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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1350 - (show annotations)
Fri Jul 26 10:03:38 2013 UTC (6 years, 2 months ago) by ph10
File size: 250148 byte(s)
Fix backup bugs with \X repeat matches.
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
56 Version 8.33 28-May-2013
57 --------------------------
58
59 1. Added 'U' to some constants that are compared to unsigned integers, to
60 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
61 variables that are added to signed variables, to ensure the result is
62 signed and can be negated.
63
64 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
65 CMake config files.
66
67 3. Revise the creation of config.h.generic so that all boolean macros are
68 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
69 overriding via -D on the command line possible.
70
71 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
72 to unsigned int is reported to make a quite noticeable speed difference in
73 a specific Windows environment. Testing on Linux did also appear to show
74 some benefit (and it is clearly not harmful). Also fixed the definition of
75 Xop which should be unsigned.
76
77 5. Related to (4), changing the definition of the intermediate variable cc
78 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
79 improvements.
80
81 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
82 unnecessary spaces.
83
84 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
85 and later.
86
87 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
88
89 9. Optimizing fast_forward_start_bits in JIT.
90
91 10. Adding support for callouts in JIT, and fixing some issues revealed
92 during this work. Namely:
93
94 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
95
96 (b) Minimum length was not checked before the matching is started.
97
98 11. The value of capture_last that is passed to callouts was incorrect in some
99 cases when there was a capture on one path that was subsequently abandoned
100 after a backtrack. Also, the capture_last value is now reset after a
101 recursion, since all captures are also reset in this case.
102
103 12. The interpreter no longer returns the "too many substrings" error in the
104 case when an overflowing capture is in a branch that is subsequently
105 abandoned after a backtrack.
106
107 13. In the pathological case when an offset vector of size 2 is used, pcretest
108 now prints out the matched string after a yield of 0 or 1.
109
110 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
111 Only supported by the JIT compiler at the moment.
112
113 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
114
115 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
116 offset of the starting point of the matching process, provided the offsets
117 vector is large enough.
118
119 17. The \A escape now records a lookbehind value of 1, though its execution
120 does not actually inspect the previous character. This is to ensure that,
121 in partial multi-segment matching, at least one character from the old
122 segment is retained when a new segment is processed. Otherwise, if there
123 are no lookbehinds in the pattern, \A might match incorrectly at the start
124 of a new segment.
125
126 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
127
128 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
129 gives some modest performance improvement in 8-bit mode.
130
131 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
132 be expressed in certain programming languages using Universal Character
133 Names.
134
135 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
136 which points out that "non characters" are not "characters that may not
137 appear in Unicode strings" but rather "characters that are reserved for
138 internal use and have only local meaning".
139
140 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
141 there was a conditional group that depended on an assertion, if the
142 assertion was false, the callout that immediately followed the alternation
143 in the condition was skipped when pcre_exec() was used for matching.
144
145 23. Allow an explicit callout to be inserted before an assertion that is the
146 condition for a conditional group, for compatibility with automatic
147 callouts, which always insert a callout at this point.
148
149 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
150 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
151
152 25. (*PRUNE) is now supported by the JIT compiler.
153
154 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
155
156 27. Fix the case where there are two or more SKIPs with arguments that may be
157 ignored.
158
159 28. (*SKIP) is now supported by the JIT compiler.
160
161 29. (*THEN) is now supported by the JIT compiler.
162
163 30. Update RunTest with additional test selector options.
164
165 31. The way PCRE handles backtracking verbs has been changed in two ways.
166
167 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
168 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
169 backtracking. In some cases this makes it more Perl-compatible, but Perl's
170 rather obscure rules do not always do the same thing.
171
172 (2) Previously, backtracking verbs were confined within assertions. This is
173 no longer the case for positive assertions, except for (*ACCEPT). Again,
174 this sometimes improves Perl compatibility, and sometimes does not.
175
176 32. A number of tests that were in test 2 because Perl did things differently
177 have been moved to test 1, because either Perl or PCRE has changed, and
178 these tests are now compatible.
179
180 32. Backtracking control verbs are now handled in the same way in JIT and
181 interpreter.
182
183 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
184 contained a forward subroutine reference caused a compile error.
185
186 34. Auto-detect and optimize limited repetitions in JIT.
187
188 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
189 blocking (*UTF) etc.
190
191 36. In the interpreter, maximizing pattern repetitions for characters and
192 character types now use tail recursion, which reduces stack usage.
193
194 37. The value of the max lookbehind was not correctly preserved if a compiled
195 and saved regex was reloaded on a host of different endianness.
196
197 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
198 of the compiled pattern block, expand the flags field from 16 to 32 bits
199 because it was almost full.
200
201 39. Try madvise first before posix_madvise.
202
203 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
204 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
205 original change locked out all matches of empty strings. This has been
206 changed so that one match of an empty string per line is recognized.
207 Subsequent searches on the same line (for colouring or for --only-matching,
208 for example) do not recognize empty strings.
209
210 41. Applied a user patch to fix a number of spelling mistakes in comments.
211
212 42. Data lines longer than 65536 caused pcretest to crash.
213
214 43. Clarified the data type for length and startoffset arguments for pcre_exec
215 and pcre_dfa_exec in the function-specific man pages, where they were
216 explicitly stated to be in bytes, never having been updated. I also added
217 some clarification to the pcreapi man page.
218
219 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
220 a segmentation fault.
221
222
223 Version 8.32 30-November-2012
224 -----------------------------
225
226 1. Improved JIT compiler optimizations for first character search and single
227 character iterators.
228
229 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
230 Patch by Daniel Richard G.
231
232 3. Single character iterator optimizations in the JIT compiler.
233
234 4. Improved JIT compiler optimizations for character ranges.
235
236 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
237 Reported by Giuseppe D'Angelo.
238
239 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
240 of a line, was being set incorrectly in cases where .* appeared inside
241 atomic brackets at the start of a pattern, or where there was a subsequent
242 *PRUNE or *SKIP.
243
244 7. Improved instruction cache flush for POWER/PowerPC.
245 Patch by Daniel Richard G.
246
247 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
248 grep:
249
250 (a) There is now no limit to the number of patterns to be matched.
251
252 (b) An error is given if a pattern is too long.
253
254 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
255 are now supported.
256
257 (d) --exclude-from and --include-from (multiple use) have been added.
258
259 (e) Exclusions and inclusions now apply to all files and directories, not
260 just to those obtained from scanning a directory recursively.
261
262 (f) Multiple uses of -f and --file-list are now supported.
263
264 (g) In a Windows environment, the default for -d has been changed from
265 "read" (the GNU grep default) to "skip", because otherwise the presence
266 of a directory in the file list provokes an error.
267
268 (h) The documentation has been revised and clarified in places.
269
270 9. Improve the matching speed of capturing brackets.
271
272 10. Changed the meaning of \X so that it now matches a Unicode extended
273 grapheme cluster.
274
275 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
276 out POSIX threads when JIT support is configured.
277
278 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
279
280 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
281 structure could go wrong in environments where size_t is not the same size
282 as int.
283
284 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
285 set.
286
287 15. The EBCDIC support had decayed; later updates to the code had included
288 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
289 general tidy up of EBCDIC-related issues, and the documentation was also
290 not quite right. There is now a test that can be run on ASCII systems to
291 check some of the EBCDIC-related things (but is it not a full test).
292
293 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
294 in a small tidy to the code.
295
296 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
297
298 18. If the --only-matching (-o) option in pcregrep is specified multiple
299 times, each one causes appropriate output. For example, -o1 -o2 outputs the
300 substrings matched by the 1st and 2nd capturing parentheses. A separating
301 string can be specified by --om-separator (default empty).
302
303 19. Improving the first n character searches.
304
305 20. Turn case lists for horizontal and vertical white space into macros so that
306 they are defined only once.
307
308 21. This set of changes together give more compatible Unicode case-folding
309 behaviour for characters that have more than one other case when UCP
310 support is available.
311
312 (a) The Unicode property table now has offsets into a new table of sets of
313 three or more characters that are case-equivalent. The MultiStage2.py
314 script that generates these tables (the pcre_ucd.c file) now scans
315 CaseFolding.txt instead of UnicodeData.txt for character case
316 information.
317
318 (b) The code for adding characters or ranges of characters to a character
319 class has been abstracted into a generalized function that also handles
320 case-independence. In UTF-mode with UCP support, this uses the new data
321 to handle characters with more than one other case.
322
323 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
324 whose other case is greater than 256 are now correctly matched
325 caselessly. Previously, the high codepoint matched the low one, but not
326 vice versa.
327
328 (d) The processing of \h, \H, \v, and \ in character classes now makes use
329 of the new class addition function, using character lists defined as
330 macros alongside the case definitions of 20 above.
331
332 (e) Caseless back references now work with characters that have more than
333 one other case.
334
335 (f) General caseless matching of characters with more than one other case
336 is supported.
337
338 22. Unicode character properties were updated from Unicode 6.2.0
339
340 23. Improved CMake support under Windows. Patch by Daniel Richard G.
341
342 24. Add support for 32-bit character strings, and UTF-32
343
344 25. Major JIT compiler update (code refactoring and bugfixing).
345 Experimental Sparc 32 support is added.
346
347 26. Applied a modified version of Daniel Richard G's patch to create
348 pcre.h.generic and config.h.generic by "make" instead of in the
349 PrepareRelease script.
350
351 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
352 pcre_compile.c when checking for a zero character.
353
354 28. Introducing a native interface for JIT. Through this interface, the compiled
355 machine code can be directly executed. The purpose of this interface is to
356 provide fast pattern matching, so several sanity checks are not performed.
357 However, feature tests are still performed. The new interface provides
358 1.4x speedup compared to the old one.
359
360 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
361 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
362 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
363
364 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
365 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
366 the "old" RFC 2279). Instead, it ended up passing a negative length to
367 pcre_exec().
368
369 31. Add support for GCC's visibility feature to hide internal functions.
370
371 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
372 "unknown -C option" after outputting 0 or 1.
373
374 33. There is now support for generating a code coverage report for the test
375 suite in environments where gcc is the compiler and lcov is installed. This
376 is mainly for the benefit of the developers.
377
378 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
379 unaddressable using valgrind annotations, allowing valgrind to detect
380 invalid memory accesses. This is mainly for the benefit of the developers.
381
382 25. (*UTF) can now be used to start a pattern in any of the three libraries.
383
384 26. Give configure error if --enable-cpp but no C++ compiler found.
385
386
387 Version 8.31 06-July-2012
388 -------------------------
389
390 1. Fixing a wrong JIT test case and some compiler warnings.
