid using
868 things like p++ in the argument of isxdigit().
869
870
871 Version 8.13 16-Aug-2011
872 ------------------------
873
874 1. The Unicode data tables have been updated to Unicode 6.0.0.
875
876 2. Two minor typos in pcre_internal.h have been fixed.
877
878 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
879 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
880 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
881
882 4. There were a number of related bugs in the code for matching backrefences
883 caselessly in UTF-8 mode when codes for the characters concerned were
884 different numbers of bytes. For example, U+023A and U+2C65 are an upper
885 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
886 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
887 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
888 2-byte code at the end of the subject (it thought there wasn't enough data
889 left).
890
891 5. Comprehensive information about what went wrong is now returned by
892 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
893 as the output vector has at least 2 elements. The offset of the start of
894 the failing character and a reason code are placed in the vector.
895
896 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
897 now returned is for the first byte of the failing character, instead of the
898 last byte inspected. This is an incompatible change, but I hope it is small
899 enough not to be a problem. It makes the returned offset consistent with
900 pcre_exec() and pcre_dfa_exec().
901
902 7. pcretest now gives a text phrase as well as the error number when
903 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
904 failure, the offset and reason code are output.
905
906 8. When \R was used with a maximizing quantifier it failed to skip backwards
907 over a \r\n pair if the subsequent match failed. Instead, it just skipped
908 back over a single character (\n). This seems wrong (because it treated the
909 two characters as a single entity when going forwards), conflicts with the
910 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
911 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
912 has been changed.
913
914 9. Some internal refactoring has changed the processing so that the handling
915 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
916 time (the PCRE_DOTALL option was changed this way some time ago: version
917 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
918 which was always a bit of a fudge. It also means that there is one less
919 argument for the match() function, which reduces its stack requirements
920 slightly. This change also fixes an incompatibility with Perl: the pattern
921 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
922
923 10. More internal refactoring has drastically reduced the number of recursive
924 calls to match() for possessively repeated groups such as (abc)++ when
925 using pcre_exec().
926
927 11. While implementing 10, a number of bugs in the handling of groups were
928 discovered and fixed:
929
930 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
931 (a|)*(?1) gave a compile-time internal error.
932 ((a|)+)+ did not notice that the outer group could match an empty string.
933 (^a|^)+ was not marked as anchored.
934 (.*a|.*)+ was not marked as matching at start or after a newline.
935
936 12. Yet more internal refactoring has removed another argument from the match()
937 function. Special calls to this function are now indicated by setting a
938 value in a variable in the "match data" data block.
939
940 13. Be more explicit in pcre_study() instead of relying on "default" for
941 opcodes that mean there is no starting character; this means that when new
942 ones are added and accidentally left out of pcre_study(), testing should
943 pick them up.
944
945 14. The -s option of pcretest has been documented for ages as being an old
946 synonym of -m (show memory usage). I have changed it to mean "force study
947 for every regex", that is, assume /S for every regex. This is similar to -i
948 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
949 using it. It makes it easier to run collections of tests with and without
950 study enabled, and thereby test pcre_study() more easily. All the standard
951 tests are now run with and without -s (but some patterns can be marked as
952 "never study" - see 20 below).
953
954 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
955 restoration of the capturing data to the outer values was not happening
956 correctly.
957
958 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
959 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
960 pattern had matched an empty string, and so incorrectly returned a no
961 match.
962
963 17. There was optimizing code for the last branch of non-capturing parentheses,
964 and also for the obeyed branch of a conditional subexpression, which used
965 tail recursion to cut down on stack usage. Unfortunately, now that there is
966 the possibility of (*THEN) occurring in these branches, tail recursion is
967 no longer possible because the return has to be checked for (*THEN). These
968 two optimizations have therefore been removed. [But see 8.20/11 above.]
969
970 18. If a pattern containing \R was studied, it was assumed that \R always
971 matched two bytes, thus causing the minimum subject length to be
972 incorrectly computed because \R can also match just one byte.
973
974 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
975 was incorrectly computed.
976
977 20. If /S is present twice on a test pattern in pcretest input, it now
978 *disables* studying, thereby overriding the use of -s on the command line
979 (see 14 above). This is necessary for one or two tests to keep the output
980 identical in both cases.
981
982 21. When (*ACCEPT) was used in an assertion that matched an empty string and
983 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
984
985 22. When an atomic group that contained a capturing parenthesis was
986 successfully matched, but the branch in which it appeared failed, the
987 capturing was not being forgotten if a higher numbered group was later
988 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
989 group 1 to "a", when in fact it should be unset. This applied to multi-
990 branched capturing and non-capturing groups, repeated or not, and also to
991 positive assertions (capturing in negative assertions does not happen
992 in PCRE) and also to nested atomic groups.
993
994 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
995 subject after a captured substring, to make it easier to tell which of a
996 number of identical substrings has been captured.
997
998 24. The way atomic groups are processed by pcre_exec() has been changed so that
999 if they are repeated, backtracking one repetition now resets captured
1000 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1001 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1002 "aaa". Previously, it would have been "a". As part of this code
1003 refactoring, the way recursive calls are handled has also been changed.
1004
1005 25. If an assertion condition captured any substrings, they were not passed
1006 back unless some other capturing happened later. For example, if
1007 (?(?=(a))a) was matched against "a", no capturing was returned.
1008
1009 26. When studying a pattern that contained subroutine calls or assertions,
1010 the code for finding the minimum length of a possible match was handling
1011 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1012 group 1 called group 2 while simultaneously a separate group 2 called group
1013 1). A stack overflow occurred in this case. I have fixed this by limiting
1014 the recursion depth to 10.
1015
1016 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1017 Fortmann. This supports explicit test numbers on the command line, and has
1018 argument validation and error reporting.
1019
1020 28. An instance of \X with an unlimited repeat could fail if at any point the
1021 first character it looked at was a mark character.
1022
1023 29. Some minor code refactoring concerning Unicode properties and scripts
1024 should reduce the stack requirement of match() slightly.
1025
1026 30. Added the '=' option to pcretest to check the setting of unused capturing
1027 slots at the end of the pattern, which are documented as being -1, but are
1028 not included in the return count.
1029
1030 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1031 compiled something random. Now it gives a compile-time error (as does
1032 Perl).
1033
1034 32. A *MARK encountered during the processing of a positive assertion is now
1035 recorded and passed back (compatible with Perl).
1036
1037 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1038 had alternative anchored branches, the search for a second match in a line
1039 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1040 matched the line "0102" twice. The same bug affected patterns that started
1041 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1042 twice.
1043
1044 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1045 does, and because of capturing effects, quantifying parenthesized
1046 assertions may at times be useful. Quantifiers are now allowed for
1047 parenthesized assertions.
1048
1049 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1050
1051 36. \g was being checked for fancy things in a character class, when it should
1052 just be a literal "g".
1053
1054 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1055 appearance of a nested POSIX class supersedes an apparent external class.
1056 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1057 unescaped square brackets may also appear as part of class names. For
1058 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1059 more like Perl. (But see 8.20/1 above.)
1060
1061 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1062 was because it thought it was \N{name}, which is not supported).
1063
1064 39. Add minix to OS list not supporting the -S option in pcretest.
1065
1066 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1067 cannot analyze patterns in sufficient detail to catch mutual recursions
1068 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1069 subgroup is called recursively as a subpattern for a second time at the
1070 same position in the subject string. In previous releases this might have
1071 been caught by the recursion limit, or it might have run out of stack.
1072
1073 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1074 happen only once. PCRE was, however incorrectly giving a compile time error
1075 "recursive call could loop indefinitely" because it cannot analyze the
1076 pattern in sufficient detail. The compile time test no longer happens when
1077 PCRE is compiling a conditional subpattern, but actual runaway loops are
1078 now caught at runtime (see 40 above).
1079
1080 42. It seems that Perl allows any characters other than a closing parenthesis
1081 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1082 has been changed to be the same.
1083
1084 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1085 as not to get warnings when autogen.sh is called. Also changed
1086 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1087
1088 44. To help people who use pcregrep to scan files containing exceedingly long
1089 lines, the following changes have been made:
1090
1091 (a) The default value of the buffer size parameter has been increased from
1092 8K to 20K. (The actual buffer used is three times this size.)
1093
1094 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1095 PCRE is built.
1096
1097 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1098 to be set at run time.
1099
1100 (d) Numerical values in pcregrep options can be followed by K or M, for
1101 example --buffer-size=50K.
1102
1103 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1104 given and the return code is set to 2.
1105
1106 45. Add a pointer to the latest mark to the callout data block.
1107
1108 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1109 partial match of an empty string instead of no match. This was specific to
1110 the use of ".".
1111
1112 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1113 complete match instead of a partial match. This bug was dependent on both
1114 the PCRE_UTF8 and PCRE_DOTALL options being set.
1115
1116 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1117 starting byte set, because \b was not being ignored.
1118
1119
1120 Version 8.12 15-Jan-2011
1121 ------------------------
1122
1123 1. Fixed some typos in the markup of the man pages, and wrote a script that
1124 checks for such things as part of the documentation building process.
1125
1126 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1127 --match-limit and --recursion-limit options (added for 8.11). In
1128 particular, this made one of the standard tests fail. (The integer value
1129 went into the wrong half of a long int.)
1130
1131 3. If the --colour option was given to pcregrep with -v (invert match), it
1132 did strange things, either producing crazy output, or crashing. It should,
1133 of course, ignore a request for colour when reporting lines that do not
1134 match.
1135
1136 4. Another pcregrep bug caused similar problems if --colour was specified with
1137 -M (multiline) and the pattern match finished with a line ending.
1138
1139 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1140 matched in multiline mode, the following line was shown as part of the
1141 match. This seems wrong, so I have changed it.
1142
1143 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1144 the check for further matches in the same line (so they could be coloured)
1145 to overrun the end of the current line. If another match was found, it was
1146 incorrectly shown (and then shown again when found in the next line).
1147
1148 7. If pcregrep was compiled under Windows, there was a reference to the
1149 function pcregrep_exit() before it was defined. I am assuming this was
1150 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1151 reported by a user. I've moved the definition above the reference.
1152
1153
1154 Version 8.11 10-Dec-2010
1155 ------------------------
1156
1157 1. (*THEN) was not working properly if there were untried alternatives prior
1158 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1159 backtracked to try for "b" instead of moving to the next alternative branch
1160 at the same level (in this case, to look for "c"). The Perl documentation
1161 is clear that when (*THEN) is backtracked onto, it goes to the "next
1162 alternative in the innermost enclosing group".
1163
1164 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1165 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1166 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1167 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1168 (*THEN).
1169
1170 3. If \s appeared in a character class, it removed the VT character from
1171 the class, even if it had been included by some previous item, for example
1172 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1173 of \s, but is part of the POSIX "space" class.)
1174
1175 4. A partial match never returns an empty string (because you can always
1176 match an empty string at the end of the subject); however the checking for
1177 an empty string was starting at the "start of match" point. This has been
1178 changed to the "earliest inspected character" point, because the returned
1179 data for a partial match starts at this character. This means that, for
1180 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1181 (previously it gave "no match").
1182
1183 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1184 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1185 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1186 has an implication that the given string is incomplete (because a partial
1187 match is preferred over a full match). For this reason, these items now
1188 give a partial match in this situation. [Aside: previously, the one case
1189 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1190 match rather than a full match, which was wrong by the old rules, but is
1191 now correct.]
1192
1193 6. There was a bug in the handling of #-introduced comments, recognized when
1194 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1195 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1196 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1197 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1198 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1199 places in pcre_compile().
1200
1201 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1202 comments when looking ahead for named forward references to subpatterns,
1203 the only newline sequence it recognized was NL. It now handles newlines
1204 according to the set newline convention.
1205
1206 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1207 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1208 cater for a lack of strerror(). These oversights have been fixed.
1209
1210 9. Added --match-limit and --recursion-limit to pcregrep.
1211
1212 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1213
1214 11. When the -o option was used, pcregrep was setting a return code of 1, even
1215 when matches were found, and --line-buffered was not being honoured.
1216
1217 12. Added an optional parentheses number to the -o and --only-matching options
1218 of pcregrep.
1219
1220 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1221 can match an empty string. The code to do it in pcretest and pcredemo
1222 needed fixing:
1223
1224 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1225 only one byte after an empty string match just before CRLF (this case
1226 just got forgotten; "any" and "anycrlf" were OK).
1227
1228 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1229 mode when an empty string match preceded an ASCII character followed by
1230 a non-ASCII character. (The code for advancing by one character rather
1231 than one byte was nonsense.)
1232
1233 (c) The pcredemo.c sample program did not have any code at all to handle
1234 the cases when CRLF is a valid newline sequence.
1235
1236 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1237 as a starting offset was within the subject string. There is now a new
1238 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1239 negative or greater than the length of the string. In order to test this,
1240 pcretest is extended to allow the setting of negative starting offsets.
1241
1242 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1243 starting offset points to the beginning of a UTF-8 character was
1244 unnecessarily clumsy. I tidied it up.
1245
1246 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1247 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1248
1249 17. Nobody had reported that the --include_dir option, which was added in
1250 release 7.7 should have been called --include-dir (hyphen, not underscore)
1251 for compatibility with GNU grep. I have changed it to --include-dir, but
1252 left --include_dir as an undocumented synonym, and the same for
1253 --exclude-dir, though that is not available in GNU grep, at least as of
1254 release 2.5.4.
1255
1256 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1257 characters from a string of bytes) have been redefined so as not to use
1258 loops, in order to improve performance in some environments. At the same
1259 time, I abstracted some of the common code into auxiliary macros to save
1260 repetition (this should not affect the compiled code).
1261
1262 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1263 compile-time error is now given if \c is not followed by an ASCII
1264 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1265 different, and any byte value is allowed.)
1266
1267 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1268 START_OPTIMIZE option, which is now allowed at compile time - but just
1269 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1270 to pcregrep and other applications that have no direct access to PCRE
1271 options. The new /Y option in pcretest sets this option when calling
1272 pcre_compile().
1273
1274 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1275 back references. Groups containing recursive back references were forced to
1276 be atomic by that change, but in the case of named groups, the amount of
1277 memory required was incorrectly computed, leading to "Failed: internal
1278 error: code overflow". This has been fixed.
1279
1280 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1281 pcretest.c, to avoid build problems in some Borland environments.
1282
1283
1284 Version 8.10 25-Jun-2010
1285 ------------------------
1286
1287 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1288 THEN.
1289
1290 2. (*ACCEPT) was not working when inside an atomic group.
1291
1292 3. Inside a character class, \B is treated as a literal by default, but
1293 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1294 causes the error). The code is unchanged, but I tidied the documentation.
1295
1296 4. Inside a character class, PCRE always treated \R and \X as literals,
1297 whereas Perl faults them if its -w option is set. I have changed PCRE so
1298 that it faults them when PCRE_EXTRA is set.
1299
1300 5. Added support for \N, which always matches any character other than
1301 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1302
1303 6. When compiling pcregrep with newer versions of gcc which may have
1304 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1305 declared with attribute warn_unused_result" were given. Just casting the
1306 result to (void) does not stop the warnings; a more elaborate fudge is
1307 needed. I've used a macro to implement this.
1308
1309 7. Minor change to pcretest.c to avoid a compiler warning.
1310
1311 8. Added four artifical Unicode properties to help with an option to make
1312 \s etc use properties (see next item). The new properties are: Xan
1313 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1314
1315 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1316 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1317 this option. Modified pcretest to add /W to test this facility. Added
1318 REG_UCP to make it available via the POSIX interface.
1319
1320 10. Added --line-buffered to pcregrep.
1321
1322 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1323 studied, and the match started with a letter with a code point greater than
1324 127 whose first byte was different to the first byte of the other case of
1325 the letter, the other case of this starting letter was not recognized
1326 (#976).
1327
1328 12. If a pattern that was studied started with a repeated Unicode property
1329 test, for example, \p{Nd}+, there was the theoretical possibility of
1330 setting up an incorrect bitmap of starting bytes, but fortunately it could
1331 not have actually happened in practice until change 8 above was made (it
1332 added property types that matched character-matching opcodes).
1333
1334 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1335 possible starting bytes for non-anchored patterns.
1336
1337 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1338 \R, and also a number of cases that involve Unicode properties, both
1339 explicit and implicit when PCRE_UCP is set.
1340
1341 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1342 input, it could crash or give wrong results if characters with values
1343 greater than 0xc0 were present in the subject string. (Detail: it assumed
1344 UTF-8 input when processing these items.)
1345
1346 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1347 size_t is 64-bit (#991).
1348
1349 17. Added a check for running out of memory when PCRE is compiled with
1350 --disable-stack-for-recursion (#990).
1351
1352 18. If the last data line in a file for pcretest does not have a newline on
1353 the end, a newline was missing in the output.
1354
1355 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1356 less than 128) in its various bitmaps. However, there is a facility for
1357 generating tables according to the current locale when PCRE is compiled. It
1358 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1359 characters, are recognized by isspace() and therefore were getting set in
1360 these tables, and indeed these tables seem to approximate to ISO 8859. This
1361 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1362 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1363 which of course cannot start UTF-8 characters. I have changed the code so
1364 that only real ASCII characters (less than 128) and the correct starting
1365 bytes for UTF-8 encodings are set for characters greater than 127 when in
1366 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1367 altogether.)
1368
1369 20. Added the /T option to pcretest so as to be able to run tests with non-
1370 standard character tables, thus making it possible to include the tests
1371 used for 19 above in the standard set of tests.
1372
1373 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1374 reference to a subpattern the other side of a comment that contains an
1375 opening parenthesis caused either an internal compiling error, or a
1376 reference to the wrong subpattern.
1377
1378
1379 Version 8.02 19-Mar-2010
1380 ------------------------
1381
1382 1. The Unicode data tables have been updated to Unicode 5.2.0.
1383
1384 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1385 configured.
1386
1387 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1388 original author of that file, following a query about its status.
1389
1390 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1391 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1392
1393 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1394 quantifier applied to a forward-referencing subroutine call, could compile
1395 incorrect code or give the error "internal error: previously-checked
1396 referenced subpattern not found".
1397
1398 6. Both MS Visual Studio and Symbian OS have problems with initializing
1399 variables to point to external functions. For these systems, therefore,
1400 pcre_malloc etc. are now initialized to local functions that call the
1401 relevant global functions.
1402
1403 7. There were two entries missing in the vectors called coptable and poptable
1404 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1405 I've fixed the data, and added a kludgy way of testing at compile time that
1406 the lengths are correct (equal to the number of opcodes).
1407
1408 8. Following on from 7, I added a similar kludge to check the length of the
1409 eint vector in pcreposix.c.