391
392 2. Removed a bashism from the RunTest script.
393
394 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
395 to unsigned type, result still unsigned" that was given by an MS compiler
396 on encountering the code "-sizeof(xxx)".
397
398 4. Partial matching support is added to the JIT compiler.
399
400 5. Fixed several bugs concerned with partial matching of items that consist
401 of more than one character:
402
403 (a) /^(..)\1/ did not partially match "aba" because checking references was
404 done on an "all or nothing" basis. This also applied to repeated
405 references.
406
407 (b) \R did not give a hard partial match if \r was found at the end of the
408 subject.
409
410 (c) \X did not give a hard partial match after matching one or more
411 characters at the end of the subject.
412
413 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
414 a partial match for the string "\r".
415
416 (e) When newline was set to CRLF, the metacharacter "." did not recognize
417 a partial match for a CR character at the end of the subject string.
418
419 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
420 running pcretest, the text "(JIT)" added to the output whenever JIT is
421 actually used to run the match.
422
423 7. Individual JIT compile options can be set in pcretest by following -s+[+]
424 or /S+[+] with a digit between 1 and 7.
425
426 8. OP_NOT now supports any UTF character not just single-byte ones.
427
428 9. (*MARK) control verb is now supported by the JIT compiler.
429
430 10. The command "./RunTest list" lists the available tests without actually
431 running any of them. (Because I keep forgetting what they all are.)
432
433 11. Add PCRE_INFO_MAXLOOKBEHIND.
434
435 12. Applied a (slightly modified) user-supplied patch that improves performance
436 when the heap is used for recursion (compiled with --disable-stack-for-
437 recursion). Instead of malloc and free for each heap frame each time a
438 logical recursion happens, frames are retained on a chain and re-used where
439 possible. This sometimes gives as much as 30% improvement.
440
441 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
442 call.
443
444 14. As documented, (*COMMIT) is now confined to within a positive assertion.
445
446 15. It is now possible to link pcretest with libedit as an alternative to
447 libreadline.
448
449 16. (*COMMIT) control verb is now supported by the JIT compiler.
450
451 17. The Unicode data tables have been updated to Unicode 6.1.0.
452
453 18. Added --file-list option to pcregrep.
454
455 19. Added binary file support to pcregrep, including the -a, --binary-files,
456 -I, and --text options.
457
458 20. The madvise function is renamed for posix_madvise for QNX compatibility
459 reasons. Fixed by Giuseppe D'Angelo.
460
461 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
462
463 22. Changed the option for creating symbolic links for 16-bit man pages from
464 -s to -sf so that re-installing does not cause issues.
465
466 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
467
468 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
469 matches in certain environments (the workspace was not being correctly
470 retained). Also added to pcre_dfa_exec() a simple plausibility check on
471 some of the workspace data at the beginning of a restart.
472
473 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
474 was not doing so when it should - probably a typo introduced by SVN 528
475 (change 8.10/14).
476
477 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
478 \w+ when the character tables indicated that \x{c4} was a word character.
479 There were several related cases, all because the tests for doing a table
480 lookup were testing for characters less than 127 instead of 255.
481
482 27. If a pattern contains capturing parentheses that are not used in a match,
483 their slots in the ovector are set to -1. For those that are higher than
484 any matched groups, this happens at the end of processing. In the case when
485 there were back references that the ovector was too small to contain
486 (causing temporary malloc'd memory to be used during matching), and the
487 highest capturing number was not used, memory off the end of the ovector
488 was incorrectly being set to -1. (It was using the size of the temporary
489 memory instead of the true size.)
490
491 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
492 ovector size, it uses memory at the end of the block that it has got.
493
494 29. Check for an overlong MARK name and give an error at compile time. The
495 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
496
497 30. JIT compiler update.
498
499 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
500 Rill for the patch.
501
502 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
503
504 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
505
506 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
507 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
508 was enabled.
509
510 35. Improve JIT code generation for greedy plus quantifier.
511
512 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
513 1 to "aa" instead of to an empty string. The bug affected repeated groups
514 that could potentially match an empty string.
515
516 37. Optimizing single character iterators in JIT.
517
518 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
519 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
520 codepoints that are too big for the mode are faulted, and in a UTF mode,
521 disallowed codepoints are also faulted.
522
523 39. If PCRE was compiled with UTF support, in three places in the DFA
524 matcher there was code that should only have been obeyed in UTF mode, but
525 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
526 processing when bytes with values greater than 127 were present. In 16-bit
527 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
528 both cases the values are those that cannot be the first data item in a UTF
529 character. The three items that might have provoked this were recursions,
530 possessively repeated groups, and atomic groups.
531
532 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
533 and pcregrep, because some OS require shared objects to be explicitly
534 passed to ld, causing the link step to fail if they are not.
535
536 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
537 mode, patterns that started with \h* or \R* might be incorrectly matched.
538
539
540 Version 8.30 04-February-2012
541 -----------------------------
542
543 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
544 name is defined in ctype.h.
545
546 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
547 only in quite long subpatterns.
548
549 3. Removed the function pcre_info(), which has been obsolete and deprecated
550 since it was replaced by pcre_fullinfo() in February 2000.
551
552 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
553 match a (*MARK), and the match failed at the start of the subject, a
554 reference to memory before the start of the subject could occur. This bug
555 was introduced by fix 17 of release 8.21.
556
557 5. A reference to an unset group with zero minimum repetition was giving
558 totally wrong answers (in non-JavaScript-compatibility mode). For example,
559 /(another)?(\1?)test/ matched against "hello world test". This bug was
560 introduced in release 8.13.
561
562 6. Add support for 16-bit character strings (a large amount of work involving
563 many changes and refactorings).
564
565 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
566 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
567 from a file.
568
569 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
570 rounding is not applied in this particular case).
571
572 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
573 if they appear, or are escaped, in patterns.
574
575 10. Get rid of a number of -Wunused-but-set-variable warnings.
576
577 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
578 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
579 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
580 also returns the mark "x". This bug applied to capturing parentheses,
581 non-capturing parentheses, and atomic parentheses. It also applied to some
582 assertions.
583
584 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
585 information out of configure.ac instead of relying on pcre.h.generic, which
586 is not stored in the repository.
587
588 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
589 -lreadline.
590
591 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
592
593 15. Applied Graycode's patch to put the top-level frame on the stack rather
594 than the heap when not using the stack for recursion. This gives a
595 performance improvement in many cases when recursion is not deep.
596
597 16. Experimental code added to "pcretest -C" to output the stack frame size.
598
599
600 Version 8.21 12-Dec-2011
601 ------------------------
602
603 1. Updating the JIT compiler.
604
605 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
606 are added as well.
607
608 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
609 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
610 calling _pcre_jit_exec. Some extra comments are added.
611
612 4. (*MARK) settings inside atomic groups that do not contain any capturing
613 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
614 was introduced by change 18 for 8.20.
615
616 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
617 ECMA-262 standard.
618
619 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
620 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
621 This bug was probably introduced by change 9 of 8.13.
622
623 7. While fixing 6 above, I noticed that a number of other items were being
624 incorrectly rejected as "not fixed length". This arose partly because newer
625 opcodes had not been added to the fixed-length checking code. I have (a)
626 corrected the bug and added tests for these items, and (b) arranged for an
627 error to occur if an unknown opcode is encountered while checking for fixed
628 length instead of just assuming "not fixed length". The items that were
629 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
630 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
631 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
632
633 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
634 being incorrectly compiled and would have given unpredicatble results.
635
636 9. A possessively repeated subpattern with minimum repeat count greater than
637 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
638 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
639 the first (A) could occur when it should not.
640
641 10. Add a cast and remove a redundant test from the code.
642
643 11. JIT should use pcre_malloc/pcre_free for allocation.
644
645 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
646 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
647 is anything other than /usr, -L is still shown).
648
649 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
650
651 14. Perl does not support \N without a following name in a [] class; PCRE now
652 also gives an error.
653
654 15. If a forward reference was repeated with an upper limit of around 2000,
655 it caused the error "internal error: overran compiling workspace". The
656 maximum number of forward references (including repeats) was limited by the
657 internal workspace, and dependent on the LINK_SIZE. The code has been
658 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
659 the default depends on LINK_SIZE. There is a new upper limit (for safety)
660 of around 200,000 forward references. While doing this, I also speeded up
661 the filling in of repeated forward references.
662
663 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
664 incorrectly expecting the subject to contain another "a" after the start.
665
666 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
667 in the match, the SKIP should be ignored. This was not happening; instead
668 the SKIP was being treated as NOMATCH. For patterns such as
669 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
670 tested.
671
672 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
673 now much more compatible with Perl, in particular in cases where the result
674 is a non-match for a non-anchored pattern. For example, if
675 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
676 "m", where previously it did not return a name. A side effect of this
677 change is that for partial matches, the last encountered mark name is
678 returned, as for non matches. A number of tests that were previously not
679 Perl-compatible have been moved into the Perl-compatible test files. The
680 refactoring has had the pleasing side effect of removing one argument from
681 the match() function, thus reducing its stack requirements.
682
683 19. If the /S+ option was used in pcretest to study a pattern using JIT,
684 subsequent uses of /S (without +) incorrectly behaved like /S+.
685
686 21. Retrieve executable code size support for the JIT compiler and fixing
687 some warnings.
688
689 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
690 not work when the shorter character appeared right at the end of the
691 subject string.
692
693 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
694 systems.
695
696 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
697 output it when the /M option is used in pcretest.
698
699 25. The CheckMan script was not being included in the distribution. Also, added
700 an explicit "perl" to run Perl scripts from the PrepareRelease script
701 because this is reportedly needed in Windows.
702
703 26. If study data was being save in a file and studying had not found a set of
704 "starts with" bytes for the pattern, the data written to the file (though
705 never used) was taken from uninitialized memory and so caused valgrind to
706 complain.
707
708 27. Updated RunTest.bat as provided by Sheri Pierce.
709
710 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
711
712 29. Computation of memory usage for the table of capturing group names was
713 giving an unnecessarily large value.
714
715
716 Version 8.20 21-Oct-2011
717 ------------------------
718
719 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
720 a POSIX class. After further experiments with Perl, which convinced me that
721 Perl has bugs and confusions, a closing square bracket is no longer allowed
722 in a POSIX name. This bug also affected patterns with classes that started
723 with full stops.
724
725 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
726 captured substring, but while checking the failing first alternative,
727 substring 1 is temporarily captured. If the output vector supplied to
728 pcre_exec() was not big enough for this capture, the yield of the function
729 was still zero ("insufficient space for captured substrings"). This cannot
730 be totally fixed without adding another stack variable, which seems a lot
731 of expense for a edge case. However, I have improved the situation in cases
732 such as /(a)(b)x|abc/ matched against "abc", where the return code
733 indicates that fewer than the maximum number of slots in the ovector have
734 been set.