1410
1411 9. Error texts for pcre_compile() are held as one long string to avoid too
1412 much relocation at load time. To find a text, the string is searched,
1413 counting zeros. There was no check for running off the end of the string,
1414 which could happen if a new error number was added without updating the
1415 string.
1416
1417 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1418
1419 11. \K was not working if it appeared in an atomic group or in a group that
1420 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1421 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1422 the assertion is positive, but not if it is negative.
1423
1424 12. Change 11 fortuitously reduced the size of the stack frame used in the
1425 "match()" function of pcre_exec.c by one pointer. Forthcoming
1426 implementation of support for (*MARK) will need an extra pointer on the
1427 stack; I have reserved it now, so that the stack frame size does not
1428 decrease.
1429
1430 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1431 item in branch that calls a recursion is a subroutine call - as in the
1432 second branch in the above example - was incorrectly given the compile-
1433 time error "recursive call could loop indefinitely" because pcre_compile()
1434 was not correctly checking the subroutine for matching a non-empty string.
1435
1436 14. The checks for overrunning compiling workspace could trigger after an
1437 overrun had occurred. This is a "should never occur" error, but it can be
1438 triggered by pathological patterns such as hundreds of nested parentheses.
1439 The checks now trigger 100 bytes before the end of the workspace.
1440
1441 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1442
1443
1444 Version 8.01 19-Jan-2010
1445 ------------------------
1446
1447 1. If a pattern contained a conditional subpattern with only one branch (in
1448 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1449 computed the wrong minimum data length (which is of course zero for such
1450 subpatterns). This could cause incorrect "no match" results.
1451
1452 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1453 the pattern is reset in the first branch, pcre_compile() failed with
1454 "internal error: code overflow at offset...". This happened only when
1455 the reset was to the original external option setting. (An optimization
1456 abstracts leading options settings into an external setting, which was the
1457 cause of this.)
1458
1459 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1460 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1461 assertion pattern did not match (meaning that the assertion was true), it
1462 was incorrectly treated as false if the SKIP had been reached during the
1463 matching. This also applied to assertions used as conditions.
1464
1465 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1466 assertion subpattern, including such a pattern used as a condition,
1467 unpredictable results occurred, instead of the error return
1468 PCRE_ERROR_DFA_UITEM.
1469
1470 5. The C++ GlobalReplace function was not working like Perl for the special
1471 situation when an empty string is matched. It now does the fancy magic
1472 stuff that is necessary.
1473
1474 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1475 removed. (These were left over from very, very early versions of PCRE.)
1476
1477 7. Some cosmetic changes to the code to make life easier when compiling it
1478 as part of something else:
1479
1480 (a) Change DEBUG to PCRE_DEBUG.
1481
1482 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1483 called "current" as "current_branch", to prevent a collision with the
1484 Linux macro when compiled as a kernel module.
1485
1486 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1487 prevent a collision with the Linux macro when compiled as a kernel
1488 module.
1489
1490 8. In pcre_compile() there are some checks for integer overflows that used to
1491 cast potentially large values to (double). This has been changed to that
1492 when building, a check for int64_t is made, and if it is found, it is used
1493 instead, thus avoiding the use of floating point arithmetic. (There is no
1494 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1495 double.
1496
1497 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1498 2005 (difference between two addresses compared to an unsigned value).
1499
1500 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1501 custom one, because of the following reported problem in Windows:
1502
1503 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1504 under Win32.
1505 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1506 therefore missing the function definition.
1507 - The compiler thus generates a "C" signature for the test function.
1508 - The linker fails to find the "C" function.
1509 - PCRE fails to configure if asked to do so against libbz2.
1510
1511 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1512 messages were output:
1513
1514 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1515 rerunning libtoolize, to keep the correct libtool macros in-tree.
1516 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1517
1518 I have done both of these things.
1519
1520 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1521 most of the time, it *can* run out if it is given a pattern that contains a
1522 runaway infinite recursion. I updated the discussion in the pcrestack man
1523 page.
1524
1525 13. Now that we have gone to the x.xx style of version numbers, the minor
1526 version may start with zero. Using 08 or 09 is a bad idea because users
1527 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1528 interpreted as invalid octal numbers. I've updated the previous comment in
1529 configure.ac, and also added a check that gives an error if 08 or 09 are
1530 used.
1531
1532 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1533 causing partial matching to fail when the end of the subject matched \W
1534 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1535
1536 15. There were some discrepancies between the declarations in pcre_internal.h
1537 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1538 their definitions. The declarations used "const uschar *" and the
1539 definitions used USPTR. Even though USPTR is normally defined as "const
1540 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1541 reported that: "This difference in casting confuses some C++ compilers, for
1542 example, SunCC recognizes above declarations as different functions and
1543 generates broken code for hbpcre." I have changed the declarations to use
1544 USPTR.
1545
1546 16. GNU libtool is named differently on some systems. The autogen.sh script now
1547 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1548 (FreeBSD).
1549
1550 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1551 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1552 comment: "Figure out how to create a longlong from a string: strtoll and
1553 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1554 instance, but it only takes 2 args instead of 3!"
1555
1556 18. A subtle bug concerned with back references has been fixed by a change of
1557 specification, with a corresponding code fix. A pattern such as
1558 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1559 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1560 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1561 same bug. Such groups have to be quantified to be useful, or contained
1562 inside another quantified group. (If there's no repetition, the reference
1563 can never match.) The problem arises because, having left the group and
1564 moved on to the rest of the pattern, a later failure that backtracks into
1565 the group uses the captured value from the final iteration of the group
1566 rather than the correct earlier one. I have fixed this in PCRE by forcing
1567 any group that contains a reference to itself to be an atomic group; that
1568 is, there cannot be any backtracking into it once it has completed. This is
1569 similar to recursive and subroutine calls.
1570
1571
1572 Version 8.00 19-Oct-09
1573 ----------------------
1574
1575 1. The table for translating pcre_compile() error codes into POSIX error codes
1576 was out-of-date, and there was no check on the pcre_compile() error code
1577 being within the table. This could lead to an OK return being given in
1578 error.
1579
1580 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1581 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1582 in a Windows environment.
1583
1584 3. The pcregrep --count option prints the count for each file even when it is
1585 zero, as does GNU grep. However, pcregrep was also printing all files when
1586 --files-with-matches was added. Now, when both options are given, it prints
1587 counts only for those files that have at least one match. (GNU grep just
1588 prints the file name in this circumstance, but including the count seems
1589 more useful - otherwise, why use --count?) Also ensured that the
1590 combination -clh just lists non-zero counts, with no names.
1591
1592 4. The long form of the pcregrep -F option was incorrectly implemented as
1593 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1594 but it seems right to fix it, and I didn't think it was worth preserving
1595 the old behaviour.
1596
1597 5. The command line items --regex=pattern and --regexp=pattern were not
1598 recognized by pcregrep, which required --regex pattern or --regexp pattern
1599 (with a space rather than an '='). The man page documented the '=' forms,
1600 which are compatible with GNU grep; these now work.
1601
1602 6. No libpcreposix.pc file was created for pkg-config; there was just
1603 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1604
1605 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1606 when UCP support is not needed, by modifying the Python script that
1607 generates it from Unicode data files. This should not matter if the module
1608 is correctly used as a library, but I received one complaint about 50K of
1609 unwanted data. My guess is that the person linked everything into his
1610 program rather than using a library. Anyway, it does no harm.
1611
1612 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1613 was a minimum greater than 1 for a wide character in a possessive
1614 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1615 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1616 character. Chaos in the form of incorrect output or a compiling loop could
1617 result.
1618
1619 9. The restrictions on what a pattern can contain when partial matching is
1620 requested for pcre_exec() have been removed. All patterns can now be
1621 partially matched by this function. In addition, if there are at least two
1622 slots in the offset vector, the offset of the earliest inspected character
1623 for the match and the offset of the end of the subject are set in them when
1624 PCRE_ERROR_PARTIAL is returned.
1625
1626 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1627 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1628 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1629 and may be more useful for multi-segment matching.
1630
1631 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1632 used to be given if ever the end of the subject was reached; now it is
1633 given only if matching could not proceed because another character was
1634 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1635 string "Z", which now yields "no match" instead of "partial match". In the
1636 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1637 final character ended with (*FAIL).
1638
1639 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1640 if the pattern had a "must contain" character that was already found in the
1641 earlier partial match, unless partial matching was again requested. For
1642 example, with the pattern /dog.(body)?/, the "must contain" character is
1643 "g". If the first part-match was for the string "dog", restarting with
1644 "sbody" failed. This bug has been fixed.
1645
1646 13. The string returned by pcre_dfa_exec() after a partial match has been
1647 changed so that it starts at the first inspected character rather than the
1648 first character of the match. This makes a difference only if the pattern
1649 starts with a lookbehind assertion or \b or \B (\K is not supported by
1650 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1651 matching functions compatible, and I think it's the right thing to do.
1652
1653 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1654 so that the demonstration program is easily available in environments where
1655 PCRE has not been installed from source.
1656
1657 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1658 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1659 library.
1660
1661 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1662 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1663 is not the first non-POSIX option to be added. Clearly some people find
1664 these options useful.
1665
1666 17. If a caller to the POSIX matching function regexec() passes a non-zero
1667 value for nmatch with a NULL value for pmatch, the value of
1668 nmatch is forced to zero.
1669
1670 18. RunGrepTest did not have a test for the availability of the -u option of
1671 the diff command, as RunTest does. It now checks in the same way as
1672 RunTest, and also checks for the -b option.
1673
1674 19. If an odd number of negated classes containing just a single character
1675 interposed, within parentheses, between a forward reference to a named
1676 subpattern and the definition of the subpattern, compilation crashed with
1677 an internal error, complaining that it could not find the referenced
1678 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1679 [The bug was that it was starting one character too far in when skipping
1680 over the character class, thus treating the ] as data rather than
1681 terminating the class. This meant it could skip too much.]
1682
1683 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1684 /g option in pcretest when the pattern contains \K, which makes it possible
1685 to have an empty string match not at the start, even when the pattern is
1686 anchored. Updated pcretest and pcredemo to use this option.
1687
1688 21. If the maximum number of capturing subpatterns in a recursion was greater
1689 than the maximum at the outer level, the higher number was returned, but
1690 with unset values at the outer level. The correct (outer level) value is
1691 now given.
1692
1693 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1694 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1695 make it do so. The string so far is captured, making this feature
1696 compatible with Perl.
1697
1698 23. The tests have been re-organized, adding tests 11 and 12, to make it
1699 possible to check the Perl 5.10 features against Perl 5.10.
1700
1701 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1702 pattern matches a fixed length string. PCRE did not allow this; now it
1703 does. Neither allows recursion.
1704
1705 25. I finally figured out how to implement a request to provide the minimum
1706 length of subject string that was needed in order to match a given pattern.
1707 (It was back references and recursion that I had previously got hung up
1708 on.) This code has now been added to pcre_study(); it finds a lower bound
1709 to the length of subject needed. It is not necessarily the greatest lower
1710 bound, but using it to avoid searching strings that are too short does give
1711 some useful speed-ups. The value is available to calling programs via
1712 pcre_fullinfo().
1713
1714 26. While implementing 25, I discovered to my embarrassment that pcretest had
1715 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1716 study optimizations had never been tested with that matching function.
1717 Oops. What is worse, even when it was passed study data, there was a bug in
1718 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1719 were also very few tests of studied patterns with pcre_dfa_exec().
1720
1721 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1722 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1723 on the other side of the coin, they are no longer allowed to have different
1724 names, because these cannot be distinguished in PCRE, and this has caused
1725 confusion. (This is a difference from Perl.)
1726
1727 28. When duplicate subpattern names are present (necessarily with different
1728 numbers, as required by 27 above), and a test is made by name in a
1729 conditional pattern, either for a subpattern having been matched, or for
1730 recursion in such a pattern, all the associated numbered subpatterns are
1731 tested, and the overall condition is true if the condition is true for any
1732 one of them. This is the way Perl works, and is also more like the way
1733 testing by number works.
1734
1735
1736 Version 7.9 11-Apr-09
1737 ---------------------
1738
1739 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1740 (pcretest), all targets were linked against these libraries. This included
1741 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1742 libraries. This caused unwanted dependencies to be created. This problem
1743 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1744 pcretest is linked with readline.
1745
1746 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1747 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1748 moved outside it again, because FALSE and TRUE are already defined in AIX,
1749 but BOOL is not.
1750
1751 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1752 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1753
1754 4. The pcregrep documentation said spaces were inserted as well as colons (or
1755 hyphens) following file names and line numbers when outputting matching
1756 lines. This is not true; no spaces are inserted. I have also clarified the
1757 wording for the --colour (or --color) option.
1758
1759 5. In pcregrep, when --colour was used with -o, the list of matching strings
1760 was not coloured; this is different to GNU grep, so I have changed it to be
1761 the same.
1762
1763 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1764 each matching line was coloured. Now it goes on to look for further matches
1765 of any of the test patterns, which is the same behaviour as GNU grep.
1766
1767 7. A pattern that could match an empty string could cause pcregrep to loop; it
1768 doesn't make sense to accept an empty string match in pcregrep, so I have
1769 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1770 seems to be how GNU grep behaves. [But see later change 40 for release
1771 8.33.]
1772
1773 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1774 start or after a newline", because the conditional assertion was not being
1775 correctly handled. The rule now is that both the assertion and what follows
1776 in the first alternative must satisfy the test.
1777
1778 9. If auto-callout was enabled in a pattern with a conditional group whose
1779 condition was an assertion, PCRE could crash during matching, both with
1780 pcre_exec() and pcre_dfa_exec().
1781
1782 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1783 used for matching.
1784
1785 11. Unicode property support in character classes was not working for
1786 characters (bytes) greater than 127 when not in UTF-8 mode.
1787
1788 12. Added the -M command line option to pcretest.
1789
1790 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1791
1792 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1793
1794 16. Added comments and documentation about mis-use of no_arg in the C++
1795 wrapper.
1796
1797 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1798 from Martin Jerabek that uses macro names for all relevant character and
1799 string constants.
1800
1801 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1802 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1803 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1804 these, but not everybody uses configure.
1805
1806 19. A conditional group that had only one branch was not being correctly
1807 recognized as an item that could match an empty string. This meant that an
1808 enclosing group might also not be so recognized, causing infinite looping
1809 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1810 with the subject "ab", where knowledge that the repeated group can match
1811 nothing is needed in order to break the loop.
1812
1813 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1814 exec(), but without supplying a callout function, matching went wrong.
1815
1816 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1817 leak if the size of the offset vector was greater than 30. When the vector
1818 is smaller, the saved offsets during recursion go onto a local stack
1819 vector, but for larger vectors malloc() is used. It was failing to free
1820 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1821 error, in fact).
1822
1823 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1824 heapframe that is used only when UTF-8 support is enabled. This caused no
1825 problem, but was untidy.
1826
1827 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1828 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1829 included within another project.
1830
1831 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1832 slightly modified by me:
1833
1834 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1835 not building pcregrep.
1836
1837 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1838 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1839
1840 25. Forward references, both numeric and by name, in patterns that made use of
1841 duplicate group numbers, could behave incorrectly or give incorrect errors,
1842 because when scanning forward to find the reference group, PCRE was not
1843 taking into account the duplicate group numbers. A pattern such as
1844 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1845
1846 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1847 the feature of a custom pointer more persuasive (as requested by a user).
1848
1849 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1850 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1851 pre-defined.
1852
1853 28. Added support for (*UTF8) at the start of a pattern.
1854
1855 29. Arrange for flags added by the "release type" setting in CMake to be shown
1856 in the configuration summary.
1857
1858
1859 Version 7.8 05-Sep-08
1860 ---------------------
1861
1862 1. Replaced UCP searching code with optimized version as implemented for Ad
1863 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1864 stage table and inline lookup instead of a function, giving speed ups of 2
1865 to 5 times on some simple patterns that I tested. Permission was given to
1866 distribute the MultiStage2.py script that generates the tables (it's not in
1867 the tarball, but is in the Subversion repository).
1868
1869 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1870 scripts.
1871
1872 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1873 a group with a zero qualifier. The result of the study could be incorrect,
1874 or the function might crash, depending on the pattern.
1875
1876 4. Caseless matching was not working for non-ASCII characters in back
1877 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1878 It now works when Unicode Property Support is available.
1879
1880 5. In pcretest, an escape such as \x{de} in the data was always generating
1881 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1882 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1883 truncation.
1884
1885 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1886
1887 7. Added two (int) casts to pcregrep when printing the difference of two
1888 pointers, in case they are 64-bit values.
1889
1890 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1891 test 2 if it fails.
1892
1893 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1894 and a #define of that name to empty if it is not externally set. This is to
1895 allow users of MSVC to set it if necessary.
1896
1897 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1898 the convenience functions in the pcre_get.c source file.
1899
1900 11. An option change at the start of a pattern that had top-level alternatives
1901 could cause overwriting and/or a crash. This command provoked a crash in
1902 some environments:
1903
1904 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1905
1906 This potential security problem was recorded as CVE-2008-2371.
1907
1908 12. For a pattern where the match had to start at the beginning or immediately
1909 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1910 pcre_dfa_exec() could read past the end of the passed subject if there was
1911 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1912 pcretest so that it places the subject at the end of its malloc-ed buffer.
1913
1914 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1915 exec() might read past the end of the data buffer in UTF-8 mode.
1916
1917 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1918 the data contained the byte 0x85 as part of a UTF-8 character within its
1919 first line. This applied both to normal and DFA matching.
1920
1921 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1922 /^[^d]*?$/8 failed to match "abc".
1923
1924 16. Added a missing copyright notice to pcrecpp_internal.h.
1925
1926 17. Make it more clear in the documentation that values returned from
1927 pcre_exec() in ovector are byte offsets, not character counts.
1928
1929 18. Tidied a few places to stop certain compilers from issuing warnings.
1930
1931 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1932 supplied by Stefan Weber. I made a further small update for 7.8 because
1933 there is a change of source arrangements: the pcre_searchfuncs.c module is
1934 replaced by pcre_ucd.c.
1935
1936
1937 Version 7.7 07-May-08
1938 ---------------------
1939
1940 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1941 a string to a long long, pretend we don't even have a long long." This is
1942 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1943
1944 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1945 pre-7.6 versions, which defined a global no_arg variable instead of putting
1946 it in the RE class. (See also #8 below.)
1947
1948 3. Remove a line of dead code, identified by coverity and reported by Nuno
1949 Lopes.
1950
1951 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1952
1953 (1) The include/exclude patterns were being applied to the whole pathnames
1954 of files, instead of just to the final components.