735
736 3. Related to (2) above: when there are more back references in a pattern than
737 slots in the output vector, pcre_exec() uses temporary memory during
738 matching, and copies in the captures as far as possible afterwards. It was
739 using the entire output vector, but this conflicts with the specification
740 that only 2/3 is used for passing back captured substrings. Now it uses
741 only the first 2/3, for compatibility. This is, of course, another edge
742 case.
743
744 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
745 main code base, and can be used by building with --enable-jit. When this is
746 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
747 runtime --no-jit option is given.
748
749 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
750 ovector, the return from the function was zero, implying that there were
751 other matches that did not fit. The correct "exactly full" value is now
752 returned.
753
754 6. If a subpattern that was called recursively or as a subroutine contained
755 (*PRUNE) or any other control that caused it to give a non-standard return,
756 invalid errors such as "Error -26 (nested recursion at the same subject
757 position)" or even infinite loops could occur.
758
759 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
760 computing the minimum length on reaching *ACCEPT, and so ended up with the
761 wrong value of 1 rather than 0. Further investigation indicates that
762 computing a minimum subject length in the presence of *ACCEPT is difficult
763 (think back references, subroutine calls), and so I have changed the code
764 so that no minimum is registered for a pattern that contains *ACCEPT.
765
766 8. If (*THEN) was present in the first (true) branch of a conditional group,
767 it was not handled as intended. [But see 16 below.]
768
769 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
770 Sheri Pierce.
771
772 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
773 the first byte in a match must be "a".
774
775 11. Change 17 for 8.13 increased the recursion depth for patterns like
776 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
777 pattern contains any instances of (*THEN). If it does not, the old
778 optimizations are restored. It would be nice to do this on a per-group
779 basis, but at the moment that is not feasible.
780
781 12. In some environments, the output of pcretest -C is CRLF terminated. This
782 broke RunTest's code that checks for the link size. A single white space
783 character after the value is now allowed for.
784
785 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
786 For "fr", it uses the Windows-specific input and output files.
787
788 14. If (*THEN) appeared in a group that was called recursively or as a
789 subroutine, it did not work as intended. [But see next item.]
790
791 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
792 pattern fragments (but not containing any | characters). If A and B are
793 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
794 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
795 D. In other words, Perl considers parentheses that do not contain any |
796 characters to be part of a surrounding alternative, whereas PCRE was
797 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
798 differently. PCRE now behaves in the same way as Perl, except in the case
799 of subroutine/recursion calls such as (?1) which have in any case always
800 been different (but PCRE had them first :-).
801
802 16. Related to 15 above: Perl does not treat the | in a conditional group as
803 creating alternatives. Such a group is treated in the same way as an
804 ordinary group without any | characters when processing (*THEN). PCRE has
805 been changed to match Perl's behaviour.
806
807 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
808 RunGrepTest script failed.
809
810 18. Change 22 for version 13 caused atomic groups to use more stack. This is
811 inevitable for groups that contain captures, but it can lead to a lot of
812 stack use in large patterns. The old behaviour has been restored for atomic
813 groups that do not contain any capturing parentheses.
814
815 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
816 suppress the check for a minimum subject length at run time. (If it was
817 given to pcre_exec() or pcre_dfa_exec() it did work.)
818
819 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
820 fail to work when decoding hex characters in data strings in EBCDIC
821 environments.
822
823 21. It appears that in at least one Mac OS environment, the isxdigit() function
824 is implemented as a macro that evaluates to its argument more than once,
825 contravening the C 90 Standard (I haven't checked a later standard). There
826 was an instance in pcretest which caused it to go wrong when processing
827 \x{...} escapes in subject strings. The has been rewritten to avoid using
828 things like p++ in the argument of isxdigit().
829
830
831 Version 8.13 16-Aug-2011
832 ------------------------
833
834 1. The Unicode data tables have been updated to Unicode 6.0.0.
835
836 2. Two minor typos in pcre_internal.h have been fixed.
837
838 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
839 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
840 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
841
842 4. There were a number of related bugs in the code for matching backrefences
843 caselessly in UTF-8 mode when codes for the characters concerned were
844 different numbers of bytes. For example, U+023A and U+2C65 are an upper
845 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
846 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
847 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
848 2-byte code at the end of the subject (it thought there wasn't enough data
849 left).
850
851 5. Comprehensive information about what went wrong is now returned by
852 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
853 as the output vector has at least 2 elements. The offset of the start of
854 the failing character and a reason code are placed in the vector.
855
856 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
857 now returned is for the first byte of the failing character, instead of the
858 last byte inspected. This is an incompatible change, but I hope it is small
859 enough not to be a problem. It makes the returned offset consistent with
860 pcre_exec() and pcre_dfa_exec().
861
862 7. pcretest now gives a text phrase as well as the error number when
863 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
864 failure, the offset and reason code are output.
865
866 8. When \R was used with a maximizing quantifier it failed to skip backwards
867 over a \r\n pair if the subsequent match failed. Instead, it just skipped
868 back over a single character (\n). This seems wrong (because it treated the
869 two characters as a single entity when going forwards), conflicts with the
870 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
871 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
872 has been changed.
873
874 9. Some internal refactoring has changed the processing so that the handling
875 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
876 time (the PCRE_DOTALL option was changed this way some time ago: version
877 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
878 which was always a bit of a fudge. It also means that there is one less
879 argument for the match() function, which reduces its stack requirements
880 slightly. This change also fixes an incompatibility with Perl: the pattern
881 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
882
883 10. More internal refactoring has drastically reduced the number of recursive
884 calls to match() for possessively repeated groups such as (abc)++ when
885 using pcre_exec().
886
887 11. While implementing 10, a number of bugs in the handling of groups were
888 discovered and fixed:
889
890 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
891 (a|)*(?1) gave a compile-time internal error.
892 ((a|)+)+ did not notice that the outer group could match an empty string.
893 (^a|^)+ was not marked as anchored.
894 (.*a|.*)+ was not marked as matching at start or after a newline.
895
896 12. Yet more internal refactoring has removed another argument from the match()
897 function. Special calls to this function are now indicated by setting a
898 value in a variable in the "match data" data block.
899
900 13. Be more explicit in pcre_study() instead of relying on "default" for
901 opcodes that mean there is no starting character; this means that when new
902 ones are added and accidentally left out of pcre_study(), testing should
903 pick them up.
904
905 14. The -s option of pcretest has been documented for ages as being an old
906 synonym of -m (show memory usage). I have changed it to mean "force study
907 for every regex", that is, assume /S for every regex. This is similar to -i
908 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
909 using it. It makes it easier to run collections of tests with and without
910 study enabled, and thereby test pcre_study() more easily. All the standard
911 tests are now run with and without -s (but some patterns can be marked as
912 "never study" - see 20 below).
913
914 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
915 restoration of the capturing data to the outer values was not happening
916 correctly.
917
918 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
919 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
920 pattern had matched an empty string, and so incorrectly returned a no
921 match.
922
923 17. There was optimizing code for the last branch of non-capturing parentheses,
924 and also for the obeyed branch of a conditional subexpression, which used
925 tail recursion to cut down on stack usage. Unfortunately, now that there is
926 the possibility of (*THEN) occurring in these branches, tail recursion is
927 no longer possible because the return has to be checked for (*THEN). These
928 two optimizations have therefore been removed. [But see 8.20/11 above.]
929
930 18. If a pattern containing \R was studied, it was assumed that \R always
931 matched two bytes, thus causing the minimum subject length to be
932 incorrectly computed because \R can also match just one byte.
933
934 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
935 was incorrectly computed.
936
937 20. If /S is present twice on a test pattern in pcretest input, it now
938 *disables* studying, thereby overriding the use of -s on the command line
939 (see 14 above). This is necessary for one or two tests to keep the output
940 identical in both cases.
941
942 21. When (*ACCEPT) was used in an assertion that matched an empty string and
943 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
944
945 22. When an atomic group that contained a capturing parenthesis was
946 successfully matched, but the branch in which it appeared failed, the
947 capturing was not being forgotten if a higher numbered group was later
948 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
949 group 1 to "a", when in fact it should be unset. This applied to multi-
950 branched capturing and non-capturing groups, repeated or not, and also to
951 positive assertions (capturing in negative assertions does not happen
952 in PCRE) and also to nested atomic groups.
953
954 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
955 subject after a captured substring, to make it easier to tell which of a
956 number of identical substrings has been captured.
957
958 24. The way atomic groups are processed by pcre_exec() has been changed so that
959 if they are repeated, backtracking one repetition now resets captured
960 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
961 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
962 "aaa". Previously, it would have been "a". As part of this code
963 refactoring, the way recursive calls are handled has also been changed.
964
965 25. If an assertion condition captured any substrings, they were not passed
966 back unless some other capturing happened later. For example, if
967 (?(?=(a))a) was matched against "a", no capturing was returned.
968
969 26. When studying a pattern that contained subroutine calls or assertions,
970 the code for finding the minimum length of a possible match was handling
971 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
972 group 1 called group 2 while simultaneously a separate group 2 called group
973 1). A stack overflow occurred in this case. I have fixed this by limiting
974 the recursion depth to 10.
975
976 27. Updated RunTest.bat in the distribution to the version supplied by Tom
977 Fortmann. This supports explicit test numbers on the command line, and has
978 argument validation and error reporting.
979
980 28. An instance of \X with an unlimited repeat could fail if at any point the
981 first character it looked at was a mark character.
982
983 29. Some minor code refactoring concerning Unicode properties and scripts
984 should reduce the stack requirement of match() slightly.
985
986 30. Added the '=' option to pcretest to check the setting of unused capturing
987 slots at the end of the pattern, which are documented as being -1, but are
988 not included in the return count.
989
990 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
991 compiled something random. Now it gives a compile-time error (as does
992 Perl).
993
994 32. A *MARK encountered during the processing of a positive assertion is now
995 recorded and passed back (compatible with Perl).
996
997 33. If --only-matching or --colour was set on a pcregrep call whose pattern
998 had alternative anchored branches, the search for a second match in a line
999 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1000 matched the line "0102" twice. The same bug affected patterns that started
1001 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1002 twice.
1003
1004 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1005 does, and because of capturing effects, quantifying parenthesized
1006 assertions may at times be useful. Quantifiers are now allowed for
1007 parenthesized assertions.
1008
1009 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1010
1011 36. \g was being checked for fancy things in a character class, when it should
1012 just be a literal "g".
1013
1014 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1015 appearance of a nested POSIX class supersedes an apparent external class.
1016 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1017 unescaped square brackets may also appear as part of class names. For
1018 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1019 more like Perl. (But see 8.20/1 above.)
1020
1021 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1022 was because it thought it was \N{name}, which is not supported).
1023
1024 39. Add minix to OS list not supporting the -S option in pcretest.
1025
1026 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1027 cannot analyze patterns in sufficient detail to catch mutual recursions
1028 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1029 subgroup is called recursively as a subpattern for a second time at the
1030 same position in the subject string. In previous releases this might have
1031 been caught by the recursion limit, or it might have run out of stack.
1032
1033 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1034 happen only once. PCRE was, however incorrectly giving a compile time error
1035 "recursive call could loop indefinitely" because it cannot analyze the
1036 pattern in sufficient detail. The compile time test no longer happens when
1037 PCRE is compiling a conditional subpattern, but actual runaway loops are
1038 now caught at runtime (see 40 above).
1039
1040 42. It seems that Perl allows any characters other than a closing parenthesis
1041 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1042 has been changed to be the same.
1043
1044 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1045 as not to get warnings when autogen.sh is called. Also changed
1046 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1047
1048 44. To help people who use pcregrep to scan files containing exceedingly long
1049 lines, the following changes have been made:
1050
1051 (a) The default value of the buffer size parameter has been increased from
1052 8K to 20K. (The actual buffer used is three times this size.)
1053
1054 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1055 PCRE is built.
1056
1057 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1058 to be set at run time.
1059
1060 (d) Numerical values in pcregrep options can be followed by K or M, for
1061 example --buffer-size=50K.
1062
1063 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1064 given and the return code is set to 2.
1065
1066 45. Add a pointer to the latest mark to the callout data block.
1067
1068 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1069 partial match of an empty string instead of no match. This was specific to
1070 the use of ".".
1071
1072 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1073 complete match instead of a partial match. This bug was dependent on both
1074 the PCRE_UTF8 and PCRE_DOTALL options being set.
1075
1076 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1077 starting byte set, because \b was not being ignored.
1078
1079
1080 Version 8.12 15-Jan-2011
1081 ------------------------
1082
1083 1. Fixed some typos in the markup of the man pages, and wrote a script that
1084 checks for such things as part of the documentation building process.
1085
1086 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1087 --match-limit and --recursion-limit options (added for 8.11). In
1088 particular, this made one of the standard tests fail. (The integer value
1089 went into the wrong half of a long int.)
1090
1091 3. If the --colour option was given to pcregrep with -v (invert match), it
1092 did strange things, either producing crazy output, or crashing. It should,
1093 of course, ignore a request for colour when reporting lines that do not
1094 match.
1095
1096 4. Another pcregrep bug caused similar problems if --colour was specified with
1097 -M (multiline) and the pattern match finished with a line ending.
1098
1099 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1100 matched in multiline mode, the following line was shown as part of the
1101 match. This seems wrong, so I have changed it.
1102
1103 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1104 the check for further matches in the same line (so they could be coloured)
1105 to overrun the end of the current line. If another match was found, it was
1106 incorrectly shown (and then shown again when found in the next line).
1107
1108 7. If pcregrep was compiled under Windows, there was a reference to the
1109 function pcregrep_exit() before it was defined. I am assuming this was
1110 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1111 reported by a user. I've moved the definition above the reference.
1112
1113
1114 Version 8.11 10-Dec-2010
1115 ------------------------
1116
1117 1. (*THEN) was not working properly if there were untried alternatives prior
1118 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1119 backtracked to try for "b" instead of moving to the next alternative branch
1120 at the same level (in this case, to look for "c"). The Perl documentation
1121 is clear that when (*THEN) is backtracked onto, it goes to the "next
1122 alternative in the innermost enclosing group".
1123
1124 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1125 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1126 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1127 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1128 (*THEN).
1129
1130 3. If \s appeared in a character class, it removed the VT character from
1131 the class, even if it had been included by some previous item, for example
1132 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1133 of \s, but is part of the POSIX "space" class.)
1134
1135 4. A partial match never returns an empty string (because you can always
1136 match an empty string at the end of the subject); however the checking for
1137 an empty string was starting at the "start of match" point. This has been
1138 changed to the "earliest inspected character" point, because the returned
1139 data for a partial match starts at this character. This means that, for
1140 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1141 (previously it gave "no match").
1142
1143 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1144 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1145 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1146 has an implication that the given string is incomplete (because a partial
1147 match is preferred over a full match). For this reason, these items now
1148 give a partial match in this situation. [Aside: previously, the one case
1149 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1150 match rather than a full match, which was wrong by the old rules, but is
1151 now correct.]
1152
1153 6. There was a bug in the handling of #-introduced comments, recognized when
1154 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1155 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1156 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1157 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1158 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1159 places in pcre_compile().
1160
1161 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1162 comments when looking ahead for named forward references to subpatterns,
1163 the only newline sequence it recognized was NL. It now handles newlines
1164 according to the set newline convention.
1165
1166 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1167 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1168 cater for a lack of strerror(). These oversights have been fixed.
1169
1170 9. Added --match-limit and --recursion-limit to pcregrep.
1171
1172 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1173
1174 11. When the -o option was used, pcregrep was setting a return code of 1, even
1175 when matches were found, and --line-buffered was not being honoured.
1176
1177 12. Added an optional parentheses number to the -o and --only-matching options
1178 of pcregrep.
1179
1180 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1181 can match an empty string. The code to do it in pcretest and pcredemo
1182 needed fixing:
1183
1184 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1185 only one byte after an empty string match just before CRLF (this case
1186 just got forgotten; "any" and "anycrlf" were OK).
1187
1188 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1189 mode when an empty string match preceded an ASCII character followed by
1190 a non-ASCII character. (The code for advancing by one character rather
1191 than one byte was nonsense.)
1192
1193 (c) The pcredemo.c sample program did not have any code at all to handle
1194 the cases when CRLF is a valid newline sequence.
1195
1196 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1197 as a starting offset was within the subject string. There is now a new
1198 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1199 negative or greater than the length of the string. In order to test this,
1200 pcretest is extended to allow the setting of negative starting offsets.
1201
1202 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1203 starting offset points to the beginning of a UTF-8 character was
1204 unnecessarily clumsy. I tidied it up.
1205
1206 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1207 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1208
1209 17. Nobody had reported that the --include_dir option, which was added in
1210 release 7.7 should have been called --include-dir (hyphen, not underscore)
1211 for compatibility with GNU grep. I have changed it to --include-dir, but
1212 left --include_dir as an undocumented synonym, and the same for
1213 --exclude-dir, though that is not available in GNU grep, at least as of
1214 release 2.5.4.
1215
1216 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1217 characters from a string of bytes) have been redefined so as not to use
1218 loops, in order to improve performance in some environments. At the same
1219 time, I abstracted some of the common code into auxiliary macros to save
1220 repetition (this should not affect the compiled code).
1221
1222 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1223 compile-time error is now given if \c is not followed by an ASCII
1224 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1225 different, and any byte value is allowed.)
1226
1227 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1228 START_OPTIMIZE option, which is now allowed at compile time - but just
1229 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1230 to pcregrep and other applications that have no direct access to PCRE
1231 options. The new /Y option in pcretest sets this option when calling
1232 pcre_compile().
1233
1234 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1235 back references. Groups containing recursive back references were forced to
1236 be atomic by that change, but in the case of named groups, the amount of
1237 memory required was incorrectly computed, leading to "Failed: internal
1238 error: code overflow". This has been fixed.
1239
1240 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1241 pcretest.c, to avoid build problems in some Borland environments.
1242
1243
1244 Version 8.10 25-Jun-2010
1245 ------------------------
1246
1247 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1248 THEN.
1249
1250 2. (*ACCEPT) was not working when inside an atomic group.
1251
1252 3. Inside a character class, \B is treated as a literal by default, but
1253 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1254 causes the error). The code is unchanged, but I tidied the documentation.
1255
1256 4. Inside a character class, PCRE always treated \R and \X as literals,
1257 whereas Perl faults them if its -w option is set. I have changed PCRE so
1258 that it faults them when PCRE_EXTRA is set.
1259
1260 5. Added support for \N, which always matches any character other than
1261 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1262
1263 6. When compiling pcregrep with newer versions of gcc which may have
1264 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1265 declared with attribute warn_unused_result" were given. Just casting the
1266 result to (void) does not stop the warnings; a more elaborate fudge is
1267 needed. I've used a macro to implement this.
1268
1269 7. Minor change to pcretest.c to avoid a compiler warning.
1270
1271 8. Added four artifical Unicode properties to help with an option to make
1272 \s etc use properties (see next item). The new properties are: Xan
1273 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1274
1275 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1276 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1277 this option. Modified pcretest to add /W to test this facility. Added
1278 REG_UCP to make it available via the POSIX interface.
1279
1280 10. Added --line-buffered to pcregrep.
1281
1282 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1283 studied, and the match started with a letter with a code point greater than
1284 127 whose first byte was different to the first byte of the other case of
1285 the letter, the other case of this starting letter was not recognized
1286 (#976).
1287
1288 12. If a pattern that was studied started with a repeated Unicode property
1289 test, for example, \p{Nd}+, there was the theoretical possibility of
1290 setting up an incorrect bitmap of starting bytes, but fortunately it could
1291 not have actually happened in practice until change 8 above was made (it
1292 added property types that matched character-matching opcodes).
1293
1294 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1295 possible starting bytes for non-anchored patterns.
1296
1297 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1298 \R, and also a number of cases that involve Unicode properties, both
1299 explicit and implicit when PCRE_UCP is set.
1300
1301 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1302 input, it could crash or give wrong results if characters with values
1303 greater than 0xc0 were present in the subject string. (Detail: it assumed
1304 UTF-8 input when processing these items.)
1305
1306 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1307 size_t is 64-bit (#991).
1308
1309 17. Added a check for running out of memory when PCRE is compiled with
1310 --disable-stack-for-recursion (#990).
1311
1312 18. If the last data line in a file for pcretest does not have a newline on
1313 the end, a newline was missing in the output.
1314
1315 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1316 less than 128) in its various bitmaps. However, there is a facility for
1317 generating tables according to the current locale when PCRE is compiled. It
1318 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1319 characters, are recognized by isspace() and therefore were getting set in
1320 these tables, and indeed these tables seem to approximate to ISO 8859. This
1321 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1322 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1323 which of course cannot start UTF-8 characters. I have changed the code so
1324 that only real ASCII characters (less than 128) and the correct starting
1325 bytes for UTF-8 encodings are set for characters greater than 127 when in
1326 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1327 altogether.)