1955
1956 (2) If there was more than one level of directory, the subdirectories were
1957 skipped unless they satisfied the include/exclude conditions. This is
1958 inconsistent with GNU grep (and could even be seen as contrary to the
1959 pcregrep specification - which I improved to make it absolutely clear).
1960 The action now is always to scan all levels of directory, and just
1961 apply the include/exclude patterns to regular files.
1962
1963 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1964 --exclude_dir in the tests to avoid scanning .svn directories.
1965
1966 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1967 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1968 doesn't support NULs in patterns.
1969
1970 7. Added some missing "const"s to declarations of static tables in
1971 pcre_compile.c and pcre_dfa_exec.c.
1972
1973 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1974 caused by fix #2 above. (Subsequently also a second patch to fix the
1975 first patch. And a third patch - this was a messy problem.)
1976
1977 9. Applied Craig's patch to remove the use of push_back().
1978
1979 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1980 matching function regexec().
1981
1982 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1983 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1984 references. PCRE supports relative numbers with this syntax (I don't think
1985 Oniguruma does).
1986
1987 12. Previously, a group with a zero repeat such as (...){0} was completely
1988 omitted from the compiled regex. However, this means that if the group
1989 was called as a subroutine from elsewhere in the pattern, things went wrong
1990 (an internal error was given). Such groups are now left in the compiled
1991 pattern, with a new opcode that causes them to be skipped at execution
1992 time.
1993
1994 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1995 to the way PCRE behaves:
1996
1997 (a) A lone ] character is dis-allowed (Perl treats it as data).
1998
1999 (b) A back reference to an unmatched subpattern matches an empty string
2000 (Perl fails the current match path).
2001
2002 (c) A data ] in a character class must be notated as \] because if the
2003 first data character in a class is ], it defines an empty class. (In
2004 Perl it is not possible to have an empty class.) The empty class []
2005 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2006 The negative empty class [^] matches any one character, independently
2007 of the DOTALL setting.
2008
2009 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2010 non-existent subpattern following a character class starting with ']' and
2011 containing () gave an internal compiling error instead of "reference to
2012 non-existent subpattern". Fortunately, when the pattern did exist, the
2013 compiled code was correct. (When scanning forwards to check for the
2014 existence of the subpattern, it was treating the data ']' as terminating
2015 the class, so got the count wrong. When actually compiling, the reference
2016 was subsequently set up correctly.)
2017
2018 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2019 it was being rejected as not supported by pcre_dfa_exec(), even though
2020 other assertions are supported. I have made pcre_dfa_exec() support
2021 (*FAIL).
2022
2023 16. The implementation of 13c above involved the invention of a new opcode,
2024 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2025 cannot be changed at match time, I realized I could make a small
2026 improvement to matching performance by compiling OP_ALLANY instead of
2027 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2028 on the OP_ANY path.
2029
2030 17. Compiling pcretest on Windows with readline support failed without the
2031 following two fixes: (1) Make the unistd.h include conditional on
2032 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2033
2034 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2035 ncurses library to be included for pcretest when ReadLine support is
2036 requested, but also to allow for it to be overridden. This patch came from
2037 Daniel Bergström.
2038
2039 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2040 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2041 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2042 spotting this.
2043
2044
2045 Version 7.6 28-Jan-08
2046 ---------------------
2047
2048 1. A character class containing a very large number of characters with
2049 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2050 overflow.
2051
2052 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2053 HAVE_LONG_LONG is not defined.
2054
2055 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2056 bring them up to date and include new features. This patch includes:
2057
2058 - Fixed PH's badly added libz and libbz2 support.
2059 - Fixed a problem with static linking.
2060 - Added pcredemo. [But later removed - see 7 below.]
2061 - Fixed dftables problem and added an option.
2062 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2063 HAVE_LONG_LONG.
2064 - Added readline support for pcretest.
2065 - Added an listing of the option settings after cmake has run.
2066
2067 4. A user submitted a patch to Makefile that makes it easy to create
2068 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2069 Makefile.am that cause it to include this special target, without
2070 affecting anything else. Note that the same mingw target plus all
2071 the other distribution libraries and programs are now supported
2072 when configuring with CMake (see 6 below) instead of with
2073 Configure/Make.
2074
2075 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2076 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2077 exported in the Windows port". It has not yet been confirmed that the patch
2078 solves the problem, but it does no harm.
2079
2080 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2081 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2082 with CMake, and also correct the comment about stack recursion.
2083
2084 7. Remove the automatic building of pcredemo from the ./configure system and
2085 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2086 of a program that users should build themselves after PCRE is installed, so
2087 building it automatically is not really right. What is more, it gave
2088 trouble in some build environments.
2089
2090 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2091
2092
2093 Version 7.5 10-Jan-08
2094 ---------------------
2095
2096 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2097 values in parens when parsing an RE using the C++ wrapper."
2098
2099 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2100 Characters greater than 255 were excluded from the class instead of being
2101 included.
2102
2103 3. The same bug as (2) above applied to negated POSIX classes such as
2104 [:^space:].
2105
2106 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2107 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2108 I have changed it.
2109
2110 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2111 first named subpattern) and a construct such as (?&a) would reference the
2112 first named subpattern whose name started with "a" (in other words, the
2113 length check was missing). Both these problems are fixed. "Subpattern name
2114 expected" is now given for (?&) (a zero-length name), and this patch also
2115 makes it give the same error for \k'' (previously it complained that that
2116 was a reference to a non-existent subpattern).
2117
2118 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2119 this is right because (?- can be followed by option settings as well as by
2120 digits. I have, however, made the messages clearer.
2121
2122 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2123 than the number used in the conditional) now cause a compile-time error.
2124 This is actually not compatible with Perl, which accepts such patterns, but
2125 treats the conditional as always being FALSE (as PCRE used to), but it
2126 seems to me that giving a diagnostic is better.
2127
2128 8. Change "alphameric" to the more common word "alphanumeric" in comments
2129 and messages.
2130
2131 9. Fix two occurrences of "backslash" in comments that should have been
2132 "backspace".
2133
2134 10. Remove two redundant lines of code that can never be obeyed (their function
2135 was moved elsewhere).
2136
2137 11. The program that makes PCRE's Unicode character property table had a bug
2138 which caused it to generate incorrect table entries for sequences of
2139 characters that have the same character type, but are in different scripts.
2140 It amalgamated them into a single range, with the script of the first of
2141 them. In other words, some characters were in the wrong script. There were
2142 thirteen such cases, affecting characters in the following ranges:
2143
2144 U+002b0 - U+002c1
2145 U+0060c - U+0060d
2146 U+0061e - U+00612
2147 U+0064b - U+0065e
2148 U+0074d - U+0076d
2149 U+01800 - U+01805
2150 U+01d00 - U+01d77
2151 U+01d9b - U+01dbf
2152 U+0200b - U+0200f
2153 U+030fc - U+030fe
2154 U+03260 - U+0327f
2155 U+0fb46 - U+0fbb1
2156 U+10450 - U+1049d
2157
2158 12. The -o option (show only the matching part of a line) for pcregrep was not
2159 compatible with GNU grep in that, if there was more than one match in a
2160 line, it showed only the first of them. It now behaves in the same way as
2161 GNU grep.
2162
2163 13. If the -o and -v options were combined for pcregrep, it printed a blank
2164 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2165 does the same. The return code can be used to tell if there were any
2166 non-matching lines.
2167
2168 14. Added --file-offsets and --line-offsets to pcregrep.
2169
2170 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2171 infinitely looping recursion. The bug was that positive lookaheads were not
2172 being skipped when checking for a possible empty match (negative lookaheads
2173 and both kinds of lookbehind were skipped).
2174
2175 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2176 inclusion of <windows.h> to before rather than after the definition of
2177 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2178
2179 17. Specifying a possessive quantifier with a specific limit for a Unicode
2180 character property caused pcre_compile() to compile bad code, which led at
2181 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2182 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2183 caused the error; without that there was no problem.
2184
2185 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2186
2187 19. Added --enable-pcretest-libreadline.
2188
2189 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2190 RE::GlobalReplace(). As a result, the number of replacements returned was
2191 double what it should be. I removed one of the increments, but Craig sent a
2192 later patch that removed the other one (the right fix) and added unit tests
2193 that check the return values (which was not done before).
2194
2195 21. Several CMake things:
2196
2197 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2198 the names libpcre and libpcreposix, not just pcre and pcreposix.
2199
2200 (2) The above change means that pcretest and pcregrep are now correctly
2201 linked with the newly-built libraries, not previously installed ones.
2202
2203 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2204
2205 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2206 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2207 UTF-8 newline character). The key issue is that the pattern starts .*;
2208 this means that the match must be either at the beginning, or after a
2209 newline. The bug was in the code for advancing after a failed match and
2210 checking that the new position followed a newline. It was not taking
2211 account of UTF-8 characters correctly.
2212
2213 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2214 character classes. PCRE was not treating the sequence [:...:] as a
2215 character class unless the ... were all letters. Perl, however, seems to
2216 allow any characters between [: and :], though of course it rejects as
2217 unknown any "names" that contain non-letters, because all the known class
2218 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2219 for example, whereas PCRE did not - it did not recognize a POSIX character
2220 class. This seemed a bit dangerous, so the code has been changed to be
2221 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2222 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2223 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2224 Perl does, and where it didn't before.
2225
2226 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2227 Windows environments %n is disabled by default.
2228
2229
2230 Version 7.4 21-Sep-07
2231 ---------------------
2232
2233 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2234 means that a class such as [\s] counted as "explicit reference to CR or
2235 LF". That isn't really right - the whole point of the change was to try to
2236 help when there was an actual mention of one of the two characters. So now
2237 the change happens only if \r or \n (or a literal CR or LF) character is
2238 encountered.
2239
2240 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2241 of both had grown to the point where there were only 3 bits left.
2242 Fortunately, there was spare space in the data structure, and so I have
2243 moved the internal flags into a new 16-bit field to free up more option
2244 bits.
2245
2246 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2247 but did not set the internal JCHANGED flag - either of these is enough to
2248 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2249 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2250 start sets both bits.
2251
2252 4. Added options (at build time, compile time, exec time) to change \R from
2253 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2254
2255 5. doc/pcresyntax.html was missing from the distribution.
2256
2257 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2258 compatibility, even though it is no longer used.
2259
2260 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2261 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2262 windows.h file is present (where different names are used). [This was
2263 reversed later after testing - see 16 below.]
2264
2265 8. Changed all #include <config.h> to #include "config.h". There were also
2266 some further <pcre.h> cases that I changed to "pcre.h".
2267
2268 9. When pcregrep was used with the --colour option, it missed the line ending
2269 sequence off the lines that it output.
2270
2271 10. It was pointed out to me that arrays of string pointers cause lots of
2272 relocations when a shared library is dynamically loaded. A technique of
2273 using a single long string with a table of offsets can drastically reduce
2274 these. I have refactored PCRE in four places to do this. The result is
2275 dramatic:
2276
2277 Originally: 290
2278 After changing UCP table: 187
2279 After changing error message table: 43
2280 After changing table of "verbs" 36
2281 After changing table of Posix names 22
2282
2283 Thanks to the folks working on Gregex for glib for this insight.
2284
2285 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2286 unicode-properties was also set.
2287
2288 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2289
2290 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2291 checked only for CRLF.
2292
2293 14. Added casts to pcretest.c to avoid compiler warnings.
2294
2295 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2296
2297 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2298 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2299 entirely. This removes changes made in 7 above.
2300
2301 17. The CMake files have been updated, and there is now more information about
2302 building with CMake in the NON-UNIX-USE document.
2303
2304
2305 Version 7.3 28-Aug-07
2306 ---------------------
2307
2308 1. In the rejigging of the build system that eventually resulted in 7.1, the
2309 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2310 brackets there is not right, since it causes compilers to look for an
2311 installed pcre.h, not the version that is in the source that is being
2312 compiled (which of course may be different). I have changed it back to:
2313
2314 #include "pcre.h"
2315
2316 I have a vague recollection that the change was concerned with compiling in
2317 different directories, but in the new build system, that is taken care of
2318 by the VPATH setting the Makefile.
2319
2320 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2321 when the subject happened to end in the byte 0x85 (e.g. if the last
2322 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2323 characters but of course it shouldn't be taken as a newline when it is part
2324 of another character. The bug was that, for an unlimited repeat of . in
2325 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2326 characters when looking for a newline.
2327
2328 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2329
2330 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2331 in debug output.
2332
2333 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2334 long printing in the pcrecpp unittest when running under MinGW.
2335
2336 6. ESC_K was left out of the EBCDIC table.
2337
2338 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2339 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2340 limit also applies to "virtual nesting" when a pattern is recursive, and in
2341 this case 1000 isn't so big. I have been able to remove this limit at the
2342 expense of backing off one optimization in certain circumstances. Normally,
2343 when pcre_exec() would call its internal match() function recursively and
2344 immediately return the result unconditionally, it uses a "tail recursion"
2345 feature to save stack. However, when a subpattern that can match an empty
2346 string has an unlimited repetition quantifier, it no longer makes this
2347 optimization. That gives it a stack frame in which to save the data for
2348 checking that an empty string has been matched. Previously this was taken
2349 from the 1000-entry workspace that had been reserved. So now there is no
2350 explicit limit, but more stack is used.
2351
2352 8. Applied Daniel's patches to solve problems with the import/export magic
2353 syntax that is required for Windows, and which was going wrong for the
2354 pcreposix and pcrecpp parts of the library. These were overlooked when this
2355 problem was solved for the main library.
2356
2357 9. There were some crude static tests to avoid integer overflow when computing
2358 the size of patterns that contain repeated groups with explicit upper
2359 limits. As the maximum quantifier is 65535, the maximum group length was
2360 set at 30,000 so that the product of these two numbers did not overflow a
2361 32-bit integer. However, it turns out that people want to use groups that
2362 are longer than 30,000 bytes (though not repeat them that many times).
2363 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2364 made it possible to implement the integer overflow checks in a much more
2365 dynamic way, which I have now done. The artificial limitation on group
2366 length has been removed - we now have only the limit on the total length of
2367 the compiled pattern, which depends on the LINK_SIZE setting.
2368
2369 10. Fixed a bug in the documentation for get/copy named substring when
2370 duplicate names are permitted. If none of the named substrings are set, the
2371 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2372 empty string.
2373
2374 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2375 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2376 because the ] is interpreted as the first data character and the
2377 terminating ] is not found. PCRE has been made compatible with Perl in this
2378 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2379 cause memory overwriting.
2380
2381 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2382 string has been matched (to stop an infinite loop). It was not recognizing
2383 a conditional subpattern that could match an empty string if that
2384 subpattern was within another subpattern. For example, it looped when
2385 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2386 condition was not nested. This bug has been fixed.
2387
2388 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2389 past the start of the subject in the presence of bytes with the top bit
2390 set, for example "\x8aBCD".
2391
2392 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2393 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2394
2395 14. Optimized (?!) to (*FAIL).
2396
2397 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2398 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2399 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2400 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2401 does: it's just the validity check that is more restrictive.
2402
2403 16. Inserted checks for integer overflows during escape sequence (backslash)
2404 processing, and also fixed erroneous offset values for syntax errors during
2405 backslash processing.
2406
2407 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2408 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2409
2410 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2411 caused an overrun.
2412
2413 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2414 something other than just ASCII characters) inside a group that had an
2415 unlimited repeat caused a loop at compile time (while checking to see
2416 whether the group could match an empty string).
2417
2418 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2419 [\P{Any}] did so. (Error in the code for printing property names.)
2420
2421 21. An orphan \E inside a character class could cause a crash.
2422
2423 22. A repeated capturing bracket such as (A)? could cause a wild memory
2424 reference during compilation.
2425
2426 23. There are several functions in pcre_compile() that scan along a compiled
2427 expression for various reasons (e.g. to see if it's fixed length for look
2428 behind). There were bugs in these functions when a repeated \p or \P was
2429 present in the pattern. These operators have additional parameters compared
2430 with \d, etc, and these were not being taken into account when moving along
2431 the compiled data. Specifically:
2432
2433 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2434 length.
2435
2436 (b) An item such as \pL+ within a repeated group could cause crashes or
2437 loops.
2438
2439 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2440 "reference to non-existent subpattern" error.
2441
2442 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2443
2444 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2445 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2446
2447 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2448
2449 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2450 character were causing crashes (broken optimization).
2451
2452 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2453 \p or \P) caused a compile-time loop.
2454
2455 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2456 break. For example, the unstudied pattern [\r\n]A does not match the string
2457 "\r\nA" because change 7.0/46 below moves the current point on by two
2458 characters after failing to match at the start. However, the pattern \nA
2459 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2460 the same is true. There doesn't seem any very clean way out of this, but
2461 what I have chosen to do makes the common cases work: PCRE now takes note
2462 of whether there can be an explicit match for \r or \n anywhere in the
2463 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2464 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2465 pattern has explicit CR or LF references.
2466
2467 29. Added (*CR) etc for changing newline setting at start of pattern.
2468
2469
2470 Version 7.2 19-Jun-07
2471 ---------------------
2472
2473 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2474 which is apparently normally available under Windows.
2475
2476 2. Re-jig the pcregrep tests with different newline settings in an attempt
2477 to make them independent of the local environment's newline setting.
2478
2479 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2480
2481 4. Some of the "internals" tests were previously cut out when the link size
2482 was not 2, because the output contained actual offsets. The recent new
2483 "Z" feature of pcretest means that these can be cut out, making the tests
2484 usable with all link sizes.
2485
2486 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2487 stack recursion. This gives a massive performance boost under BSD, but just
2488 a small improvement under Linux. However, it saves one field in the frame
2489 in all cases.
2490
2491 6. Added more features from the forthcoming Perl 5.10:
2492
2493 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2494 recursion call. It refers to the nth most recently opened parentheses.
2495
2496 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2497 to be opened parentheses.
2498
2499 (c) Conditions that refer to capturing parentheses can be specified
2500 relatively, for example, (?(-2)... or (?(+3)...
2501
2502 (d) \K resets the start of the current match so that everything before
2503 is not part of it.
2504
2505 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2506
2507 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2508 reference syntax.
2509
2510 (g) (?| introduces a group in which the numbering of parentheses in each
2511 alternative starts with the same number.
2512
2513 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2514
2515 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2516 PCRE_INFO_JCHANGED.
2517
2518 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2519 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2520 for detecting groups that can match an empty string.