1328
1329 20. Added the /T option to pcretest so as to be able to run tests with non-
1330 standard character tables, thus making it possible to include the tests
1331 used for 19 above in the standard set of tests.
1332
1333 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1334 reference to a subpattern the other side of a comment that contains an
1335 opening parenthesis caused either an internal compiling error, or a
1336 reference to the wrong subpattern.
1337
1338
1339 Version 8.02 19-Mar-2010
1340 ------------------------
1341
1342 1. The Unicode data tables have been updated to Unicode 5.2.0.
1343
1344 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1345 configured.
1346
1347 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1348 original author of that file, following a query about its status.
1349
1350 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1351 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1352
1353 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1354 quantifier applied to a forward-referencing subroutine call, could compile
1355 incorrect code or give the error "internal error: previously-checked
1356 referenced subpattern not found".
1357
1358 6. Both MS Visual Studio and Symbian OS have problems with initializing
1359 variables to point to external functions. For these systems, therefore,
1360 pcre_malloc etc. are now initialized to local functions that call the
1361 relevant global functions.
1362
1363 7. There were two entries missing in the vectors called coptable and poptable
1364 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1365 I've fixed the data, and added a kludgy way of testing at compile time that
1366 the lengths are correct (equal to the number of opcodes).
1367
1368 8. Following on from 7, I added a similar kludge to check the length of the
1369 eint vector in pcreposix.c.
1370
1371 9. Error texts for pcre_compile() are held as one long string to avoid too
1372 much relocation at load time. To find a text, the string is searched,
1373 counting zeros. There was no check for running off the end of the string,
1374 which could happen if a new error number was added without updating the
1375 string.
1376
1377 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1378
1379 11. \K was not working if it appeared in an atomic group or in a group that
1380 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1381 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1382 the assertion is positive, but not if it is negative.
1383
1384 12. Change 11 fortuitously reduced the size of the stack frame used in the
1385 "match()" function of pcre_exec.c by one pointer. Forthcoming
1386 implementation of support for (*MARK) will need an extra pointer on the
1387 stack; I have reserved it now, so that the stack frame size does not
1388 decrease.
1389
1390 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1391 item in branch that calls a recursion is a subroutine call - as in the
1392 second branch in the above example - was incorrectly given the compile-
1393 time error "recursive call could loop indefinitely" because pcre_compile()
1394 was not correctly checking the subroutine for matching a non-empty string.
1395
1396 14. The checks for overrunning compiling workspace could trigger after an
1397 overrun had occurred. This is a "should never occur" error, but it can be
1398 triggered by pathological patterns such as hundreds of nested parentheses.
1399 The checks now trigger 100 bytes before the end of the workspace.
1400
1401 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1402
1403
1404 Version 8.01 19-Jan-2010
1405 ------------------------
1406
1407 1. If a pattern contained a conditional subpattern with only one branch (in
1408 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1409 computed the wrong minimum data length (which is of course zero for such
1410 subpatterns). This could cause incorrect "no match" results.
1411
1412 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1413 the pattern is reset in the first branch, pcre_compile() failed with
1414 "internal error: code overflow at offset...". This happened only when
1415 the reset was to the original external option setting. (An optimization
1416 abstracts leading options settings into an external setting, which was the
1417 cause of this.)
1418
1419 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1420 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1421 assertion pattern did not match (meaning that the assertion was true), it
1422 was incorrectly treated as false if the SKIP had been reached during the
1423 matching. This also applied to assertions used as conditions.
1424
1425 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1426 assertion subpattern, including such a pattern used as a condition,
1427 unpredictable results occurred, instead of the error return
1428 PCRE_ERROR_DFA_UITEM.
1429
1430 5. The C++ GlobalReplace function was not working like Perl for the special
1431 situation when an empty string is matched. It now does the fancy magic
1432 stuff that is necessary.
1433
1434 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1435 removed. (These were left over from very, very early versions of PCRE.)
1436
1437 7. Some cosmetic changes to the code to make life easier when compiling it
1438 as part of something else:
1439
1440 (a) Change DEBUG to PCRE_DEBUG.
1441
1442 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1443 called "current" as "current_branch", to prevent a collision with the
1444 Linux macro when compiled as a kernel module.
1445
1446 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1447 prevent a collision with the Linux macro when compiled as a kernel
1448 module.
1449
1450 8. In pcre_compile() there are some checks for integer overflows that used to
1451 cast potentially large values to (double). This has been changed to that
1452 when building, a check for int64_t is made, and if it is found, it is used
1453 instead, thus avoiding the use of floating point arithmetic. (There is no
1454 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1455 double.
1456
1457 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1458 2005 (difference between two addresses compared to an unsigned value).
1459
1460 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1461 custom one, because of the following reported problem in Windows:
1462
1463 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1464 under Win32.
1465 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1466 therefore missing the function definition.
1467 - The compiler thus generates a "C" signature for the test function.
1468 - The linker fails to find the "C" function.
1469 - PCRE fails to configure if asked to do so against libbz2.
1470
1471 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1472 messages were output:
1473
1474 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1475 rerunning libtoolize, to keep the correct libtool macros in-tree.
1476 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1477
1478 I have done both of these things.
1479
1480 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1481 most of the time, it *can* run out if it is given a pattern that contains a
1482 runaway infinite recursion. I updated the discussion in the pcrestack man
1483 page.
1484
1485 13. Now that we have gone to the x.xx style of version numbers, the minor
1486 version may start with zero. Using 08 or 09 is a bad idea because users
1487 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1488 interpreted as invalid octal numbers. I've updated the previous comment in
1489 configure.ac, and also added a check that gives an error if 08 or 09 are
1490 used.
1491
1492 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1493 causing partial matching to fail when the end of the subject matched \W
1494 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1495
1496 15. There were some discrepancies between the declarations in pcre_internal.h
1497 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1498 their definitions. The declarations used "const uschar *" and the
1499 definitions used USPTR. Even though USPTR is normally defined as "const
1500 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1501 reported that: "This difference in casting confuses some C++ compilers, for
1502 example, SunCC recognizes above declarations as different functions and
1503 generates broken code for hbpcre." I have changed the declarations to use
1504 USPTR.
1505
1506 16. GNU libtool is named differently on some systems. The autogen.sh script now
1507 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1508 (FreeBSD).
1509
1510 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1511 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1512 comment: "Figure out how to create a longlong from a string: strtoll and
1513 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1514 instance, but it only takes 2 args instead of 3!"
1515
1516 18. A subtle bug concerned with back references has been fixed by a change of
1517 specification, with a corresponding code fix. A pattern such as
1518 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1519 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1520 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1521 same bug. Such groups have to be quantified to be useful, or contained
1522 inside another quantified group. (If there's no repetition, the reference
1523 can never match.) The problem arises because, having left the group and
1524 moved on to the rest of the pattern, a later failure that backtracks into
1525 the group uses the captured value from the final iteration of the group
1526 rather than the correct earlier one. I have fixed this in PCRE by forcing
1527 any group that contains a reference to itself to be an atomic group; that
1528 is, there cannot be any backtracking into it once it has completed. This is
1529 similar to recursive and subroutine calls.
1530
1531
1532 Version 8.00 19-Oct-09
1533 ----------------------
1534
1535 1. The table for translating pcre_compile() error codes into POSIX error codes
1536 was out-of-date, and there was no check on the pcre_compile() error code
1537 being within the table. This could lead to an OK return being given in
1538 error.
1539
1540 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1541 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1542 in a Windows environment.
1543
1544 3. The pcregrep --count option prints the count for each file even when it is
1545 zero, as does GNU grep. However, pcregrep was also printing all files when
1546 --files-with-matches was added. Now, when both options are given, it prints
1547 counts only for those files that have at least one match. (GNU grep just
1548 prints the file name in this circumstance, but including the count seems
1549 more useful - otherwise, why use --count?) Also ensured that the
1550 combination -clh just lists non-zero counts, with no names.
1551
1552 4. The long form of the pcregrep -F option was incorrectly implemented as
1553 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1554 but it seems right to fix it, and I didn't think it was worth preserving
1555 the old behaviour.
1556
1557 5. The command line items --regex=pattern and --regexp=pattern were not
1558 recognized by pcregrep, which required --regex pattern or --regexp pattern
1559 (with a space rather than an '='). The man page documented the '=' forms,
1560 which are compatible with GNU grep; these now work.
1561
1562 6. No libpcreposix.pc file was created for pkg-config; there was just
1563 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1564
1565 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1566 when UCP support is not needed, by modifying the Python script that
1567 generates it from Unicode data files. This should not matter if the module
1568 is correctly used as a library, but I received one complaint about 50K of
1569 unwanted data. My guess is that the person linked everything into his
1570 program rather than using a library. Anyway, it does no harm.
1571
1572 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1573 was a minimum greater than 1 for a wide character in a possessive
1574 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1575 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1576 character. Chaos in the form of incorrect output or a compiling loop could
1577 result.
1578
1579 9. The restrictions on what a pattern can contain when partial matching is
1580 requested for pcre_exec() have been removed. All patterns can now be
1581 partially matched by this function. In addition, if there are at least two
1582 slots in the offset vector, the offset of the earliest inspected character
1583 for the match and the offset of the end of the subject are set in them when
1584 PCRE_ERROR_PARTIAL is returned.
1585
1586 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1587 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1588 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1589 and may be more useful for multi-segment matching.
1590
1591 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1592 used to be given if ever the end of the subject was reached; now it is
1593 given only if matching could not proceed because another character was
1594 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1595 string "Z", which now yields "no match" instead of "partial match". In the
1596 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1597 final character ended with (*FAIL).
1598
1599 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1600 if the pattern had a "must contain" character that was already found in the
1601 earlier partial match, unless partial matching was again requested. For
1602 example, with the pattern /dog.(body)?/, the "must contain" character is
1603 "g". If the first part-match was for the string "dog", restarting with
1604 "sbody" failed. This bug has been fixed.
1605
1606 13. The string returned by pcre_dfa_exec() after a partial match has been
1607 changed so that it starts at the first inspected character rather than the
1608 first character of the match. This makes a difference only if the pattern
1609 starts with a lookbehind assertion or \b or \B (\K is not supported by
1610 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1611 matching functions compatible, and I think it's the right thing to do.
1612
1613 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1614 so that the demonstration program is easily available in environments where
1615 PCRE has not been installed from source.
1616
1617 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1618 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1619 library.
1620
1621 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1622 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1623 is not the first non-POSIX option to be added. Clearly some people find
1624 these options useful.