2521
2522 9. A pattern with a very large number of alternatives (more than several
2523 hundred) was running out of internal workspace during the pre-compile
2524 phase, where pcre_compile() figures out how much memory will be needed. A
2525 bit of new cunning has reduced the workspace needed for groups with
2526 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2527 workspace instead of running out of the 4096 that are available.
2528
2529 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2530
2531 11. Applied patch from Google to remove an optimization that didn't quite work.
2532 The report of the bug said:
2533
2534 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2535 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2536 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2537
2538 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2539 it matched the wrong number of bytes.
2540
2541
2542 Version 7.1 24-Apr-07
2543 ---------------------
2544
2545 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2546 that is more "standard", making use of automake and other Autotools. There
2547 is some re-arrangement of the files and adjustment of comments consequent
2548 on this.
2549
2550 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2551 for recursive directory scanning broke on some systems because the files
2552 are not scanned in any specific order and on different systems the order
2553 was different. A call to "sort" has been inserted into RunGrepTest for the
2554 approprate test as a short-term fix. In the longer term there may be an
2555 alternative.
2556
2557 3. I had an email from Eric Raymond about problems translating some of PCRE's
2558 man pages to HTML (despite the fact that I distribute HTML pages, some
2559 people do their own conversions for various reasons). The problems
2560 concerned the use of low-level troff macros .br and .in. I have therefore
2561 removed all such uses from the man pages (some were redundant, some could
2562 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2563 HTML has been updated to handle .nf/.fi and to complain if it encounters
2564 .br or .in.
2565
2566 4. Updated comments in configure.ac that get placed in config.h.in and also
2567 arranged for config.h to be included in the distribution, with the name
2568 config.h.generic, for the benefit of those who have to compile without
2569 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2570
2571 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2572 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2573 makevp.bat for the current PCRE, using the additional files
2574 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2575
2576 6. A Windows user reported a minor discrepancy with test 2, which turned out
2577 to be caused by a trailing space on an input line that had got lost in his
2578 copy. The trailing space was an accident, so I've just removed it.
2579
2580 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2581 that is needed.
2582
2583 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2584 as "const" (a) because they are and (b) because it helps the PHP
2585 maintainers who have recently made a script to detect big data structures
2586 in the php code that should be moved to the .rodata section. I remembered
2587 to update Builducptable as well, so it won't revert if ucptable.h is ever
2588 re-created.
2589
2590 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2591 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2592 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2593 support is not required. This saves 1.5-2K of code, which is important in
2594 some applications.
2595
2596 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2597 so as not to refer to the tables, even though these functions will never be
2598 called when UTF-8 support is disabled. Otherwise there are problems with a
2599 shared library.
2600
2601 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2602
2603 (a) It was defining its arguments as char * instead of void *.
2604
2605 (b) It was assuming that all moves were upwards in memory; this was true
2606 a long time ago when I wrote it, but is no longer the case.
2607
2608 The emulated memove() is provided for those environments that have neither
2609 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2610 is clearly not the case, as these two bugs were recently reported.
2611
2612 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2613 and Detrail to create the HTML documentation, the .txt form of the man
2614 pages, and it removes trailing spaces from listed files. It also creates
2615 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2616 case, it wraps all the #defines with #ifndefs. This script should be run
2617 before "make dist".
2618
2619 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2620 with Unicode property support.
2621
2622 (a) For a maximizing quantifier, if the two different cases of the
2623 character were of different lengths in their UTF-8 codings (there are
2624 some cases like this - I found 11), and the matching function had to
2625 back up over a mixture of the two cases, it incorrectly assumed they
2626 were both the same length.
2627
2628 (b) When PCRE was configured to use the heap rather than the stack for
2629 recursion during matching, it was not correctly preserving the data for
2630 the other case of a UTF-8 character when checking ahead for a match
2631 while processing a minimizing repeat. If the check also involved
2632 matching a wide character, but failed, corruption could cause an
2633 erroneous result when trying to check for a repeat of the original
2634 character.
2635
2636 13. Some tidying changes to the testing mechanism:
2637
2638 (a) The RunTest script now detects the internal link size and whether there
2639 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2640 values substituted by "configure". (The RunGrepTest script already did
2641 this for UTF-8.) The configure.ac script no longer substitutes the
2642 relevant variables.
2643
2644 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2645 with length and offset values. This means that the output is different
2646 for different internal link sizes. Test 2 is skipped for link sizes
2647 other than 2 because of this, bypassing the problem. Unfortunately,
2648 there was also a test in test 3 (the locale tests) that used /B and
2649 failed for link sizes other than 2. Rather than cut the whole test out,
2650 I have added a new /Z option to pcretest that replaces the length and
2651 offset values with spaces. This is now used to make test 3 independent
2652 of link size. (Test 2 will be tidied up later.)
2653
2654 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2655 segmentation fault instead of returning the appropriate error message.
2656
2657 15. In multiline mode when the newline sequence was set to "any", the pattern
2658 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2659 This doesn't seem right; it now treats the CRLF combination as the line
2660 ending, and so does not match in that case. It's only a pattern such as ^$
2661 that would hit this one: something like ^ABC$ would have failed after \r
2662 and then tried again after \r\n.
2663
2664 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2665 in an attempt to make files that differ only in their line terminators
2666 compare equal. This works on Linux.
2667
2668 17. Under certain error circumstances pcregrep might try to free random memory
2669 as it exited. This is now fixed, thanks to valgrind.
2670
2671 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2672 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2673 was because its rules for how to advance for /g after matching an empty
2674 string at the end of a line did not allow for this case. They now check for
2675 it specially.
2676
2677 20. pcretest is supposed to handle patterns and data of any length, by
2678 extending its buffers when necessary. It was getting this wrong when the
2679 buffer for a data line had to be extended.
2680
2681 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2682 CRLF as a newline sequence.
2683
2684 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2685 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2686 I have nevertheless tidied it up.
2687
2688 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2689
2690 24. Added a man page for pcre-config.
2691
2692
2693 Version 7.0 19-Dec-06
2694 ---------------------
2695
2696 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2697 moving to gcc 4.1.1.
2698
2699 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2700 sys/time.h, which is documented as needed for this function. It doesn't
2701 seem to matter on Linux, but it showed up on some releases of OS X.
2702
2703 3. It seems that there are systems where bytes whose values are greater than
2704 127 match isprint() in the "C" locale. The "C" locale should be the
2705 default when a C program starts up. In most systems, only ASCII printing
2706 characters match isprint(). This difference caused the output from pcretest
2707 to vary, making some of the tests fail. I have changed pcretest so that:
2708
2709 (a) When it is outputting text in the compiled version of a pattern, bytes
2710 other than 32-126 are always shown as hex escapes.
2711
2712 (b) When it is outputting text that is a matched part of a subject string,
2713 it does the same, unless a different locale has been set for the match
2714 (using the /L modifier). In this case, it uses isprint() to decide.
2715
2716 4. Fixed a major bug that caused incorrect computation of the amount of memory
2717 required for a compiled pattern when options that changed within the
2718 pattern affected the logic of the preliminary scan that determines the
2719 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2720 that the computed length was too small. The symptoms of this bug were
2721 either the PCRE error "internal error: code overflow" from pcre_compile(),
2722 or a glibc crash with a message such as "pcretest: free(): invalid next
2723 size (fast)". Examples of patterns that provoked this bug (shown in
2724 pcretest format) are:
2725
2726 /(?-x: )/x
2727 /(?x)(?-x: \s*#\s*)/
2728 /((?i)[\x{c0}])/8
2729 /(?i:[\x{c0}])/8
2730
2731 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2732 is now done differently.
2733
2734 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2735 wrapper classes; (b) implement a new function in the C++ scanner that is
2736 more efficient than the old way of doing things because it avoids levels of
2737 recursion in the regex matching; (c) add a paragraph to the documentation
2738 for the FullMatch() function.
2739
2740 6. The escape sequence \n was being treated as whatever was defined as
2741 "newline". Not only was this contrary to the documentation, which states
2742 that \n is character 10 (hex 0A), but it also went horribly wrong when
2743 "newline" was defined as CRLF. This has been fixed.
2744
2745 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2746 was being set to -1 for the "end of line" case (supposedly a value that no
2747 character can have). Though this value is never used (the check for end of
2748 line is "zero bytes in current character"), it caused compiler complaints.
2749 I've changed it to 0xffffffff.
2750
2751 8. In pcre_version.c, the version string was being built by a sequence of
2752 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2753 string (as it is for production releases) called a macro with an empty
2754 argument. The C standard says the result of this is undefined. The gcc
2755 compiler treats it as an empty string (which was what was wanted) but it is
2756 reported that Visual C gives an error. The source has been hacked around to
2757 avoid this problem.
2758
2759 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2760 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2761 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2762 of them did).
2763
2764 10. Originally, pcretest opened its input and output without "b"; then I was
2765 told that "b" was needed in some environments, so it was added for release
2766 5.0 to both the input and output. (It makes no difference on Unix-like
2767 systems.) Later I was told that it is wrong for the input on Windows. I've
2768 now abstracted the modes into two macros, to make it easier to fiddle with
2769 them, and removed "b" from the input mode under Windows.
2770
2771 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2772
2773 12. Added -help and --help to pcretest as an official way of being reminded
2774 of the options.
2775
2776 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2777 and pcrecpp.cc because they annoy compilers at high warning levels.
2778
2779 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2780
2781 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2782 scripts are not C programs :-) and which was not noticed because it works
2783 on Linux.
2784
2785 16. pcretest is supposed to handle any length of pattern and data line (as one
2786 line or as a continued sequence of lines) by extending its input buffer if
2787 necessary. This feature was broken for very long pattern lines, leading to
2788 a string of junk being passed to pcre_compile() if the pattern was longer
2789 than about 50K.
2790
2791 17. I have done a major re-factoring of the way pcre_compile() computes the
2792 amount of memory needed for a compiled pattern. Previously, there was code
2793 that made a preliminary scan of the pattern in order to do this. That was
2794 OK when PCRE was new, but as the facilities have expanded, it has become
2795 harder and harder to keep it in step with the real compile phase, and there
2796 have been a number of bugs (see for example, 4 above). I have now found a
2797 cunning way of running the real compile function in a "fake" mode that
2798 enables it to compute how much memory it would need, while actually only
2799 ever using a few hundred bytes of working memory and without too many
2800 tests of the mode. This should make future maintenance and development
2801 easier. A side effect of this work is that the limit of 200 on the nesting
2802 depth of parentheses has been removed (though this was never a serious
2803 limitation, I suspect). However, there is a downside: pcre_compile() now
2804 runs more slowly than before (30% or more, depending on the pattern). I
2805 hope this isn't a big issue. There is no effect on runtime performance.
2806
2807 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2808 newline (only possible for the last line of a file) and it was a
2809 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2810
2811 19. Added additional timing features to pcretest. (1) The -tm option now times
2812 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2813 separate command line item, by a number that specifies the number of
2814 repeats to use when timing. The default is 50000; this gives better
2815 precision, but takes uncomfortably long for very large patterns.
2816
2817 20. Extended pcre_study() to be more clever in cases where a branch of a
2818 subpattern has no definite first character. For example, (a*|b*)[cd] would
2819 previously give no result from pcre_study(). Now it recognizes that the
2820 first character must be a, b, c, or d.
2821
2822 21. There was an incorrect error "recursive call could loop indefinitely" if
2823 a subpattern (or the entire pattern) that was being tested for matching an
2824 empty string contained only one non-empty item after a nested subpattern.
2825 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2826 incorrectly, because the \d was being skipped in the check.
2827
2828 22. The pcretest program now has a new pattern option /B and a command line
2829 option -b, which is equivalent to adding /B to every pattern. This causes
2830 it to show the compiled bytecode, without the additional information that
2831 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2832 is the same as /B/I).
2833
2834 23. A new optimization is now able automatically to treat some sequences such
2835 as a*b as a*+b. More specifically, if something simple (such as a character
2836 or a simple class like \d) has an unlimited quantifier, and is followed by
2837 something that cannot possibly match the quantified thing, the quantifier
2838 is automatically "possessified".
2839
2840 24. A recursive reference to a subpattern whose number was greater than 39
2841 went wrong under certain circumstances in UTF-8 mode. This bug could also
2842 have affected the operation of pcre_study().
2843
2844 25. Realized that a little bit of performance could be had by replacing
2845 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2846
2847 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2848
2849 27. Possessive quantifiers such as a++ were previously implemented by turning
2850 them into atomic groups such as ($>a+). Now they have their own opcodes,
2851 which improves performance. This includes the automatically created ones
2852 from 23 above.
2853
2854 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2855 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2856 the first matched character to be a colon. This applied both to named and
2857 numbered groups.
2858
2859 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2860
2861 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2862 building PCRE on Windows easier, so I have included it in the distribution.
2863
2864 31. There is now a check in pcretest against a ridiculously large number being
2865 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2866 loop, the loop is abandoned.
2867
2868 32. Forward references to subpatterns in conditions such as (?(2)...) where
2869 subpattern 2 is defined later cause pcre_compile() to search forwards in
2870 the pattern for the relevant set of parentheses. This search went wrong
2871 when there were unescaped parentheses in a character class, parentheses
2872 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2873
2874 33. "Subroutine" calls and backreferences were previously restricted to
2875 referencing subpatterns earlier in the regex. This restriction has now
2876 been removed.
2877
2878 34. Added a number of extra features that are going to be in Perl 5.10. On the
2879 whole, these are just syntactic alternatives for features that PCRE had
2880 previously implemented using the Python syntax or my own invention. The
2881 other formats are all retained for compatibility.
2882
2883 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2884 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2885 also .NET compatible.
2886
2887 (b) A recursion or subroutine call to a named group can now be defined as
2888 (?&name) as well as (?P>name).
2889
2890 (c) A backreference to a named group can now be defined as \k<name> or
2891 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2892 5.10, are also .NET compatible.
2893
2894 (d) A conditional reference to a named group can now use the syntax
2895 (?(<name>) or (?('name') as well as (?(name).
2896
2897 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2898 groups (named and numbered) that are never evaluated inline, but can be
2899 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2900 is always false. There may be only one alternative in such a group.
2901
2902 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2903 as the simple (?(R). The condition is true only if the most recent
2904 recursion is that of the given number or name. It does not search out
2905 through the entire recursion stack.
2906
2907 (g) The escape \gN or \g{N} has been added, where N is a positive or
2908 negative number, specifying an absolute or relative reference.
2909
2910 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2911 some "unreachable code" warnings.
2912
2913 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2914 things, this adds five new scripts.
2915
2916 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2917 There were also incompatibilities regarding the handling of \Q..\E inside
2918 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2919 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2920
2921 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2922 matches an empty string, and forcibly breaks the loop. There were bugs in
2923 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2924 against aaaa the result was just "a" rather than "aaaa", for example. Two
2925 separate and independent bugs (that affected different cases) have been
2926 fixed.
2927
2928 39. Refactored the code to abolish the use of different opcodes for small
2929 capturing bracket numbers. This is a tidy that I avoided doing when I
2930 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2931 The new approach is not only tidier, it makes it possible to reduce the
2932 memory needed to fix the previous bug (38).
2933
2934 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2935 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2936 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2937 mode.
2938
2939 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2940 report.
2941
2942 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2943 copy construction and assignment in the C++ wrapper.
2944
2945 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2946 couple of bugs that could have given wrong results in the "--newline=crlf"
2947 case.
2948
2949 44. Added a number of casts and did some reorganization of signed/unsigned int
2950 variables following suggestions from Dair Grant. Also renamed the variable
2951 "this" as "item" because it is a C++ keyword.
2952
2953 45. Arranged for dftables to add
2954
2955 #include "pcre_internal.h"
2956
2957 to pcre_chartables.c because without it, gcc 4.x may remove the array
2958 definition from the final binary if PCRE is built into a static library and
2959 dead code stripping is activated.
2960
2961 46. For an unanchored pattern, if a match attempt fails at the start of a
2962 newline sequence, and the newline setting is CRLF or ANY, and the next two
2963 characters are CRLF, advance by two characters instead of one.
2964
2965
2966 Version 6.7 04-Jul-06
2967 ---------------------
2968
2969 1. In order to handle tests when input lines are enormously long, pcretest has
2970 been re-factored so that it automatically extends its buffers when
2971 necessary. The code is crude, but this _is_ just a test program. The
2972 default size has been increased from 32K to 50K.
2973
2974 2. The code in pcre_study() was using the value of the re argument before
2975 testing it for NULL. (Of course, in any sensible call of the function, it
2976 won't be NULL.)
2977
2978 3. The memmove() emulation function in pcre_internal.h, which is used on
2979 systems that lack both memmove() and bcopy() - that is, hardly ever -
2980 was missing a "static" storage class specifier.
2981
2982 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2983 containing an extended class (one that cannot be represented by a bitmap
2984 because it contains high-valued characters or Unicode property items, e.g.
2985 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2986 pattern, but PCRE should not loop if you do not (it no longer does).
2987 [Detail: two cases were found: (a) a repeated subpattern containing an
2988 extended class; (b) a recursive reference to a subpattern that followed a
2989 previous extended class. It wasn't skipping over the extended class
2990 correctly when UTF-8 mode was not set.]
2991
2992 5. A negated single-character class was not being recognized as fixed-length
2993 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2994 compile error "lookbehind assertion is not fixed length".
2995
2996 6. The RunPerlTest auxiliary script was showing an unexpected difference
2997 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2998 write a Perl script that can interpret lines of an input file either as
2999 byte characters or as UTF-8, which is what "perltest" was being required to
3000 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3001 can't do is switch easily at run time between having the "use utf8;" pragma
3002 or not. In the end, I fudged it by using the RunPerlTest script to insert
3003 "use utf8;" explicitly for the UTF-8 tests.
3004
3005 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3006 the end of the subject string, contrary to the documentation and to what
3007 Perl does. This was true of both matching functions. Now it matches only at
3008 the start of the subject and immediately after *internal* newlines.
3009
3010 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3011 a pointer to an int instead of a pointer to an unsigned long int. This
3012 caused problems on 64-bit systems.
3013
3014 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3015 instance of the 'standard' template library not being so standard".
3016
3017 10. There was no check on the number of named subpatterns nor the maximum
3018 length of a subpattern name. The product of these values is used to compute
3019 the size of the memory block for a compiled pattern. By supplying a very
3020 long subpattern name and a large number of named subpatterns, the size
3021 computation could be caused to overflow. This is now prevented by limiting
3022 the length of names to 32 characters, and the number of named subpatterns
3023 to 10,000.
3024
3025 11. Subpatterns that are repeated with specific counts have to be replicated in
3026 the compiled pattern. The size of memory for this was computed from the
3027 length of the subpattern and the repeat count. The latter is limited to
3028 65535, but there was no limit on the former, meaning that integer overflow
3029 could in principle occur. The compiled length of a repeated subpattern is
3030 now limited to 30,000 bytes in order to prevent this.
3031
3032 12. Added the optional facility to have named substrings with the same name.
3033
3034 13. Added the ability to use a named substring as a condition, using the
3035 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3036 are numbers (not recommended). Forward references are permitted.
3037
3038 14. Added forward references in named backreferences (if you see what I mean).
3039
3040 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3041 pattern could run off the end of the subject. For example, the pattern
3042 "(?s)(.{1,5})"8 did this with the subject "ab".
3043
3044 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3045 PCRE_CASELESS was set when matching characters that were quantified with ?
3046 or *.
3047
3048 17. A character class other than a single negated character that had a minimum
3049 but no maximum quantifier - for example [ab]{6,} - was not handled
3050 correctly by pce_dfa_exec(). It would match only one character.
3051
3052 18. A valid (though odd) pattern that looked like a POSIX character
3053 class but used an invalid character after [ (for example [[,abc,]]) caused
3054 pcre_compile() to give the error "Failed: internal error: code overflow" or
3055 in some cases to crash with a glibc free() error. This could even happen if
3056 the pattern terminated after [[ but there just happened to be a sequence of
3057 letters, a binary zero, and a closing ] in the memory that followed.
3058
3059 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3060 over the years. Originally (before any Unicode support), just the bottom 8
3061 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3062 output from "man perlunicode" includes this:
3063
3064 The regular expression compiler produces polymorphic opcodes. That
3065 is, the pattern adapts to the data and automatically switches to
3066 the Unicode character scheme when presented with Unicode data--or
3067 instead uses a traditional byte scheme when presented with byte
3068 data.
3069
3070 Sadly, a wide octal escape does not cause a switch, and in a string with
3071 no other multibyte characters, these octal escapes are treated as before.
3072 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3073 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3074 Unicode string.
3075
3076 I have not perpetrated such confusion in PCRE. Up till now, it took just
3077 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3078 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3079 translate to the appropriate multibyte character.
3080
3081 29. Applied some refactoring to reduce the number of warnings from Microsoft
3082 and Borland compilers. This has included removing the fudge introduced
3083 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3084 a warning about an unused variable.
3085
3086 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3087 characters since release 4.0, because Perl (from release 5.004) does not.
3088 [Or at least, is documented not to: some releases seem to be in conflict
3089 with the documentation.] However, when a pattern was studied with
3090 pcre_study() and all its branches started with \s, PCRE still included VT
3091 as a possible starting character. Of course, this did no harm; it just
3092 caused an unnecessary match attempt.
3093
3094 22. Removed a now-redundant internal flag bit that recorded the fact that case
3095 dependency changed within the pattern. This was once needed for "required
3096 byte" processing, but is no longer used. This recovers a now-scarce options
3097 bit. Also moved the least significant internal flag bit to the most-
3098 significant bit of the word, which was not previously used (hangover from
3099 the days when it was an int rather than a uint) to free up another bit for
3100 the future.
3101
3102 23. Added support for CRLF line endings as well as CR and LF. As well as the
3103 default being selectable at build time, it can now be changed at runtime
3104 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3105 specify that it is scanning data with non-default line endings.
3106
3107 24. Changed the definition of CXXLINK to make it agree with the definition of
3108 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3109
3110 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3111 recursions. This makes a big different to stack usage for some patterns.
3112
3113 26. If a subpattern containing a named recursion or subroutine reference such
3114 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3115 the space required for the compiled pattern went wrong and gave too small a
3116 value. Depending on the environment, this could lead to "Failed: internal
3117 error: code overflow at offset 49" or "glibc detected double free or
3118 corruption" errors.
3119
3120 27. Applied patches from Google (a) to support the new newline modes and (b) to
3121 advance over multibyte UTF-8 characters in GlobalReplace.
3122
3123 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3124 difference for some implementation of PCRE in some Windows version.
3125
3126 29. Added some extra testing facilities to pcretest:
3127
3128 \q<number> in a data line sets the "match limit" value
3129 \Q<number> in a data line sets the "match recursion limt" value
3130 -S <number> sets the stack size, where <number> is in megabytes
3131
3132 The -S option isn't available for Windows.
3133
3134
3135 Version 6.6 06-Feb-06
3136 ---------------------
3137
3138 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3139 in pcreposix.h. I have copied the definition from pcre.h.
3140
3141 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3142 because pcre.h is no longer a built file.
3143
3144 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3145 not normally included in the compiled code.
3146
3147
3148 Version 6.5 01-Feb-06
3149 ---------------------
3150
3151 1. When using the partial match feature with pcre_dfa_exec(), it was not
3152 anchoring the second and subsequent partial matches at the new starting
3153 point. This could lead to incorrect results. For example, with the pattern
3154 /1234/, partially matching against "123" and then "a4" gave a match.
3155
3156 2. Changes to pcregrep:
3157
3158 (a) All non-match returns from pcre_exec() were being treated as failures
3159 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3160 error message is output. Some extra information is given for the
3161 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3162 probably the only errors that are likely to be caused by users (by
3163 specifying a regex that has nested indefinite repeats, for instance).
3164 If there are more than 20 of these errors, pcregrep is abandoned.
3165
3166 (b) A binary zero was treated as data while matching, but terminated the
3167 output line if it was written out. This has been fixed: binary zeroes
3168 are now no different to any other data bytes.
3169
3170 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3171 used to set a locale for matching. The --locale=xxxx long option has
3172 been added (no short equivalent) to specify a locale explicitly on the
3173 pcregrep command, overriding the environment variables.
3174
3175 (d) When -B was used with -n, some line numbers in the output were one less
3176 than they should have been.
3177
3178 (e) Added the -o (--only-matching) option.
3179
3180 (f) If -A or -C was used with -c (count only), some lines of context were
3181 accidentally printed for the final match.
3182
3183 (g) Added the -H (--with-filename) option.
3184
3185 (h) The combination of options -rh failed to suppress file names for files
3186 that were found from directory arguments.
3187
3188 (i) Added the -D (--devices) and -d (--directories) options.
3189
3190 (j) Added the -F (--fixed-strings) option.
3191
3192 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3193
3194 (l) Added the --colo(u)r option.
3195
3196 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3197 is not present by default.
3198
3199 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3200 items such as (?R) or (?1), when the recursion could match a number of
3201 alternatives. If it matched one of the alternatives, but subsequently,
3202 outside the recursion, there was a failure, the code tried to back up into
3203 the recursion. However, because of the way PCRE is implemented, this is not
3204 possible, and the result was an incorrect result from the match.
3205
3206 In order to prevent this happening, the specification of recursion has
3207 been changed so that all such subpatterns are automatically treated as
3208 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3209
3210 4. I had overlooked the fact that, in some locales, there are characters for
3211 which isalpha() is true but neither isupper() nor islower() are true. In
3212 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3213 and ordfeminine) are like this. This affected the treatment of \w and \W
3214 when they appeared in character classes, but not when they appeared outside
3215 a character class. The bit map for "word" characters is now created
3216 separately from the results of isalnum() instead of just taking it from the
3217 upper, lower, and digit maps. (Plus the underscore character, of course.)
3218
3219 5. The above bug also affected the handling of POSIX character classes such as
3220 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3221 permanent tables. Instead, the bit maps for such a class were previously
3222 created as the appropriate unions of the upper, lower, and digit bitmaps.
3223 Now they are created by subtraction from the [[:word:]] class, which has
3224 its own bitmap.
3225
3226 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3227 It is created by subtracting the vertical space characters (\x09, \x0a,
3228 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3229 subtraction was done in the overall bitmap for a character class, meaning
3230 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3231 be recognized. This bug has been fixed.
3232
3233 7. Patches from the folks at Google:
3234
3235 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3236 real life, but is still worth protecting against".
3237
3238 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3239 regular expressions".
3240
3241 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3242 have it.
3243
3244 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3245 "configure" and the latter not, in order to fix a problem somebody had
3246 with compiling the Arg class on HP-UX.
3247
3248 (e) Improve the error-handling of the C++ wrapper a little bit.
3249
3250 (f) New tests for checking recursion limiting.
3251
3252 8. The pcre_memmove() function, which is used only if the environment does not
3253 have a standard memmove() function (and is therefore rarely compiled),
3254 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3255 returning a result (though PCRE never actually uses the result).
3256
3257 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3258 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3259 returned instead of calling malloc() with an overflowing number that would
3260 most likely cause subsequent chaos.
3261
3262 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3263
3264 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3265 with this option is matched, the nmatch and pmatch options of regexec() are
3266 ignored.
3267
3268 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3269 provided in case anyone wants to the the POSIX interface with UTF-8
3270 strings.
3271
3272 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3273 C++ linking (needed for some HP-UX environments).
3274
3275 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3276 (unused parameter) and in the pcre_printint() function (omitted "default"
3277 switch label when the default is to do nothing).
3278
3279 15. Added some code to make it possible, when PCRE is compiled as a C++
3280 library, to replace subject pointers for pcre_exec() with a smart pointer
3281 class, thus making it possible to process discontinuous strings.
3282
3283 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3284 much the same function. They were added by different people who were trying
3285 to make PCRE easy to compile on non-Unix systems. It has been suggested
3286 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3287 for compiling on Windows systems. I have therefore replaced it with
3288 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3289 defaults to "extern" for C or "extern C" for C++, which works fine on
3290 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3291 SCOPE with something explicit in config.h. In addition:
3292
3293 (a) pcreposix.h still had just "extern" instead of either of these macros;
3294 I have replaced it with PCRE_DATA_SCOPE.
3295
3296 (b) Functions such as _pcre_xclass(), which are internal to the library,
3297 but external in the C sense, all had PCRE_EXPORT in their definitions.
3298 This is apparently wrong for the Windows case, so I have removed it.
3299 (It makes no difference on Unix-like systems.)
3300
3301 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3302 of recursive calls to match(). This is different to MATCH_LIMIT because
3303 that limits the total number of calls to match(), not all of which increase
3304 the depth of recursion. Limiting the recursion depth limits the amount of
3305 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3306 when PCRE is compiled, and changed at run time. A patch from Google adds
3307 this functionality to the C++ interface.
3308
3309 18. Changes to the handling of Unicode character properties:
3310
3311 (a) Updated the table to Unicode 4.1.0.
3312
3313 (b) Recognize characters that are not in the table as "Cn" (undefined).
3314
3315 (c) I revised the way the table is implemented to a much improved format
3316 which includes recognition of ranges. It now supports the ranges that
3317 are defined in UnicodeData.txt, and it also amalgamates other
3318 characters into ranges. This has reduced the number of entries in the
3319 table from around 16,000 to around 3,000, thus reducing its size
3320 considerably. I realized I did not need to use a tree structure after
3321 all - a binary chop search is just as efficient. Having reduced the
3322 number of entries, I extended their size from 6 bytes to 8 bytes to
3323 allow for more data.
3324
3325 (d) Added support for Unicode script names via properties such as \p{Han}.
3326
3327 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3328 matching that character.
3329
3330 20. When matching a repeated Unicode property with a minimum greater than zero,
3331 (for example \pL{2,}), PCRE could look past the end of the subject if it
3332 reached it while seeking the minimum number of characters. This could
3333 happen only if some of the characters were more than one byte long, because
3334 there is a check for at least the minimum number of bytes.
3335
3336 21. Refactored the implementation of \p and \P so as to be more general, to
3337 allow for more different types of property in future. This has changed the
3338 compiled form incompatibly. Anybody with saved compiled patterns that use
3339 \p or \P will have to recompile them.
3340
3341 22. Added "Any" and "L&" to the supported property types.
3342
3343 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3344 but give a compile time error if the value is greater than 0xff.
3345
3346 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3347 accidentally not being installed or uninstalled.
3348
3349 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3350 made were to insert the current release number. This seemed silly, because
3351 it made things harder for people building PCRE on systems that don't run
3352 "configure". I have turned pcre.h into a distributed file, no longer built
3353 by "configure", with the version identification directly included. There is
3354 no longer a pcre.h.in file.
3355
3356 However, this change necessitated a change to the pcre-config script as
3357 well. It is built from pcre-config.in, and one of the substitutions was the
3358 release number. I have updated configure.ac so that ./configure now finds
3359 the release number by grepping pcre.h.
3360
3361 26. Added the ability to run the tests under valgrind.
3362
3363
3364 Version 6.4 05-Sep-05
3365 ---------------------
3366
3367 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3368 "--" to be printed when multiple files were scanned, even when none of the
3369 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3370 consider it to be a bug, and have restored the previous behaviour.
3371
3372 2. A couple of code tidies to get rid of compiler warnings.
3373
3374 3. The pcretest program used to cheat by referring to symbols in the library
3375 whose names begin with _pcre_. These are internal symbols that are not
3376 really supposed to be visible externally, and in some environments it is
3377 possible to suppress them. The cheating is now confined to including
3378 certain files from the library's source, which is a bit cleaner.
3379
3380 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3381 file's purpose clearer.
3382
3383 5. Reorganized pcre_ucp_findchar().
3384
3385
3386 Version 6.3 15-Aug-05
3387 ---------------------
3388
3389 1. The file libpcre.pc.in did not have general read permission in the tarball.
3390
3391 2. There were some problems when building without C++ support:
3392
3393 (a) If C++ support was not built, "make install" and "make test" still
3394 tried to test it.
3395
3396 (b) There were problems when the value of CXX was explicitly set. Some
3397 changes have been made to try to fix these, and ...
3398
3399 (c) --disable-cpp can now be used to explicitly disable C++ support.
3400
3401 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3402 backslash in a target when C++ was disabled. This confuses some
3403 versions of "make", apparently. Using an intermediate variable solves
3404 this. (Same for CPP_LOBJ.)
3405
3406 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3407 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3408 necessary on certain architectures.
3409
3410 4. Added a setting of -export-symbols-regex to the link command to remove
3411 those symbols that are exported in the C sense, but actually are local
3412 within the library, and not documented. Their names all begin with
3413 "_pcre_". This is not a perfect job, because (a) we have to except some
3414 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3415 available (and never for static libraries). I have made a note to try to
3416 find a way round (a) in the future.
3417
3418
3419 Version 6.2 01-Aug-05
3420 ---------------------
3421
3422 1. There was no test for integer overflow of quantifier values. A construction
3423 such as {1111111111111111} would give undefined results. What is worse, if
3424 a minimum quantifier for a parenthesized subpattern overflowed and became
3425 negative, the calculation of the memory size went wrong. This could have
3426 led to memory overwriting.
3427
3428 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3429
3430 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3431 operating environments where this matters.
3432
3433 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3434 PCRE options from within the C++ wrapper.
3435
3436 5. Named capturing subpatterns were not being correctly counted when a pattern
3437 was compiled. This caused two problems: (a) If there were more than 100
3438 such subpatterns, the calculation of the memory needed for the whole
3439 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3440 back references of the form \12, where the number was greater than 9, were
3441 not recognized as back references, even though there were sufficient
3442 previous subpatterns.
3443
3444 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3445 versions of gcc, e.g. 2.95.4.
3446
3447
3448 Version 6.1 21-Jun-05
3449 ---------------------
3450
3451 1. There was one reference to the variable "posix" in pcretest.c that was not
3452 surrounded by "#if !defined NOPOSIX".
3453
3454 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3455 the cross-check on the old pcre_info() function, for the benefit of the
3456 cut-down version of PCRE that is currently imported into Exim.
3457
3458 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3459 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3460 patterns that start (?i) but I don't think that matters. The use of (?i) is
3461 just an example; this all applies to the other options as well.
3462
3463 4. Since libtool seems to echo the compile commands it is issuing, the output
3464 from "make" can be reduced a bit by putting "@" in front of each libtool
3465 compile command.
3466
3467 5. Patch from the folks at Google for configure.in to be a bit more thorough
3468 in checking for a suitable C++ installation before trying to compile the
3469 C++ stuff. This should fix a reported problem when a compiler was present,
3470 but no suitable headers.
3471
3472 6. The man pages all had just "PCRE" as their title. I have changed them to
3473 be the relevant file name. I have also arranged that these names are
3474 retained in the file doc/pcre.txt, which is a concatenation in text format
3475 of all the man pages except the little individual ones for each function.
3476
3477 7. The NON-UNIX-USE file had not been updated for the different set of source
3478 files that come with release 6. I also added a few comments about the C++
3479 wrapper.
3480
3481
3482 Version 6.0 07-Jun-05
3483 ---------------------
3484
3485 1. Some minor internal re-organization to help with my DFA experiments.
3486
3487 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3488 didn't matter for the library itself when fully configured, but did matter
3489 when compiling without UCP support, or within Exim, where the ucp files are
3490 not imported.
3491
3492 3. Refactoring of the library code to split up the various functions into
3493 different source modules. The addition of the new DFA matching code (see
3494 below) to a single monolithic source would have made it really too
3495 unwieldy, quite apart from causing all the code to be include in a
3496 statically linked application, when only some functions are used. This is
3497 relevant even without the DFA addition now that patterns can be compiled in
3498 one application and matched in another.
3499
3500 The downside of splitting up is that there have to be some external
3501 functions and data tables that are used internally in different modules of
3502 the library but which are not part of the API. These have all had their
3503 names changed to start with "_pcre_" so that they are unlikely to clash
3504 with other external names.
3505
3506 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3507 a different (DFA) algorithm. Although it is slower than the original
3508 function, it does have some advantages for certain types of matching
3509 problem.
3510
3511 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3512 including restarting after a partial match.
3513
3514 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3515 defined when compiling for Windows was sent to me. I have put it into the
3516 code, though I have no means of testing or verifying it.
3517
3518 7. Added the pcre_refcount() auxiliary function.
3519
3520 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3521 match before or at the first newline in the subject string. In pcretest,
3522 the /f option on a pattern can be used to set this.
3523
3524 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3525 would behave wrongly. This has been present in PCRE since release 4.0.
3526
3527 10. A number of changes to the pcregrep command:
3528
3529 (a) Refactored how -x works; insert ^(...)$ instead of setting
3530 PCRE_ANCHORED and checking the length, in preparation for adding
3531 something similar for -w.
3532
3533 (b) Added the -w (match as a word) option.
3534
3535 (c) Refactored the way lines are read and buffered so as to have more
3536 than one at a time available.
3537
3538 (d) Implemented a pcregrep test script.
3539
3540 (e) Added the -M (multiline match) option. This allows patterns to match
3541 over several lines of the subject. The buffering ensures that at least
3542 8K, or the rest of the document (whichever is the shorter) is available
3543 for matching (and similarly the previous 8K for lookbehind assertions).