1625
1626 17. If a caller to the POSIX matching function regexec() passes a non-zero
1627 value for nmatch with a NULL value for pmatch, the value of
1628 nmatch is forced to zero.
1629
1630 18. RunGrepTest did not have a test for the availability of the -u option of
1631 the diff command, as RunTest does. It now checks in the same way as
1632 RunTest, and also checks for the -b option.
1633
1634 19. If an odd number of negated classes containing just a single character
1635 interposed, within parentheses, between a forward reference to a named
1636 subpattern and the definition of the subpattern, compilation crashed with
1637 an internal error, complaining that it could not find the referenced
1638 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1639 [The bug was that it was starting one character too far in when skipping
1640 over the character class, thus treating the ] as data rather than
1641 terminating the class. This meant it could skip too much.]
1642
1643 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1644 /g option in pcretest when the pattern contains \K, which makes it possible
1645 to have an empty string match not at the start, even when the pattern is
1646 anchored. Updated pcretest and pcredemo to use this option.
1647
1648 21. If the maximum number of capturing subpatterns in a recursion was greater
1649 than the maximum at the outer level, the higher number was returned, but
1650 with unset values at the outer level. The correct (outer level) value is
1651 now given.
1652
1653 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1654 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1655 make it do so. The string so far is captured, making this feature
1656 compatible with Perl.
1657
1658 23. The tests have been re-organized, adding tests 11 and 12, to make it
1659 possible to check the Perl 5.10 features against Perl 5.10.
1660
1661 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1662 pattern matches a fixed length string. PCRE did not allow this; now it
1663 does. Neither allows recursion.
1664
1665 25. I finally figured out how to implement a request to provide the minimum
1666 length of subject string that was needed in order to match a given pattern.
1667 (It was back references and recursion that I had previously got hung up
1668 on.) This code has now been added to pcre_study(); it finds a lower bound
1669 to the length of subject needed. It is not necessarily the greatest lower
1670 bound, but using it to avoid searching strings that are too short does give
1671 some useful speed-ups. The value is available to calling programs via
1672 pcre_fullinfo().
1673
1674 26. While implementing 25, I discovered to my embarrassment that pcretest had
1675 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1676 study optimizations had never been tested with that matching function.
1677 Oops. What is worse, even when it was passed study data, there was a bug in
1678 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1679 were also very few tests of studied patterns with pcre_dfa_exec().
1680
1681 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1682 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1683 on the other side of the coin, they are no longer allowed to have different
1684 names, because these cannot be distinguished in PCRE, and this has caused
1685 confusion. (This is a difference from Perl.)
1686
1687 28. When duplicate subpattern names are present (necessarily with different
1688 numbers, as required by 27 above), and a test is made by name in a
1689 conditional pattern, either for a subpattern having been matched, or for
1690 recursion in such a pattern, all the associated numbered subpatterns are
1691 tested, and the overall condition is true if the condition is true for any
1692 one of them. This is the way Perl works, and is also more like the way
1693 testing by number works.
1694
1695
1696 Version 7.9 11-Apr-09
1697 ---------------------
1698
1699 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1700 (pcretest), all targets were linked against these libraries. This included
1701 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1702 libraries. This caused unwanted dependencies to be created. This problem
1703 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1704 pcretest is linked with readline.
1705
1706 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1707 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1708 moved outside it again, because FALSE and TRUE are already defined in AIX,
1709 but BOOL is not.
1710
1711 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1712 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1713
1714 4. The pcregrep documentation said spaces were inserted as well as colons (or
1715 hyphens) following file names and line numbers when outputting matching
1716 lines. This is not true; no spaces are inserted. I have also clarified the
1717 wording for the --colour (or --color) option.
1718
1719 5. In pcregrep, when --colour was used with -o, the list of matching strings
1720 was not coloured; this is different to GNU grep, so I have changed it to be
1721 the same.
1722
1723 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1724 each matching line was coloured. Now it goes on to look for further matches
1725 of any of the test patterns, which is the same behaviour as GNU grep.
1726
1727 7. A pattern that could match an empty string could cause pcregrep to loop; it
1728 doesn't make sense to accept an empty string match in pcregrep, so I have
1729 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1730 seems to be how GNU grep behaves. [But see later change 40 for release
1731 8.33.]
1732
1733 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1734 start or after a newline", because the conditional assertion was not being
1735 correctly handled. The rule now is that both the assertion and what follows
1736 in the first alternative must satisfy the test.
1737
1738 9. If auto-callout was enabled in a pattern with a conditional group whose
1739 condition was an assertion, PCRE could crash during matching, both with
1740 pcre_exec() and pcre_dfa_exec().
1741
1742 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1743 used for matching.
1744
1745 11. Unicode property support in character classes was not working for
1746 characters (bytes) greater than 127 when not in UTF-8 mode.
1747
1748 12. Added the -M command line option to pcretest.
1749
1750 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1751
1752 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1753
1754 16. Added comments and documentation about mis-use of no_arg in the C++
1755 wrapper.
1756
1757 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1758 from Martin Jerabek that uses macro names for all relevant character and
1759 string constants.
1760
1761 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1762 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1763 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1764 these, but not everybody uses configure.
1765
1766 19. A conditional group that had only one branch was not being correctly
1767 recognized as an item that could match an empty string. This meant that an
1768 enclosing group might also not be so recognized, causing infinite looping
1769 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1770 with the subject "ab", where knowledge that the repeated group can match
1771 nothing is needed in order to break the loop.
1772
1773 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1774 exec(), but without supplying a callout function, matching went wrong.
1775
1776 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1777 leak if the size of the offset vector was greater than 30. When the vector
1778 is smaller, the saved offsets during recursion go onto a local stack
1779 vector, but for larger vectors malloc() is used. It was failing to free
1780 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1781 error, in fact).
1782
1783 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1784 heapframe that is used only when UTF-8 support is enabled. This caused no
1785 problem, but was untidy.
1786
1787 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1788 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1789 included within another project.
1790
1791 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1792 slightly modified by me:
1793
1794 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1795 not building pcregrep.
1796
1797 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1798 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1799
1800 25. Forward references, both numeric and by name, in patterns that made use of
1801 duplicate group numbers, could behave incorrectly or give incorrect errors,
1802 because when scanning forward to find the reference group, PCRE was not
1803 taking into account the duplicate group numbers. A pattern such as
1804 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1805
1806 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1807 the feature of a custom pointer more persuasive (as requested by a user).
1808
1809 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1810 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1811 pre-defined.
1812
1813 28. Added support for (*UTF8) at the start of a pattern.
1814
1815 29. Arrange for flags added by the "release type" setting in CMake to be shown
1816 in the configuration summary.
1817
1818
1819 Version 7.8 05-Sep-08
1820 ---------------------
1821
1822 1. Replaced UCP searching code with optimized version as implemented for Ad
1823 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1824 stage table and inline lookup instead of a function, giving speed ups of 2
1825 to 5 times on some simple patterns that I tested. Permission was given to
1826 distribute the MultiStage2.py script that generates the tables (it's not in
1827 the tarball, but is in the Subversion repository).
1828
1829 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1830 scripts.
1831
1832 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1833 a group with a zero qualifier. The result of the study could be incorrect,
1834 or the function might crash, depending on the pattern.
1835
1836 4. Caseless matching was not working for non-ASCII characters in back
1837 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1838 It now works when Unicode Property Support is available.
1839
1840 5. In pcretest, an escape such as \x{de} in the data was always generating
1841 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1842 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1843 truncation.
1844
1845 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1846
1847 7. Added two (int) casts to pcregrep when printing the difference of two
1848 pointers, in case they are 64-bit values.
1849
1850 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1851 test 2 if it fails.
1852
1853 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1854 and a #define of that name to empty if it is not externally set. This is to
1855 allow users of MSVC to set it if necessary.
1856
1857 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1858 the convenience functions in the pcre_get.c source file.
1859
1860 11. An option change at the start of a pattern that had top-level alternatives
1861 could cause overwriting and/or a crash. This command provoked a crash in
1862 some environments:
1863
1864 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1865
1866 This potential security problem was recorded as CVE-2008-2371.
1867
1868 12. For a pattern where the match had to start at the beginning or immediately
1869 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1870 pcre_dfa_exec() could read past the end of the passed subject if there was
1871 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1872 pcretest so that it places the subject at the end of its malloc-ed buffer.
1873
1874 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1875 exec() might read past the end of the data buffer in UTF-8 mode.
1876
1877 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1878 the data contained the byte 0x85 as part of a UTF-8 character within its
1879 first line. This applied both to normal and DFA matching.
1880
1881 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1882 /^[^d]*?$/8 failed to match "abc".
1883
1884 16. Added a missing copyright notice to pcrecpp_internal.h.
1885
1886 17. Make it more clear in the documentation that values returned from
1887 pcre_exec() in ovector are byte offsets, not character counts.
1888
1889 18. Tidied a few places to stop certain compilers from issuing warnings.
1890
1891 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1892 supplied by Stefan Weber. I made a further small update for 7.8 because
1893 there is a change of source arrangements: the pcre_searchfuncs.c module is
1894 replaced by pcre_ucd.c.
1895
1896
1897 Version 7.7 07-May-08
1898 ---------------------
1899
1900 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1901 a string to a long long, pretend we don't even have a long long." This is
1902 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1903
1904 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1905 pre-7.6 versions, which defined a global no_arg variable instead of putting
1906 it in the RE class. (See also #8 below.)
1907
1908 3. Remove a line of dead code, identified by coverity and reported by Nuno
1909 Lopes.
1910
1911 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1912
1913 (1) The include/exclude patterns were being applied to the whole pathnames
1914 of files, instead of just to the final components.
1915
1916 (2) If there was more than one level of directory, the subdirectories were
1917 skipped unless they satisfied the include/exclude conditions. This is
1918 inconsistent with GNU grep (and could even be seen as contrary to the
1919 pcregrep specification - which I improved to make it absolutely clear).
1920 The action now is always to scan all levels of directory, and just
1921 apply the include/exclude patterns to regular files.
1922
1923 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1924 --exclude_dir in the tests to avoid scanning .svn directories.
1925
1926 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1927 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1928 doesn't support NULs in patterns.
1929
1930 7. Added some missing "const"s to declarations of static tables in
1931 pcre_compile.c and pcre_dfa_exec.c.
1932
1933 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1934 caused by fix #2 above. (Subsequently also a second patch to fix the
1935 first patch. And a third patch - this was a messy problem.)
1936
1937 9. Applied Craig's patch to remove the use of push_back().
1938
1939 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1940 matching function regexec().