3544
3545 (f) Changed the --help output so that it now says
3546
3547 -w, --word-regex(p)
3548
3549 instead of two lines, one with "regex" and the other with "regexp"
3550 because that confused at least one person since the short forms are the
3551 same. (This required a bit of code, as the output is generated
3552 automatically from a table. It wasn't just a text change.)
3553
3554 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3555 option but starts with a hyphen. Could be a pattern or a path name
3556 starting with a hyphen, for instance.
3557
3558 (h) "-" can be given as a file name to represent stdin.
3559
3560 (i) When file names are being printed, "(standard input)" is used for
3561 the standard input, for compatibility with GNU grep. Previously
3562 "<stdin>" was used.
3563
3564 (j) The option --label=xxx can be used to supply a name to be used for
3565 stdin when file names are being printed. There is no short form.
3566
3567 (k) Re-factored the options decoding logic because we are going to add
3568 two more options that take data. Such options can now be given in four
3569 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3570
3571 (l) Added the -A, -B, and -C options for requesting that lines of context
3572 around matches be printed.
3573
3574 (m) Added the -L option to print the names of files that do not contain
3575 any matching lines, that is, the complement of -l.
3576
3577 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3578 continue to scan other files.
3579
3580 (o) The -s option was incorrectly implemented. For compatibility with other
3581 greps, it now suppresses the error message for a non-existent or non-
3582 accessible file (but not the return code). There is a new option called
3583 -q that suppresses the output of matching lines, which was what -s was
3584 previously doing.
3585
3586 (p) Added --include and --exclude options to specify files for inclusion
3587 and exclusion when recursing.
3588
3589 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3590 Hopefully, it now does.
3591
3592 12. Missing cast in pcre_study().
3593
3594 13. Added an "uninstall" target to the makefile.
3595
3596 14. Replaced "extern" in the function prototypes in Makefile.in with
3597 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3598 world, but is set differently for Windows.
3599
3600 15. Added a second compiling function called pcre_compile2(). The only
3601 difference is that it has an extra argument, which is a pointer to an
3602 integer error code. When there is a compile-time failure, this is set
3603 non-zero, in addition to the error test pointer being set to point to an
3604 error message. The new argument may be NULL if no error number is required
3605 (but then you may as well call pcre_compile(), which is now just a
3606 wrapper). This facility is provided because some applications need a
3607 numeric error indication, but it has also enabled me to tidy up the way
3608 compile-time errors are handled in the POSIX wrapper.
3609
3610 16. Added VPATH=.libs to the makefile; this should help when building with one
3611 prefix path and installing with another. (Or so I'm told by someone who
3612 knows more about this stuff than I do.)
3613
3614 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3615 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3616 match everything, including newlines. This is not POSIX-compatible, but
3617 somebody wanted the feature. From pcretest it can be activated by using
3618 both the P and the s flags.
3619
3620 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3621
3622 19. libpcre.pc was being incorrectly installed as executable.
3623
3624 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3625 it now also looks for '\r' so that it will work unmodified on Windows.
3626
3627 21. Added Google's contributed C++ wrapper to the distribution.
3628
3629 22. Added some untidy missing memory free() calls in pcretest, to keep
3630 Electric Fence happy when testing.
3631
3632
3633
3634 Version 5.0 13-Sep-04
3635 ---------------------
3636
3637 1. Internal change: literal characters are no longer packed up into items
3638 containing multiple characters in a single byte-string. Each character
3639 is now matched using a separate opcode. However, there may be more than one
3640 byte in the character in UTF-8 mode.
3641
3642 2. The pcre_callout_block structure has two new fields: pattern_position and
3643 next_item_length. These contain the offset in the pattern to the next match
3644 item, and its length, respectively.
3645
3646 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3647 insertion of callouts before each pattern item. Added the /C option to
3648 pcretest to make use of this.
3649
3650 4. On the advice of a Windows user, the lines
3651
3652 #if defined(_WIN32) || defined(WIN32)
3653 _setmode( _fileno( stdout ), 0x8000 );
3654 #endif /* defined(_WIN32) || defined(WIN32) */
3655
3656 have been added to the source of pcretest. This apparently does useful
3657 magic in relation to line terminators.
3658
3659 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3660 for the benefit of those environments where the "b" makes a difference.
3661
3662 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3663 to know about it. I have put a hack into configure.in that adds in code
3664 to set GCC=yes if CC=icc. This seems to end up at a point in the
3665 generated configure script that is early enough to affect the setting of
3666 compiler options, which is what is needed, but I have no means of testing
3667 whether it really works. (The user who reported this had patched the
3668 generated configure script, which of course I cannot do.)
3669
3670 LATER: After change 22 below (new libtool files), the configure script
3671 seems to know about icc (and also ecc). Therefore, I have commented out
3672 this hack in configure.in.
3673
3674 7. Added support for pkg-config (2 patches were sent in).
3675
3676 8. Negated POSIX character classes that used a combination of internal tables
3677 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3678 [[:^ascii]]. Typically, they would match almost any characters. The other
3679 POSIX classes were not broken in this way.
3680
3681 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3682 to find the match, as PCRE was deluded into thinking that the match had to
3683 start at the start point or following a newline. The same bug applied to
3684 patterns with negative forward assertions or any backward assertions
3685 preceding ".*" at the start, unless the pattern required a fixed first
3686 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3687
3688 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3689 starting at the last subject character, bytes beyond the end of the subject
3690 string were read.
3691
3692 11. Renamed the variable "class" as "classbits" to make life easier for C++
3693 users. (Previously there was a macro definition, but it apparently wasn't
3694 enough.)
3695
3696 12. Added the new field "tables" to the extra data so that tables can be passed
3697 in at exec time, or the internal tables can be re-selected. This allows
3698 a compiled regex to be saved and re-used at a later time by a different
3699 program that might have everything at different addresses.
3700
3701 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3702 -R library as well as a -L library.
3703
3704 14. The debugging options of pcretest (-d on the command line or D on a
3705 pattern) showed incorrect output for anything following an extended class
3706 that contained multibyte characters and which was followed by a quantifier.
3707
3708 15. Added optional support for general category Unicode character properties
3709 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3710 support. It adds about 90K to the size of the library. The meanings of the
3711 inbuilt class escapes such as \d and \s have NOT been changed.
3712
3713 16. Updated pcredemo.c to include calls to free() to release the memory for the
3714 compiled pattern.
3715
3716 17. The generated file chartables.c was being created in the source directory
3717 instead of in the building directory. This caused the build to fail if the
3718 source directory was different from the building directory, and was
3719 read-only.
3720
3721 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3722 file. No doubt somebody will tell me if they don't make sense... Also added
3723 Dan Mooney's comments about building on OpenVMS.
3724
3725 19. Added support for partial matching via the PCRE_PARTIAL option for
3726 pcre_exec() and the \P data escape in pcretest.
3727
3728 20. Extended pcretest with 3 new pattern features:
3729
3730 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3731 write the compiled pattern to the file whose name is "rest-of-line".
3732 This is a straight binary dump of the data, with the saved pointer to
3733 the character tables forced to be NULL. The study data, if any, is
3734 written too. After writing, pcretest reads a new pattern.
3735
3736 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3737 compiled pattern from the given file. There must not be any
3738 occurrences of "<" in the file name (pretty unlikely); if there are,
3739 pcretest will instead treat the initial "<" as a pattern delimiter.
3740 After reading in the pattern, pcretest goes on to read data lines as
3741 usual.
3742
3743 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3744 and 16-bit fields in a compiled pattern, to simulate a pattern that
3745 was compiled on a host of opposite endianness.
3746
3747 21. The pcre-exec() function can now cope with patterns that were compiled on
3748 hosts of opposite endianness, with this restriction:
3749
3750 As for any compiled expression that is saved and used later, the tables
3751 pointer field cannot be preserved; the extra_data field in the arguments
3752 to pcre_exec() should be used to pass in a tables address if a value
3753 other than the default internal tables were used at compile time.
3754
3755 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3756 now diagnosed as an error. Previously, most of the time, a negative number
3757 would have been treated as zero, but if in addition "ovector" was passed as
3758 NULL, a crash could occur.
3759
3760 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3761 new versions from the libtool 1.5 distribution (the last one is a copy of
3762 a file called libtool.m4). This seems to have fixed the need to patch
3763 "configure" to support Darwin 1.3 (which I used to do). However, I still
3764 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3765 workstation).
3766
3767 24. Changed the PCRE licence to be the more standard "BSD" licence.
3768
3769
3770 Version 4.5 01-Dec-03
3771 ---------------------
3772
3773 1. There has been some re-arrangement of the code for the match() function so
3774 that it can be compiled in a version that does not call itself recursively.
3775 Instead, it keeps those local variables that need separate instances for
3776 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3777 needs to "recurse". Keeping track of where control must go is done by means
3778 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3779 hide most of the details from the main code, and operates only if
3780 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3781 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3782 operating.
3783
3784 To make it easier for callers to provide specially tailored get/free
3785 functions for this usage, two new functions, pcre_stack_malloc, and
3786 pcre_stack_free, are used. They are always called in strict stacking order,
3787 and the size of block requested is always the same.
3788
3789 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3790 PCRE has been compiled to use the stack or the heap for recursion. The
3791 -C option of pcretest uses this to show which version is compiled.
3792
3793 A new data escape \S, is added to pcretest; it causes the amounts of store
3794 obtained and freed by both kinds of malloc/free at match time to be added
3795 to the output.
3796
3797 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3798 what's available on my current Linux desktop machine.
3799
3800 3. When matching a UTF-8 string, the test for a valid string at the start has
3801 been extended. If start_offset is not zero, PCRE now checks that it points
3802 to a byte that is the start of a UTF-8 character. If not, it returns
3803 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3804 this is necessary because there may be backward assertions in the pattern.
3805 When matching the same subject several times, it may save resources to use
3806 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3807
3808 4. The code for checking the validity of UTF-8 strings has been tightened so
3809 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3810 containing "overlong sequences".
3811
3812 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3813 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3814 should have been "&", but it just so happened that all the cases this let
3815 through by mistake were picked up later in the function.
3816
3817 6. I had used a variable called "isblank" - this is a C99 function, causing
3818 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3819
3820 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3821 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3822 so the version is shown; (c) stop "make test" echoing "./RunTest".
3823
3824 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3825
3826 9. The prototype for memmove() for systems that don't have it was using
3827 size_t, but the inclusion of the header that defines size_t was later. I've
3828 moved the #includes for the C headers earlier to avoid this.
3829
3830 10. Added some adjustments to the code to make it easier to compiler on certain
3831 special systems:
3832
3833 (a) Some "const" qualifiers were missing.
3834 (b) Added the macro EXPORT before all exported functions; by default this
3835 is defined to be empty.
3836 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3837 that it reads its output file name as an argument instead of writing
3838 to the standard output and assuming this can be redirected.
3839
3840 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3841 class containing characters with values greater than 255, PCRE compilation
3842 went into a loop.
3843
3844 12. A recursive reference to a subpattern that was within another subpattern
3845 that had a minimum quantifier of zero caused PCRE to crash. For example,
3846 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3847 recursion. If the recursively-called subpattern itself had a zero repeat,
3848 that was OK.
3849
3850 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3851 buffer into which it was copied (for escape processing) was still set at
3852 1024, so long lines caused crashes.
3853
3854 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3855 "internal error: code overflow...". This applied to any character class
3856 that was followed by a possessive quantifier.
3857
3858 15. Modified the Makefile to add libpcre.la as a prerequisite for
3859 libpcreposix.la because I was told this is needed for a parallel build to
3860 work.
3861
3862 16. If a pattern that contained .* following optional items at the start was
3863 studied, the wrong optimizing data was generated, leading to matching
3864 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3865 matching string must start with a or b or c. The correct conclusion for
3866 this pattern is that a match can start with any character.
3867
3868
3869 Version 4.4 13-Aug-03
3870 ---------------------
3871
3872 1. In UTF-8 mode, a character class containing characters with values between
3873 127 and 255 was not handled correctly if the compiled pattern was studied.
3874 In fixing this, I have also improved the studying algorithm for such
3875 classes (slightly).
3876
3877 2. Three internal functions had redundant arguments passed to them. Removal
3878 might give a very teeny performance improvement.
3879
3880 3. Documentation bug: the value of the capture_top field in a callout is *one
3881 more than* the number of the hightest numbered captured substring.
3882
3883 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3884 in incorrectly linking with a previously installed version. They now link
3885 explicitly with libpcre.la.
3886
3887 5. configure.in no longer needs to recognize Cygwin specially.
3888
3889 6. A problem in pcre.in for Windows platforms is fixed.
3890
3891 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3892 pcretest, it used to include the size of the study block as part of its
3893 output. Unfortunately, the structure contains a field that has a different
3894 size on different hardware architectures. This meant that the tests that
3895 showed this size failed. As the block is currently always of a fixed size,
3896 this information isn't actually particularly useful in pcretest output, so
3897 I have just removed it.
3898
3899 8. Three pre-processor statements accidentally did not start in column 1.
3900 Sadly, there are *still* compilers around that complain, even though
3901 standard C has not required this for well over a decade. Sigh.
3902
3903 9. In pcretest, the code for checking callouts passed small integers in the
3904 callout_data field, which is a void * field. However, some picky compilers
3905 complained about the casts involved for this on 64-bit systems. Now
3906 pcretest passes the address of the small integer instead, which should get
3907 rid of the warnings.
3908
3909 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3910 both compile and run time, and gives an error if an invalid UTF-8 sequence
3911 is found. There is a option for disabling this check in cases where the
3912 string is known to be correct and/or the maximum performance is wanted.
3913
3914 11. In response to a bug report, I changed one line in Makefile.in from
3915
3916 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3917 to
3918 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3919
3920 to look similar to other lines, but I have no way of telling whether this
3921 is the right thing to do, as I do not use Windows. No doubt I'll get told
3922 if it's wrong...
3923
3924
3925 Version 4.3 21-May-03
3926 ---------------------
3927
3928 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3929 Makefile.
3930
3931 2. Some refactoring to improve the quality of the code:
3932
3933 (i) The utf8_table... variables are now declared "const".
3934
3935 (ii) The code for \cx, which used the "case flipping" table to upper case
3936 lower case letters, now just substracts 32. This is ASCII-specific,
3937 but the whole concept of \cx is ASCII-specific, so it seems
3938 reasonable.
3939
3940 (iii) PCRE was using its character types table to recognize decimal and
3941 hexadecimal digits in the pattern. This is silly, because it handles
3942 only 0-9, a-f, and A-F, but the character types table is locale-
3943 specific, which means strange things might happen. A private
3944 table is now used for this - though it costs 256 bytes, a table is
3945 much faster than multiple explicit tests. Of course, the standard
3946 character types table is still used for matching digits in subject
3947 strings against \d.
3948
3949 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3950 ending in _t are). So I've renamed it as ESC_tee.
3951
3952 3. The first argument for regexec() in the POSIX wrapper should have been
3953 defined as "const".
3954
3955 4. Changed pcretest to use malloc() for its buffers so that they can be
3956 Electric Fenced for debugging.
3957
3958 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3959 to read one or more bytes before the start of the subject string. Often this
3960 had no effect on PCRE's behaviour, but in some circumstances it could
3961 provoke a segmentation fault.
3962
3963 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3964 to try to read one or more bytes before the start of the subject string.
3965
3966 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3967 UTF-8 support could misbehave in various ways if the subject string
3968 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3969 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3970 back over UTF-8 characters.)
3971
3972
3973 Version 4.2 14-Apr-03
3974 ---------------------
3975
3976 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3977
3978 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3979 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3980 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3981 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3982 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3983 and BUILD_EXEEXT
3984 Note: automatic setting of the BUILD variables is not yet working
3985 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3986 compile-time but not at link-time
3987 [LINK]: use for linking executables only
3988 make different versions for Windows and non-Windows
3989 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3990 libraries
3991 [LINK_FOR_BUILD]: new variable
3992 [OBJEXT]: use throughout
3993 [EXEEXT]: use throughout
3994 <winshared>: new target
3995 <wininstall>: new target
3996 <dftables.o>: use native compiler
3997 <dftables>: use native linker
3998 <install>: handle Windows platform correctly
3999 <clean>: ditto
4000 <check>: ditto
4001 copy DLL to top builddir before testing
4002
4003 As part of these changes, -no-undefined was removed again. This was reported
4004 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4005 in any case.
4006
4007 3. Some tidies to get rid of compiler warnings:
4008
4009 . In the match_data structure, match_limit was an unsigned long int, whereas
4010 match_call_count was an int. I've made them both unsigned long ints.
4011
4012 . In pcretest the fact that a const uschar * doesn't automatically cast to
4013 a void * provoked a warning.
4014
4015 . Turning on some more compiler warnings threw up some "shadow" variables
4016 and a few more missing casts.
4017
4018 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4019 option, a class that contained a single character with a value between 128
4020 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4021
4022 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4023 option, a class that contained several characters, but with at least one
4024 whose value was between 128 and 255 caused PCRE to crash.
4025
4026
4027 Version 4.1 12-Mar-03
4028 ---------------------
4029
4030 1. Compiling with gcc -pedantic found a couple of places where casts were
4031 needed, and a string in dftables.c that was longer than standard compilers are
4032 required to support.
4033
4034 2. Compiling with Sun's compiler found a few more places where the code could
4035 be tidied up in order to avoid warnings.
4036
4037 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4038 first of these names is deprecated in the latest Autoconf in favour of the name
4039 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4040 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4041 analogy I have changed it to CFLAGS_FOR_BUILD.
4042
4043 4. Added -no-undefined to the linking command in the Makefile, because this is
4044 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4045 linking step for the pcreposix library.
4046
4047 5. PCRE was failing to diagnose the case of two named groups with the same
4048 name.
4049
4050 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4051 literal character that is needed in the subject for a match, and scans along to
4052 ensure that it is present before embarking on the full matching process. This
4053 saves time in cases of nested unlimited repeats that are never going to match.
4054 Problem: the scan can take a lot of time if the subject is very long (e.g.
4055 megabytes), thus penalizing straightforward matches. It is now done only if the
4056 amount of subject to be scanned is less than 1000 bytes.
4057
4058 7. A lesser problem with the same optimization is that it was recording the
4059 first character of an anchored pattern as "needed", thus provoking a search
4060 right along the subject, even when the first match of the pattern was going to
4061 fail. The "needed" character is now not set for anchored patterns, unless it
4062 follows something in the pattern that is of non-fixed length. Thus, it still
4063 fulfils its original purpose of finding quick non-matches in cases of nested
4064 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4065
4066
4067 Version 4.0 17-Feb-03
4068 ---------------------
4069
4070 1. If a comment in an extended regex that started immediately after a meta-item
4071 extended to the end of string, PCRE compiled incorrect data. This could lead to
4072 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4073
4074 2. Moved to autoconf 2.53 and libtool 1.4.2.
4075
4076 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4077 the special perltest8 script is no longer needed - all the tests can be run
4078 from a single perltest script.
4079
4080 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4081 by \s. It has now been removed in PCRE. This means it isn't recognized as
4082 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4083 class [:space:] *does* include VT, thereby creating a mess.
4084
4085 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4086 space and tab.
4087
4088 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4089 its new features into the main test script, reducing the number of scripts.
4090
4091 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4092 were backward compatible, and made the (?i) apply to the whole pattern, as if
4093 /i were given. Now it behaves more logically, and applies the option setting
4094 only to what follows. PCRE has been changed to follow suit. However, if it
4095 finds options settings right at the start of the pattern, it extracts them into
4096 the global options, as before. Thus, they show up in the info data.
4097
4098 8. Added support for the \Q...\E escape sequence. Characters in between are
4099 treated as literals. This is slightly different from Perl in that $ and @ are
4100 also handled as literals inside the quotes. In Perl, they will cause variable
4101 interpolation. Note the following examples:
4102
4103 Pattern PCRE matches Perl matches
4104
4105 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4106 \Qabc\$xyz\E abc\$xyz abc\$xyz
4107 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4108
4109 For compatibility with Perl, \Q...\E sequences are recognized inside character
4110 classes as well as outside them.
4111
4112 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4113 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4114 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4115 signed/unsigned warnings.
4116
4117 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4118 option for pcretest, so I've replaced it by a simple function that does just
4119 that job.
4120
4121 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4122 "pcregrep -".
4123
4124 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4125 Java package. This provides some syntactic sugar for simple cases of what my
4126 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4127 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4128 item, you can use this simplified notation. Note that only makes sense with
4129 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4130 greediness, whatever the setting of the PCRE_UNGREEDY option.
4131
4132 13. A change of greediness default within a pattern was not taking effect at
4133 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4134 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4135 was abstracted outside.
4136
4137 14. PCRE now supports the \G assertion. It is true when the current matching
4138 position is at the start point of the match. This differs from \A when the
4139 starting offset is non-zero. Used with the /g option of pcretest (or similar
4140 code), it works in the same way as it does for Perl's /g option. If all
4141 alternatives of a regex begin with \G, the expression is anchored to the start
4142 match position, and the "anchored" flag is set in the compiled expression.
4143
4144 15. Some bugs concerning the handling of certain option changes within patterns
4145 have been fixed. These applied to options other than (?ims). For example,
4146 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4147 been the other way round. Some of this was related to change 7 above.
4148
4149 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4150 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4151 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4152 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4153
4154 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4155 mode. Unlike ".", it always matches newline, whatever the setting of
4156 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4157 assertions. Perl allows it, but it doesn't (in general) work because it can't
4158 calculate the length of the lookbehind. At least, that's the case for Perl
4159 5.8.0 - I've been told they are going to document that it doesn't work in
4160 future.
4161
4162 18. Added an error diagnosis for escapes that PCRE does not support: these are
4163 \L, \l, \N, \P, \p, \U, \u, and \X.
4164
4165 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4166 reading past the end of the pattern in cases such as /[abcd/.
4167
4168 20. PCRE was getting more memory than necessary for patterns with classes that
4169 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4170
4171 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4172 compiling PCRE for use with Virtual Pascal.
4173
4174 22. Small fix to the Makefile to make it work properly if the build is done
4175 outside the source tree.
4176
4177 23. Added a new extension: a condition to go with recursion. If a conditional
4178 subpattern starts with (?(R) the "true" branch is used if recursion has
4179 happened, whereas the "false" branch is used only at the top level.
4180
4181 24. When there was a very long string of literal characters (over 255 bytes
4182 without UTF support, over 250 bytes with UTF support), the computation of how
4183 much memory was required could be incorrect, leading to segfaults or other
4184 strange effects.
4185
4186 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4187 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4188 there was a subsequent back reference to those brackets. This meant that, for
4189 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4190 possible to check for precisely this case. All we can do is abandon the
4191 optimization if .* occurs inside capturing brackets when there are any back
4192 references whatsoever. (See below for a better fix that came later.)
4193
4194 26. The handling of the optimization for finding the first character of a
4195 non-anchored pattern, and for finding a character that is required later in the
4196 match were failing in some cases. This didn't break the matching; it just
4197 failed to optimize when it could. The way this is done has been re-implemented.
4198
4199 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4200
4201 28. Added a new feature that provides some of the functionality that Perl
4202 provides with (?{...}). The facility is termed a "callout". The way it is done
4203 in PCRE is for the caller to provide an optional function, by setting
4204 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4205 global variable. By default it is unset, which disables all calling out. To get
4206 the function called, the regex must include (?C) at appropriate points. This
4207 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4208 This provides a means of identifying different callout points. When PCRE
4209 reaches such a point in the regex, if pcre_callout has been set, the external
4210 function is called. It is provided with data in a structure called
4211 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4212 matching continues; if it returns a non-zero value, the match at the current
4213 point fails. However, backtracking will occur if possible. [This was changed
4214 later and other features added - see item 49 below.]
4215
4216 29. pcretest is upgraded to test the callout functionality. It provides a
4217 callout function that displays information. By default, it shows the start of
4218 the match and the current position in the text. There are some new data escapes
4219 to vary what happens:
4220
4221 \C+ in addition, show current contents of captured substrings
4222 \C- do not supply a callout function
4223 \C!n return 1 when callout number n is reached
4224 \C!n!m return 1 when callout number n is reached for the mth time
4225
4226 30. If pcregrep was called with the -l option and just a single file name, it
4227 output "<stdin>" if a match was found, instead of the file name.
4228
4229 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4230 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4231 pcre_exec(). This saves a malloc/free per call. The default value of
4232 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4233 when configuring.
4234
4235 32. The default maximum size of a compiled pattern is 64K. There have been a
4236 few cases of people hitting this limit. The code now uses macros to handle the
4237 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4238 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4239 configuring. Tests 2 and 5 work only with 2-byte links because they output
4240 debugging information about compiled patterns.
4241
4242 33. Internal code re-arrangements:
4243
4244 (a) Moved the debugging function for printing out a compiled regex into
4245 its own source file (printint.c) and used #include to pull it into
4246 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4247 separate copies.
4248
4249 (b) Defined the list of op-code names for debugging as a macro in
4250 internal.h so that it is next to the definition of the opcodes.
4251
4252 (c) Defined a table of op-code lengths for simpler skipping along compiled
4253 code. This is again a macro in internal.h so that it is next to the
4254 definition of the opcodes.
4255
4256 34. Added support for recursive calls to individual subpatterns, along the
4257 lines of Robin Houston's patch (but implemented somewhat differently).
4258
4259 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4260 allow it to read and process whole directories in Win32. This code was
4261 contributed by Lionel Fourquaux; it has not been tested by me.
4262
4263 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4264 used to name a group. Names consist of alphanumerics and underscores, and must
4265 be unique. Back references use the syntax (?P=name) and recursive calls use
4266 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4267 numbers. The function pcre_fullinfo() can be used after compilation to extract
4268 a name/number map. There are three relevant calls:
4269
4270 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4271 PCRE_INFO_NAMECOUNT yields the number of entries
4272 PCRE_INFO_NAMETABLE yields a pointer to the map.
4273
4274 The map is a vector of fixed-size entries. The size of each entry depends on
4275 the length of the longest name used. The first two bytes of each entry are the
4276 group number, most significant byte first. There follows the corresponding
4277 name, zero terminated. The names are in alphabetical order.
4278
4279 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4280 case instead of 255. Making it the same both with and without UTF-8 support
4281 means that the same test output works with both.
4282
4283 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4284 calling malloc() with a zero argument.
4285
4286 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4287 optimization. I've improved things by keeping a bitmap of backreferences with
4288 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4289 fact referenced, the optimization can be applied. It is unlikely that a
4290 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4291 the match to follow \n) will appear inside brackets with a number greater than
4292 31, but if it does, any back reference > 31 suppresses the optimization.
4293
4294 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4295 of disabling numbered capturing parentheses. Any opening parenthesis that is
4296 not followed by ? behaves as if it were followed by ?: but named parentheses
4297 can still be used for capturing (and they will acquire numbers in the usual
4298 way).
4299
4300 41. Redesigned the return codes from the match() function into yes/no/error so
4301 that errors can be passed back from deep inside the nested calls. A malloc
4302 failure while inside a recursive subpattern call now causes the
4303 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4304
4305 42. It is now possible to set a limit on the number of times the match()
4306 function is called in a call to pcre_exec(). This facility makes it possible to
4307 limit the amount of recursion and backtracking, though not in a directly
4308 obvious way, because the match() function is used in a number of different
4309 circumstances. The count starts from zero for each position in the subject
4310 string (for non-anchored patterns). The default limit is, for compatibility, a
4311 large number, namely 10 000 000. You can change this in two ways:
4312
4313 (a) When configuring PCRE before making, you can use --with-match-limit=n
4314 to set a default value for the compiled library.
4315
4316 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4317 a different value is set. See 45 below.
4318
4319 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4320
4321 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4322 of things that can be changed at compile time. The first argument specifies
4323 what is wanted and the second points to where the information is to be placed.
4324 The current list of available information is:
4325
4326 PCRE_CONFIG_UTF8
4327
4328 The output is an integer that is set to one if UTF-8 support is available;
4329 otherwise it is set to zero.
4330
4331 PCRE_CONFIG_NEWLINE
4332
4333 The output is an integer that it set to the value of the code that is used for
4334 newline. It is either LF (10) or CR (13).
4335
4336 PCRE_CONFIG_LINK_SIZE
4337
4338 The output is an integer that contains the number of bytes used for internal
4339 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4340
4341 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4342
4343 The output is an integer that contains the threshold above which the POSIX
4344 interface uses malloc() for output vectors. See item 31 above.
4345
4346 PCRE_CONFIG_MATCH_LIMIT
4347
4348 The output is an unsigned integer that contains the default limit of the number
4349 of match() calls in a pcre_exec() execution. See 42 above.
4350
4351 44. pcretest has been upgraded by the addition of the -C option. This causes it
4352 to extract all the available output from the new pcre_config() function, and to
4353 output it. The program then exits immediately.
4354
4355 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4356 order to support additional features. One way would have been to define
4357 pcre_exec2() (for example) with extra arguments, but this would not have been
4358 extensible, and would also have required all calls to the original function to
4359 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4360 is used for passing in "extra" data from pcre_study().
4361
4362 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4363 contains the following fields:
4364
4365 flags a bitmap indicating which of the following fields are set
4366 study_data opaque data from pcre_study()
4367 match_limit a way of specifying a limit on match() calls for a specific
4368 call to pcre_exec()
4369 callout_data data for callouts (see 49 below)
4370
4371 The flag bits are also defined in pcre.h, and are
4372
4373 PCRE_EXTRA_STUDY_DATA
4374 PCRE_EXTRA_MATCH_LIMIT
4375 PCRE_EXTRA_CALLOUT_DATA
4376
4377 The pcre_study() function now returns one of these new pcre_extra blocks, with
4378 the actual study data pointed to by the study_data field, and the
4379 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4380 before. That is, this change is entirely upwards-compatible and requires no
4381 change to existing code.
4382
4383 If you want to pass in additional data to pcre_exec(), you can either place it
4384 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4385 block.
4386
4387 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4388 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4389 times with different match limits, until it finds the minimum value needed for
4390 pcre_exec() to complete. The value is then output. This can be instructive; for
4391 most simple matches the number is quite small, but for pathological cases it
4392 gets very large very quickly.
4393
4394 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4395 returns the size of the data block pointed to by the study_data field in a
4396 pcre_extra block, that is, the value that was passed as the argument to
4397 pcre_malloc() when PCRE was getting memory in which to place the information
4398 created by pcre_study(). The fourth argument should point to a size_t variable.
4399 pcretest has been extended so that this information is shown after a successful
4400 pcre_study() call when information about the compiled regex is being displayed.
4401
4402 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4403 because what follows is always an absolute path. (Later: it turns out that this
4404 is more than cosmetic for MinGW, because it doesn't like empty path
4405 components.)
4406
4407 49. Some changes have been made to the callout feature (see 28 above):
4408
4409 (i) A callout function now has three choices for what it returns:
4410
4411 0 => success, carry on matching
4412 > 0 => failure at this point, but backtrack if possible
4413 < 0 => serious error, return this value from pcre_exec()
4414
4415 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4416 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4417 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4418 use by callout functions. It will never be used by PCRE itself.
4419
4420 (ii) The pcre_extra structure (see 45 above) has a void * field called
4421 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4422 pcre_callout_block structure has a field of the same name. The contents of
4423 the field passed in the pcre_extra structure are passed to the callout
4424 function in the corresponding field in the callout block. This makes it
4425 easier to use the same callout-containing regex from multiple threads. For
4426 testing, the pcretest program has a new data escape
4427
4428 \C*n pass the number n (may be negative) as callout_data
4429
4430 If the callout function in pcretest receives a non-zero value as
4431 callout_data, it returns that value.
4432
4433 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4434 there were some redundant $(CFLAGS) in commands that are now specified as
4435 $(LINK), which already includes $(CFLAGS).
4436
4437 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4438 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4439 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4440 one-byte characters throughout. Note that case-insensitive matching applies
4441 only to characters whose values are less than 256. PCRE doesn't support the
4442 notion of cases for higher-valued characters.
4443
4444 (i) A character class whose characters are all within 0-255 is handled as
4445 a bit map, and the map is inverted for negative classes. Previously, a
4446 character > 255 always failed to match such a class; however it should
4447 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4448
4449 (ii) A negated character class with a single character < 255 is coded as
4450 "not this character" (OP_NOT). This wasn't working properly when the test
4451 character was multibyte, either singly or repeated.
4452
4453 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4454 mode, for example: \x{100}{2,3}.
4455
4456 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4457 singly or repeated) now correctly test multibyte characters. However,
4458 PCRE doesn't recognize any characters with values greater than 255 as
4459 digits, spaces, or word characters. Such characters always match \D, \S,
4460 and \W, and never match \d, \s, or \w.
4461
4462 (v) Classes may now contain characters and character ranges with values
4463 greater than 255. For example: [ab\x{100}-\x{400}].
4464
4465 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4466 PCRE in UTF-8 mode.
4467
4468 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4469 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4470 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4471 value.)
4472
4473 53. The single man page has become too large. I have therefore split it up into
4474 a number of separate man pages. These also give rise to individual HTML pages;
4475 these are now put in a separate directory, and there is an index.html page that
4476 lists them all. Some hyperlinking between the pages has been installed.
4477
4478 54. Added convenience functions for handling named capturing parentheses.
4479
4480 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4481 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4482 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4483 are faulted.
4484
4485 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4486 calling configure. These values are used when compiling the dftables.c program
4487 which is run to generate the source of the default character tables. They
4488 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4489 you will need to set these values.
4490
4491 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4492
4493
4494 Version 3.9 02-Jan-02
4495 ---------------------
4496
4497 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4498
4499 2. If --disable-static was given, the building process failed when trying to
4500 build pcretest and pcregrep. (For some reason it was using libtool to compile
4501 them, which is not right, as they aren't part of the library.)
4502
4503
4504 Version 3.8 18-Dec-01
4505 ---------------------
4506
4507 1. The experimental UTF-8 code was completely screwed up. It was packing the
4508 bytes in the wrong order. How dumb can you get?
4509
4510
4511 Version 3.7 29-Oct-01
4512 ---------------------
4513
4514 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4515 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4516 this didn't happen under Solaris 8, where I normally test things.
4517
4518 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4519 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4520 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4521 this fix an hour or so after the initial 3.7 release.)
4522
4523
4524 Version 3.6 23-Oct-01
4525 ---------------------
4526
4527 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4528 offsets passed as NULL with zero offset count.
4529
4530 2. The config.guess and config.sub files had not been updated when I moved to
4531 the latest autoconf.
4532
4533
4534 Version 3.5 15-Aug-01
4535 ---------------------
4536
4537 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4538 had been forgotten.
4539
4540 2. By using declared but undefined structures, we can avoid using "void"
4541 definitions in pcre.h while keeping the internal definitions of the structures
4542 private.
4543
4544 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4545 user point of view, this means that both static and shared libraries are built
4546 by default, but this can be individually controlled. More of the work of
4547 handling this static/shared cases is now inside libtool instead of PCRE's make
4548 file.
4549
4550 4. The pcretest utility is now installed along with pcregrep because it is
4551 useful for users (to test regexs) and by doing this, it automatically gets
4552 relinked by libtool. The documentation has been turned into a man page, so
4553 there are now .1, .txt, and .html versions in /doc.
4554
4555 5. Upgrades to pcregrep:
4556 (i) Added long-form option names like gnu grep.
4557 (ii) Added --help to list all options with an explanatory phrase.
4558 (iii) Added -r, --recursive to recurse into sub-directories.
4559 (iv) Added -f, --file to read patterns from a file.
4560
4561 6. pcre_exec() was referring to its "code" argument before testing that
4562 argument for NULL (and giving an error if it was NULL).
4563
4564 7. Upgraded Makefile.in to allow for compiling in a different directory from
4565 the source directory.
4566
4567 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4568 options bits, the pointer it was passed was to an int instead of to an unsigned
4569 long int. This mattered only on 64-bit systems.
4570
4571 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4572 generated) instead of pcre.in, which it its source. Also made the same change
4573 in several of the .c files.
4574
4575 10. A new release of gcc defines printf() as a macro, which broke pcretest
4576 because it had an ifdef in the middle of a string argument for printf(). Fixed
4577 by using separate calls to printf().
4578
4579 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4580 script, to force use of CR or LF instead of \n in the source. On non-Unix
4581 systems, the value can be set in config.h.
4582
4583 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4584 absolute limit. Changed the text of the error message to make this clear, and
4585 likewise updated the man page.
4586
4587 13. The limit of 99 on the number of capturing subpatterns has been removed.
4588 The new limit is 65535, which I hope will not be a "real" limit.
4589
4590
4591 Version 3.4 22-Aug-00
4592 ---------------------
4593
4594 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4595
4596 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4597
4598
4599 Version 3.3 01-Aug-00
4600 ---------------------
4601
4602 1. If an octal character was given, but the value was greater than \377, it
4603 was not getting masked to the least significant bits, as documented. This could
4604 lead to crashes in some systems.