1941
1942 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1943 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1944 references. PCRE supports relative numbers with this syntax (I don't think
1945 Oniguruma does).
1946
1947 12. Previously, a group with a zero repeat such as (...){0} was completely
1948 omitted from the compiled regex. However, this means that if the group
1949 was called as a subroutine from elsewhere in the pattern, things went wrong
1950 (an internal error was given). Such groups are now left in the compiled
1951 pattern, with a new opcode that causes them to be skipped at execution
1952 time.
1953
1954 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1955 to the way PCRE behaves:
1956
1957 (a) A lone ] character is dis-allowed (Perl treats it as data).
1958
1959 (b) A back reference to an unmatched subpattern matches an empty string
1960 (Perl fails the current match path).
1961
1962 (c) A data ] in a character class must be notated as \] because if the
1963 first data character in a class is ], it defines an empty class. (In
1964 Perl it is not possible to have an empty class.) The empty class []
1965 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1966 The negative empty class [^] matches any one character, independently
1967 of the DOTALL setting.
1968
1969 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1970 non-existent subpattern following a character class starting with ']' and
1971 containing () gave an internal compiling error instead of "reference to
1972 non-existent subpattern". Fortunately, when the pattern did exist, the
1973 compiled code was correct. (When scanning forwards to check for the
1974 existence of the subpattern, it was treating the data ']' as terminating
1975 the class, so got the count wrong. When actually compiling, the reference
1976 was subsequently set up correctly.)
1977
1978 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1979 it was being rejected as not supported by pcre_dfa_exec(), even though
1980 other assertions are supported. I have made pcre_dfa_exec() support
1981 (*FAIL).
1982
1983 16. The implementation of 13c above involved the invention of a new opcode,
1984 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1985 cannot be changed at match time, I realized I could make a small
1986 improvement to matching performance by compiling OP_ALLANY instead of
1987 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1988 on the OP_ANY path.
1989
1990 17. Compiling pcretest on Windows with readline support failed without the
1991 following two fixes: (1) Make the unistd.h include conditional on
1992 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1993
1994 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1995 ncurses library to be included for pcretest when ReadLine support is
1996 requested, but also to allow for it to be overridden. This patch came from
1997 Daniel Bergström.
1998
1999 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2000 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2001 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2002 spotting this.
2003
2004
2005 Version 7.6 28-Jan-08
2006 ---------------------
2007
2008 1. A character class containing a very large number of characters with
2009 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2010 overflow.
2011
2012 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2013 HAVE_LONG_LONG is not defined.
2014
2015 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2016 bring them up to date and include new features. This patch includes:
2017
2018 - Fixed PH's badly added libz and libbz2 support.
2019 - Fixed a problem with static linking.
2020 - Added pcredemo. [But later removed - see 7 below.]
2021 - Fixed dftables problem and added an option.
2022 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2023 HAVE_LONG_LONG.
2024 - Added readline support for pcretest.
2025 - Added an listing of the option settings after cmake has run.
2026
2027 4. A user submitted a patch to Makefile that makes it easy to create
2028 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2029 Makefile.am that cause it to include this special target, without
2030 affecting anything else. Note that the same mingw target plus all
2031 the other distribution libraries and programs are now supported
2032 when configuring with CMake (see 6 below) instead of with
2033 Configure/Make.
2034
2035 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2036 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2037 exported in the Windows port". It has not yet been confirmed that the patch
2038 solves the problem, but it does no harm.
2039
2040 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2041 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2042 with CMake, and also correct the comment about stack recursion.
2043
2044 7. Remove the automatic building of pcredemo from the ./configure system and
2045 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2046 of a program that users should build themselves after PCRE is installed, so
2047 building it automatically is not really right. What is more, it gave
2048 trouble in some build environments.
2049
2050 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2051
2052
2053 Version 7.5 10-Jan-08
2054 ---------------------
2055
2056 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2057 values in parens when parsing an RE using the C++ wrapper."
2058
2059 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2060 Characters greater than 255 were excluded from the class instead of being
2061 included.
2062
2063 3. The same bug as (2) above applied to negated POSIX classes such as
2064 [:^space:].
2065
2066 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2067 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2068 I have changed it.
2069
2070 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2071 first named subpattern) and a construct such as (?&a) would reference the
2072 first named subpattern whose name started with "a" (in other words, the
2073 length check was missing). Both these problems are fixed. "Subpattern name
2074 expected" is now given for (?&) (a zero-length name), and this patch also
2075 makes it give the same error for \k'' (previously it complained that that
2076 was a reference to a non-existent subpattern).
2077
2078 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2079 this is right because (?- can be followed by option settings as well as by
2080 digits. I have, however, made the messages clearer.
2081
2082 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2083 than the number used in the conditional) now cause a compile-time error.
2084 This is actually not compatible with Perl, which accepts such patterns, but
2085 treats the conditional as always being FALSE (as PCRE used to), but it
2086 seems to me that giving a diagnostic is better.
2087
2088 8. Change "alphameric" to the more common word "alphanumeric" in comments
2089 and messages.
2090
2091 9. Fix two occurrences of "backslash" in comments that should have been
2092 "backspace".
2093
2094 10. Remove two redundant lines of code that can never be obeyed (their function
2095 was moved elsewhere).
2096
2097 11. The program that makes PCRE's Unicode character property table had a bug
2098 which caused it to generate incorrect table entries for sequences of
2099 characters that have the same character type, but are in different scripts.
2100 It amalgamated them into a single range, with the script of the first of
2101 them. In other words, some characters were in the wrong script. There were
2102 thirteen such cases, affecting characters in the following ranges:
2103
2104 U+002b0 - U+002c1
2105 U+0060c - U+0060d
2106 U+0061e - U+00612
2107 U+0064b - U+0065e
2108 U+0074d - U+0076d
2109 U+01800 - U+01805
2110 U+01d00 - U+01d77
2111 U+01d9b - U+01dbf
2112 U+0200b - U+0200f
2113 U+030fc - U+030fe
2114 U+03260 - U+0327f
2115 U+0fb46 - U+0fbb1
2116 U+10450 - U+1049d
2117
2118 12. The -o option (show only the matching part of a line) for pcregrep was not
2119 compatible with GNU grep in that, if there was more than one match in a
2120 line, it showed only the first of them. It now behaves in the same way as
2121 GNU grep.
2122
2123 13. If the -o and -v options were combined for pcregrep, it printed a blank
2124 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2125 does the same. The return code can be used to tell if there were any
2126 non-matching lines.
2127
2128 14. Added --file-offsets and --line-offsets to pcregrep.
2129
2130 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2131 infinitely looping recursion. The bug was that positive lookaheads were not
2132 being skipped when checking for a possible empty match (negative lookaheads
2133 and both kinds of lookbehind were skipped).
2134
2135 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2136 inclusion of <windows.h> to before rather than after the definition of
2137 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2138
2139 17. Specifying a possessive quantifier with a specific limit for a Unicode
2140 character property caused pcre_compile() to compile bad code, which led at
2141 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2142 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2143 caused the error; without that there was no problem.
2144
2145 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2146
2147 19. Added --enable-pcretest-libreadline.
2148
2149 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2150 RE::GlobalReplace(). As a result, the number of replacements returned was
2151 double what it should be. I removed one of the increments, but Craig sent a
2152 later patch that removed the other one (the right fix) and added unit tests
2153 that check the return values (which was not done before).
2154
2155 21. Several CMake things:
2156
2157 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2158 the names libpcre and libpcreposix, not just pcre and pcreposix.
2159
2160 (2) The above change means that pcretest and pcregrep are now correctly
2161 linked with the newly-built libraries, not previously installed ones.
2162
2163 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2164
2165 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2166 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2167 UTF-8 newline character). The key issue is that the pattern starts .*;
2168 this means that the match must be either at the beginning, or after a
2169 newline. The bug was in the code for advancing after a failed match and
2170 checking that the new position followed a newline. It was not taking
2171 account of UTF-8 characters correctly.
2172
2173 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2174 character classes. PCRE was not treating the sequence [:...:] as a
2175 character class unless the ... were all letters. Perl, however, seems to
2176 allow any characters between [: and :], though of course it rejects as
2177 unknown any "names" that contain non-letters, because all the known class
2178 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2179 for example, whereas PCRE did not - it did not recognize a POSIX character
2180 class. This seemed a bit dangerous, so the code has been changed to be
2181 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2182 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2183 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2184 Perl does, and where it didn't before.
2185
2186 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2187 Windows environments %n is disabled by default.
2188
2189
2190 Version 7.4 21-Sep-07
2191 ---------------------
2192
2193 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2194 means that a class such as [\s] counted as "explicit reference to CR or
2195 LF". That isn't really right - the whole point of the change was to try to
2196 help when there was an actual mention of one of the two characters. So now
2197 the change happens only if \r or \n (or a literal CR or LF) character is
2198 encountered.
2199
2200 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2201 of both had grown to the point where there were only 3 bits left.
2202 Fortunately, there was spare space in the data structure, and so I have
2203 moved the internal flags into a new 16-bit field to free up more option
2204 bits.
2205
2206 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2207 but did not set the internal JCHANGED flag - either of these is enough to
2208 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2209 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2210 start sets both bits.
2211
2212 4. Added options (at build time, compile time, exec time) to change \R from
2213 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2214
2215 5. doc/pcresyntax.html was missing from the distribution.
2216
2217 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2218 compatibility, even though it is no longer used.
2219
2220 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2221 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2222 windows.h file is present (where different names are used). [This was
2223 reversed later after testing - see 16 below.]
2224
2225 8. Changed all #include <config.h> to #include "config.h". There were also
2226 some further <pcre.h> cases that I changed to "pcre.h".
2227
2228 9. When pcregrep was used with the --colour option, it missed the line ending
2229 sequence off the lines that it output.
2230
2231 10. It was pointed out to me that arrays of string pointers cause lots of
2232 relocations when a shared library is dynamically loaded. A technique of
2233 using a single long string with a table of offsets can drastically reduce
2234 these. I have refactored PCRE in four places to do this. The result is
2235 dramatic:
2236
2237 Originally: 290
2238 After changing UCP table: 187
2239 After changing error message table: 43
2240 After changing table of "verbs" 36
2241 After changing table of Posix names 22
2242
2243 Thanks to the folks working on Gregex for glib for this insight.
2244
2245 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2246 unicode-properties was also set.
2247
2248 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2249
2250 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2251 checked only for CRLF.
2252
2253 14. Added casts to pcretest.c to avoid compiler warnings.
2254
2255 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2256
2257 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2258 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2259 entirely. This removes changes made in 7 above.
2260
2261 17. The CMake files have been updated, and there is now more information about
2262 building with CMake in the NON-UNIX-USE document.