4605
4606 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4607 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4608
4609 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4610 These just pass their arguments on to (pcre_free)(), but they are provided
4611 because some uses of PCRE bind it to non-C systems that can call its functions,
4612 but cannot call free() or pcre_free() directly.
4613
4614 4. Add "make test" as a synonym for "make check". Corrected some comments in
4615 the Makefile.
4616
4617 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4618 Makefile.
4619
4620 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4621 command called pgrep for grepping around the active processes.
4622
4623 7. Added the beginnings of support for UTF-8 character strings.
4624
4625 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4626 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4627 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4628 out for the ar command.)
4629
4630
4631 Version 3.2 12-May-00
4632 ---------------------
4633
4634 This is purely a bug fixing release.
4635
4636 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4637 of ZA. This was just one example of several cases that could provoke this bug,
4638 which was introduced by change 9 of version 2.00. The code for breaking
4639 infinite loops after an iteration that matches an empty string was't working
4640 correctly.
4641
4642 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4643 when matched against abbab (for example). After matching an empty string, it
4644 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4645 caused it to match further down the string than it should.
4646
4647 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4648 was there because it doesn't seem to be needed, and it causes trouble on some
4649 systems, as it is not a Standard C header. It has been removed.
4650
4651 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4652 were reported on the Macintosh. The changes were from
4653
4654 while ((c = *(++ptr)) != 0 && c != '\n');
4655 to
4656 while ((c = *(++ptr)) != 0 && c != '\n') ;
4657
4658 Totally extraordinary, but if that's what it takes...
4659
4660 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4661 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4662 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4663 assumes the way PCRE uses memmove() (always moving upwards).
4664
4665 6. PCRE is being used in one environment where strchr() is not available. There
4666 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4667 faster code anyway.
4668
4669
4670 Version 3.1 09-Feb-00
4671 ---------------------
4672
4673 The only change in this release is the fixing of some bugs in Makefile.in for
4674 the "install" target:
4675
4676 (1) It was failing to install pcreposix.h.
4677
4678 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4679
4680
4681 Version 3.0 01-Feb-00
4682 ---------------------
4683
4684 1. Add support for the /+ modifier to perltest (to output $` like it does in
4685 pcretest).
4686
4687 2. Add support for the /g modifier to perltest.
4688
4689 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4690 matches null strings.
4691
4692 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4693 pattern. Perl treats empty patterns specially - it reuses the most recent
4694 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4695 effect.
4696
4697 5. The POSIX interface was broken in that it was just handing over the POSIX
4698 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4699 required a bigger vector, with some working space on the end. This means that
4700 the POSIX wrapper now has to get and free some memory, and copy the results.
4701
4702 6. Added some simple autoconf support, placing the test data and the
4703 documentation in separate directories, re-organizing some of the
4704 information files, and making it build pcre-config (a GNU standard). Also added
4705 libtool support for building PCRE as a shared library, which is now the
4706 default.
4707
4708 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4709 09 are not valid octal constants. Single digits will be used for minor values
4710 less than 10.
4711
4712 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4713 existing programs that set these in the POSIX interface can use PCRE without
4714 modification.
4715
4716 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4717 return all that pcre_info() returns, plus additional data. The pcre_info()
4718 function is retained for compatibility, but is considered to be obsolete.
4719
4720 10. Added experimental recursion feature (?R) to handle one common case that
4721 Perl 5.6 will be able to do with (?p{...}).
4722
4723 11. Added support for POSIX character classes like [:alpha:], which Perl is
4724 adopting.
4725
4726
4727 Version 2.08 31-Aug-99
4728 ----------------------
4729
4730 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4731 trying to match at the startoffset position, but instead was moving forward to
4732 the next newline as if a previous match had failed.
4733
4734 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4735 and could get into a loop if a null string was matched other than at the start
4736 of the subject.
4737
4738 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4739 be distinguished at compile time, and for completeness also added PCRE_DATE.
4740
4741 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4742 in GnuWin32 environments.
4743
4744
4745 Version 2.07 29-Jul-99
4746 ----------------------
4747
4748 1. The documentation is now supplied in plain text form and HTML as well as in
4749 the form of man page sources.
4750
4751 2. C++ compilers don't like assigning (void *) values to other pointer types.
4752 In particular this affects malloc(). Although there is no problem in Standard
4753 C, I've put in casts to keep C++ compilers happy.
4754
4755 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4756 should be (const char *).
4757
4758 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4759 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4760 However, I haven't made this a standard facility. The documentation doesn't
4761 mention it, and the Makefile doesn't support it.
4762
4763 5. The Makefile now contains an "install" target, with editable destinations at
4764 the top of the file. The pcretest program is not installed.
4765
4766 6. pgrep -V now gives the PCRE version number and date.
4767
4768 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4769 causing the entire string to be ignored, instead of just the last character.
4770
4771 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4772 non-matching string, it can take a very, very long time, even for strings of
4773 quite modest length, because of the nested recursion. PCRE now does better in
4774 some of these cases. It does this by remembering the last required literal
4775 character in the pattern, and pre-searching the subject to ensure it is present
4776 before running the real match. In other words, it applies a heuristic to detect
4777 some types of certain failure quickly, and in the above example, if presented
4778 with a string that has no trailing " it gives "no match" very quickly.
4779
4780 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4781 other alternatives are tried instead.
4782
4783
4784 Version 2.06 09-Jun-99
4785 ----------------------
4786
4787 1. Change pcretest's output for amount of store used to show just the code
4788 space, because the remainder (the data block) varies in size between 32-bit and
4789 64-bit systems.
4790
4791 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4792 start matching at. This allows lookbehinds to work when searching for multiple
4793 occurrences in a string.
4794
4795 3. Added additional options to pcretest for testing multiple occurrences:
4796
4797 /+ outputs the rest of the string that follows a match
4798 /g loops for multiple occurrences, using the new startoffset argument
4799 /G loops for multiple occurrences by passing an incremented pointer
4800
4801 4. PCRE wasn't doing the "first character" optimization for patterns starting
4802 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4803 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4804 the letter 'x'. On long subject strings, this gives a significant speed-up.
4805
4806
4807 Version 2.05 21-Apr-99
4808 ----------------------
4809
4810 1. Changed the type of magic_number from int to long int so that it works
4811 properly on 16-bit systems.
4812
4813 2. Fixed a bug which caused patterns starting with .* not to work correctly
4814 when the subject string contained newline characters. PCRE was assuming
4815 anchoring for such patterns in all cases, which is not correct because .* will
4816 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4817 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4818 must be retried after every newline in the subject.
4819
4820
4821 Version 2.04 18-Feb-99
4822 ----------------------
4823
4824 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4825 computation of the store needed to hold the pattern was incorrect (too large).
4826 If such patterns were nested a few deep, this could multiply and become a real
4827 problem.
4828
4829 2. Added /M option to pcretest to show the memory requirement of a specific
4830 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4831
4832 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4833 compiled in such a way that the backtracking after subsequent failure was
4834 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4835 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4836
4837
4838 Version 2.03 02-Feb-99
4839 ----------------------
4840
4841 1. Fixed typo and small mistake in man page.
4842
4843 2. Added 4th condition (GPL supersedes if conflict) and created separate
4844 LICENCE file containing the conditions.
4845
4846 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4847 Perl, that is the internal \ allows the delimiter to be included in the
4848 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4849 the final delimiter, add \ to the end of the pattern (to test the error).
4850
4851 4. Added the convenience functions for extracting substrings after a successful
4852 match. Updated pcretest to make it able to test these functions.
4853
4854
4855 Version 2.02 14-Jan-99
4856 ----------------------
4857
4858 1. Initialized the working variables associated with each extraction so that
4859 their saving and restoring doesn't refer to uninitialized store.
4860
4861 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4862 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4863 fix the problem.
4864
4865 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4866 calls, and wasn't printing the correct value for compiling calls. Increased the
4867 default value of LOOPREPEAT, and the number of significant figures in the
4868 times.
4869
4870 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4871
4872 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4873 a building problem on Windows NT with a FAT file system.
4874
4875
4876 Version 2.01 21-Oct-98
4877 ----------------------
4878
4879 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4880 to character tables built by pcre_maketables() in the current locale. If NULL
4881 is passed, the default tables are used.
4882
4883
4884 Version 2.00 24-Sep-98
4885 ----------------------
4886
4887 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4888 it any more.
4889
4890 2. Allow quantification of (?>) groups, and make it work correctly.
4891
4892 3. The first character computation wasn't working for (?>) groups.
4893
4894 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4895 end of the subject) and add 5.005's \z, which really does match only at the
4896 very end of the subject.
4897
4898 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4899
4900 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4901 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4902 localized options. All options to pcre_study() were also removed.
4903
4904 7. Add other new features from 5.005:
4905
4906 $(?<= positive lookbehind
4907 $(?<! negative lookbehind
4908 (?imsx-imsx) added the unsetting capability
4909 such a setting is global if at outer level; local otherwise
4910 (?imsx-imsx:) non-capturing groups with option setting
4911 (?(cond)re|re) conditional pattern matching
4912
4913 A backreference to itself in a repeated group matches the previous
4914 captured string.
4915
4916 8. General tidying up of studying (both automatic and via "study")
4917 consequential on the addition of new assertions.
4918
4919 9. As in 5.005, unlimited repeated groups that could match an empty substring
4920 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4921 runtime if any iteration does actually match an empty substring.
4922
4923 10. Include the RunTest script in the distribution.
4924
4925 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4926 discrepancies, some of which were old and were also with respect to 5.004. They
4927 have now been fixed.
4928
4929
4930 Version 1.09 28-Apr-98
4931 ----------------------
4932
4933 1. A negated single character class followed by a quantifier with a minimum
4934 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4935 program crashes, or just wrong answers. This did not apply to negated classes
4936 containing more than one character, or to minima other than one.
4937
4938
4939 Version 1.08 27-Mar-98
4940 ----------------------
4941
4942 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4943
4944 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4945 latter must appear before anything that relies on it in the pattern.
4946
4947
4948 Version 1.07 16-Feb-98
4949 ----------------------
4950
4951 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4952 repeat of a potentially empty string).
4953
4954
4955 Version 1.06 23-Jan-98
4956 ----------------------
4957
4958 1. Added Markus Oberhumer's little patches for C++.
4959
4960 2. Literal strings longer than 255 characters were broken.
4961
4962
4963 Version 1.05 23-Dec-97
4964 ----------------------
4965
4966 1. Negated character classes containing more than one character were failing if
4967 PCRE_CASELESS was set at run time.
4968
4969
4970 Version 1.04 19-Dec-97
4971 ----------------------
4972
4973 1. Corrected the man page, where some "const" qualifiers had been omitted.
4974
4975 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4976 input syntax.
4977
4978 3. Fixed memory leak which occurred when a regex with back references was
4979 matched with an offsets vector that wasn't big enough. The temporary memory
4980 that is used in this case wasn't being freed if the match failed.
4981
4982 4. Tidied pcretest to ensure it frees memory that it gets.
4983
4984 5. Temporary memory was being obtained in the case where the passed offsets
4985 vector was exactly big enough.
4986
4987 6. Corrected definition of offsetof() from change 5 below.
4988
4989 7. I had screwed up change 6 below and broken the rules for the use of
4990 setjmp(). Now fixed.
4991
4992
4993 Version 1.03 18-Dec-97
4994 ----------------------
4995
4996 1. A erroneous regex with a missing opening parenthesis was correctly
4997 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4998 on some systems.
4999
5000 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5001 it was reported that one broken compiler failed on the former because "code" is
5002 also an independent variable.
5003
5004 3. The erroneous regex a[]b caused an array overrun reference.
5005
5006 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5007 fail on data ending with that character. (It was going on too far, and checking
5008 the next character, typically a binary zero.) This was specific to the
5009 optimized code for single-character negative classes.
5010
5011 5. Added a contributed patch from the TIN world which does the following:
5012
5013 + Add an undef for memmove, in case the the system defines a macro for it.
5014
5015 + Add a definition of offsetof(), in case there isn't one. (I don't know
5016 the reason behind this - offsetof() is part of the ANSI standard - but
5017 it does no harm).
5018
5019 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5020 most of the places where whitespace preceded '#'. I have given up and
5021 allowed the remaining 2 cases to be at the margin.
5022
5023 + Rename some variables in pcre to eliminate shadowing. This seems very
5024 pedantic, but does no harm, of course.
5025
5026 6. Moved the call to setjmp() into its own function, to get rid of warnings
5027 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5028
5029 7. Constructs such as \d{8,} were compiling into the equivalent of
5030 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5031 outcome, but in this particular case used more store than had been allocated,
5032 which caused the bug to be discovered because it threw up an internal error.
5033
5034 8. The debugging code in both pcre and pcretest for outputting the compiled
5035 form of a regex was going wrong in the case of back references followed by
5036 curly-bracketed repeats.
5037
5038
5039 Version 1.02 12-Dec-97
5040 ----------------------
5041
5042 1. Typos in pcre.3 and comments in the source fixed.
5043
5044 2. Applied a contributed patch to get rid of places where it used to remove
5045 'const' from variables, and fixed some signed/unsigned and uninitialized
5046 variable warnings.
5047
5048 3. Added the "runtest" target to Makefile.
5049
5050 4. Set default compiler flag to -O2 rather than just -O.
5051
5052
5053 Version 1.01 19-Nov-97
5054 ----------------------
5055
5056 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5057 like /([ab]*)*/, that is, for classes with more than one character in them.
5058
5059 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5060 as /((?>a*))*/ (a PCRE_EXTRA facility).
5061
5062
5063 Version 1.00 18-Nov-97
5064 ----------------------
5065
5066 1. Added compile-time macros to support systems such as SunOS4 which don't have
5067 memmove() or strerror() but have other things that can be used instead.
5068
5069 2. Arranged that "make clean" removes the executables.
5070
5071
5072 Version 0.99 27-Oct-97
5073 ----------------------
5074
5075 1. Fixed bug in code for optimizing classes with only one character. It was
5076 initializing a 32-byte map regardless, which could cause it to run off the end
5077 of the memory it had got.
5078
5079 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5080
5081
5082 Version 0.98 22-Oct-97
5083 ----------------------
5084
5085 1. Fixed bug in code for handling temporary memory usage when there are more
5086 back references than supplied space in the ovector. This could cause segfaults.
5087
5088
5089 Version 0.97 21-Oct-97
5090 ----------------------
5091
5092 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5093
5094 2. Optimized negated single characters not to use a bit map.
5095
5096 3. Brought error texts together as macro definitions; clarified some of them;
5097 fixed one that was wrong - it said "range out of order" when it meant "invalid
5098 escape sequence".
5099
5100 4. Changed some char * arguments to const char *.
5101
5102 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5103
5104 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5105 pcretest.
5106
5107
5108 Version 0.96 16-Oct-97
5109 ----------------------
5110
5111 1. Added a simple "pgrep" utility to the distribution.
5112
5113 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5114 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5115 where "ddd" means "one or more decimal digits".
5116
5117 3. Fixed serious bug. If a pattern had a back reference, but the call to
5118 pcre_exec() didn't supply a large enough ovector to record the related
5119 identifying subpattern, the match always failed. PCRE now remembers the number
5120 of the largest back reference, and gets some temporary memory in which to save
5121 the offsets during matching if necessary, in order to ensure that
5122 backreferences always work.
5123
5124 4. Increased the compatibility with Perl in a number of ways:
5125
5126 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5127 to request this handling. The option can be set at compile or exec time.
5128
5129 (b) $ matches before a terminating newline by default; an option
5130 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5131 mode). The option can be set at compile or exec time.
5132
5133 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5134 the same as Perl's. If the decimal number it represents is less than 10
5135 or there aren't that many previous left capturing parentheses, an octal
5136 escape is read. Inside a character class, it's always an octal escape,
5137 even if it is a single digit.
5138
5139 (d) An escaped but undefined alphabetic character is taken as a literal,
5140 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5141 escapes.
5142
5143 (e) {0} is now permitted. (The previous item is removed from the compiled
5144 pattern).
5145
5146 5. Changed all the names of code files so that the basic parts are no longer
5147 than 10 characters, and abolished the teeny "globals.c" file.
5148
5149 6. Changed the handling of character classes; they are now done with a 32-byte
5150 bit map always.
5151
5152 7. Added the -d and /D options to pcretest to make it possible to look at the
5153 internals of compilation without having to recompile pcre.
5154
5155
5156 Version 0.95 23-Sep-97
5157 ----------------------
5158
5159 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5160 \x20 at the start of a run of normal characters. These were being treated as
5161 real characters, instead of the source characters being re-checked.
5162
5163
5164 Version 0.94 18-Sep-97
5165 ----------------------
5166
5167 1. The functions are now thread-safe, with the caveat that the global variables
5168 containing pointers to malloc() and free() or alternative functions are the
5169 same for all threads.
5170
5171 2. Get pcre_study() to generate a bitmap of initial characters for non-
5172 anchored patterns when this is possible, and use it if passed to pcre_exec().
5173
5174
5175 Version 0.93 15-Sep-97
5176 ----------------------
5177
5178 1. /(b)|(:+)/ was computing an incorrect first character.
5179
5180 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5181 but not actually doing anything yet.
5182
5183 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5184 as Perl does (e.g. [-az] or [az-]).
5185
5186 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5187 all possible positions.
5188
5189 5. Split up into different modules to avoid including unneeded functions in a
5190 compiled binary. However, compile and exec are still in one module. The "study"
5191 function is split off.
5192
5193 6. The character tables are now in a separate module whose source is generated
5194 by an auxiliary program - but can then be edited by hand if required. There are
5195 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5196 toupper() in the code.
5197
5198 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5199 make them global. Abolish the function for setting them, as the caller can now
5200 set them directly.
5201
5202
5203 Version 0.92 11-Sep-97
5204 ----------------------
5205
5206 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5207 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5208
5209 2. Caseless matching was not working in character classes if the characters in
5210 the pattern were in upper case.
5211
5212 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5213
5214 4. Make PCRE_ANCHORED public and accept as a compile option.
5215
5216 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5217 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5218 pass them.
5219
5220 6. Give an error if bad option bits passed at compile or run time.
5221
5222 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5223 pcretest to cause it to pass that flag.
5224
5225 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5226 options, and the first character, if set.
5227
5228 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5229
5230
5231 Version 0.91 10-Sep-97
5232 ----------------------
5233
5234 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5235 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5236
5237 2. PCRE was looping on encountering an indefinitely repeated back reference to
5238 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5239 Perl does - treats the match as successful.
5240
5241 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5