2263
2264
2265 Version 7.3 28-Aug-07
2266 ---------------------
2267
2268 1. In the rejigging of the build system that eventually resulted in 7.1, the
2269 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2270 brackets there is not right, since it causes compilers to look for an
2271 installed pcre.h, not the version that is in the source that is being
2272 compiled (which of course may be different). I have changed it back to:
2273
2274 #include "pcre.h"
2275
2276 I have a vague recollection that the change was concerned with compiling in
2277 different directories, but in the new build system, that is taken care of
2278 by the VPATH setting the Makefile.
2279
2280 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2281 when the subject happened to end in the byte 0x85 (e.g. if the last
2282 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2283 characters but of course it shouldn't be taken as a newline when it is part
2284 of another character. The bug was that, for an unlimited repeat of . in
2285 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2286 characters when looking for a newline.
2287
2288 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2289
2290 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2291 in debug output.
2292
2293 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2294 long printing in the pcrecpp unittest when running under MinGW.
2295
2296 6. ESC_K was left out of the EBCDIC table.
2297
2298 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2299 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2300 limit also applies to "virtual nesting" when a pattern is recursive, and in
2301 this case 1000 isn't so big. I have been able to remove this limit at the
2302 expense of backing off one optimization in certain circumstances. Normally,
2303 when pcre_exec() would call its internal match() function recursively and
2304 immediately return the result unconditionally, it uses a "tail recursion"
2305 feature to save stack. However, when a subpattern that can match an empty
2306 string has an unlimited repetition quantifier, it no longer makes this
2307 optimization. That gives it a stack frame in which to save the data for
2308 checking that an empty string has been matched. Previously this was taken
2309 from the 1000-entry workspace that had been reserved. So now there is no
2310 explicit limit, but more stack is used.
2311
2312 8. Applied Daniel's patches to solve problems with the import/export magic
2313 syntax that is required for Windows, and which was going wrong for the
2314 pcreposix and pcrecpp parts of the library. These were overlooked when this
2315 problem was solved for the main library.
2316
2317 9. There were some crude static tests to avoid integer overflow when computing
2318 the size of patterns that contain repeated groups with explicit upper
2319 limits. As the maximum quantifier is 65535, the maximum group length was
2320 set at 30,000 so that the product of these two numbers did not overflow a
2321 32-bit integer. However, it turns out that people want to use groups that
2322 are longer than 30,000 bytes (though not repeat them that many times).
2323 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2324 made it possible to implement the integer overflow checks in a much more
2325 dynamic way, which I have now done. The artificial limitation on group
2326 length has been removed - we now have only the limit on the total length of
2327 the compiled pattern, which depends on the LINK_SIZE setting.
2328
2329 10. Fixed a bug in the documentation for get/copy named substring when
2330 duplicate names are permitted. If none of the named substrings are set, the
2331 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2332 empty string.
2333
2334 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2335 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2336 because the ] is interpreted as the first data character and the
2337 terminating ] is not found. PCRE has been made compatible with Perl in this
2338 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2339 cause memory overwriting.
2340
2341 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2342 string has been matched (to stop an infinite loop). It was not recognizing
2343 a conditional subpattern that could match an empty string if that
2344 subpattern was within another subpattern. For example, it looped when
2345 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2346 condition was not nested. This bug has been fixed.
2347
2348 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2349 past the start of the subject in the presence of bytes with the top bit
2350 set, for example "\x8aBCD".
2351
2352 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2353 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2354
2355 14. Optimized (?!) to (*FAIL).
2356
2357 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2358 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2359 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2360 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2361 does: it's just the validity check that is more restrictive.
2362
2363 16. Inserted checks for integer overflows during escape sequence (backslash)
2364 processing, and also fixed erroneous offset values for syntax errors during
2365 backslash processing.
2366
2367 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2368 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2369
2370 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2371 caused an overrun.
2372
2373 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2374 something other than just ASCII characters) inside a group that had an
2375 unlimited repeat caused a loop at compile time (while checking to see
2376 whether the group could match an empty string).
2377
2378 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2379 [\P{Any}] did so. (Error in the code for printing property names.)
2380
2381 21. An orphan \E inside a character class could cause a crash.
2382
2383 22. A repeated capturing bracket such as (A)? could cause a wild memory
2384 reference during compilation.
2385
2386 23. There are several functions in pcre_compile() that scan along a compiled
2387 expression for various reasons (e.g. to see if it's fixed length for look
2388 behind). There were bugs in these functions when a repeated \p or \P was
2389 present in the pattern. These operators have additional parameters compared
2390 with \d, etc, and these were not being taken into account when moving along
2391 the compiled data. Specifically:
2392
2393 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2394 length.
2395
2396 (b) An item such as \pL+ within a repeated group could cause crashes or
2397 loops.
2398
2399 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2400 "reference to non-existent subpattern" error.
2401
2402 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2403
2404 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2405 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2406
2407 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2408
2409 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2410 character were causing crashes (broken optimization).
2411
2412 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2413 \p or \P) caused a compile-time loop.
2414
2415 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2416 break. For example, the unstudied pattern [\r\n]A does not match the string
2417 "\r\nA" because change 7.0/46 below moves the current point on by two
2418 characters after failing to match at the start. However, the pattern \nA
2419 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2420 the same is true. There doesn't seem any very clean way out of this, but
2421 what I have chosen to do makes the common cases work: PCRE now takes note
2422 of whether there can be an explicit match for \r or \n anywhere in the
2423 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2424 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2425 pattern has explicit CR or LF references.
2426
2427 29. Added (*CR) etc for changing newline setting at start of pattern.
2428
2429
2430 Version 7.2 19-Jun-07
2431 ---------------------
2432
2433 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2434 which is apparently normally available under Windows.
2435
2436 2. Re-jig the pcregrep tests with different newline settings in an attempt
2437 to make them independent of the local environment's newline setting.
2438
2439 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2440
2441 4. Some of the "internals" tests were previously cut out when the link size
2442 was not 2, because the output contained actual offsets. The recent new
2443 "Z" feature of pcretest means that these can be cut out, making the tests
2444 usable with all link sizes.
2445
2446 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2447 stack recursion. This gives a massive performance boost under BSD, but just
2448 a small improvement under Linux. However, it saves one field in the frame
2449 in all cases.
2450
2451 6. Added more features from the forthcoming Perl 5.10:
2452
2453 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2454 recursion call. It refers to the nth most recently opened parentheses.
2455
2456 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2457 to be opened parentheses.
2458
2459 (c) Conditions that refer to capturing parentheses can be specified
2460 relatively, for example, (?(-2)... or (?(+3)...
2461
2462 (d) \K resets the start of the current match so that everything before
2463 is not part of it.
2464
2465 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2466
2467 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2468 reference syntax.
2469
2470 (g) (?| introduces a group in which the numbering of parentheses in each
2471 alternative starts with the same number.
2472
2473 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2474
2475 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2476 PCRE_INFO_JCHANGED.
2477
2478 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2479 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2480 for detecting groups that can match an empty string.
2481
2482 9. A pattern with a very large number of alternatives (more than several
2483 hundred) was running out of internal workspace during the pre-compile
2484 phase, where pcre_compile() figures out how much memory will be needed. A
2485 bit of new cunning has reduced the workspace needed for groups with
2486 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2487 workspace instead of running out of the 4096 that are available.
2488
2489 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2490
2491 11. Applied patch from Google to remove an optimization that didn't quite work.
2492 The report of the bug said:
2493
2494 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2495 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2496 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2497
2498 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2499 it matched the wrong number of bytes.
2500
2501
2502 Version 7.1 24-Apr-07
2503 ---------------------
2504
2505 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2506 that is more "standard", making use of automake and other Autotools. There
2507 is some re-arrangement of the files and adjustment of comments consequent
2508 on this.
2509
2510 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2511 for recursive directory scanning broke on some systems because the files
2512 are not scanned in any specific order and on different systems the order
2513 was different. A call to "sort" has been inserted into RunGrepTest for the
2514 approprate test as a short-term fix. In the longer term there may be an
2515 alternative.
2516
2517 3. I had an email from Eric Raymond about problems translating some of PCRE's
2518 man pages to HTML (despite the fact that I distribute HTML pages, some
2519 people do their own conversions for various reasons). The problems
2520 concerned the use of low-level troff macros .br and .in. I have therefore
2521 removed all such uses from the man pages (some were redundant, some could
2522 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2523 HTML has been updated to handle .nf/.fi and to complain if it encounters
2524 .br or .in.
2525
2526 4. Updated comments in configure.ac that get placed in config.h.in and also
2527 arranged for config.h to be included in the distribution, with the name
2528 config.h.generic, for the benefit of those who have to compile without
2529 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2530
2531 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2532 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2533 makevp.bat for the current PCRE, using the additional files
2534 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2535
2536 6. A Windows user reported a minor discrepancy with test 2, which turned out
2537 to be caused by a trailing space on an input line that had got lost in his
2538 copy. The trailing space was an accident, so I've just removed it.
2539
2540 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2541 that is needed.
2542
2543 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2544 as "const" (a) because they are and (b) because it helps the PHP
2545 maintainers who have recently made a script to detect big data structures
2546 in the php code that should be moved to the .rodata section. I remembered
2547 to update Builducptable as well, so it won't revert if ucptable.h is ever
2548 re-created.
2549
2550 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2551 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2552 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2553 support is not required. This saves 1.5-2K of code, which is important in
2554 some applications.
2555
2556 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2557 so as not to refer to the tables, even though these functions will never be
2558 called when UTF-8 support is disabled. Otherwise there are problems with a
2559 shared library.
2560
2561 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2562
2563 (a) It was defining its arguments as char * instead of void *.
2564
2565 (b) It was assuming that all moves were upwards in memory; this was true
2566 a long time ago when I wrote it, but is no longer the case.
2567
2568 The emulated memove() is provided for those environments that have neither
2569 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2570 is clearly not the case, as these two bugs were recently reported.
2571
2572 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2573 and Detrail to create the HTML documentation, the .txt form of the man
2574 pages, and it removes trailing spaces from listed files. It also creates
2575 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2576 case, it wraps all the #defines with #ifndefs. This script should be run
2577 before "make dist".
2578
2579 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2580 with Unicode property support.
2581
2582 (a) For a maximizing quantifier, if the two different cases of the
2583 character were of different lengths in their UTF-8 codings (there are
2584 some cases like this - I found 11), and the matching function had to
2585 back up over a mixture of the two cases, it incorrectly assumed they
2586 were both the same length.
2587
2588 (b) When PCRE was configured to use the heap rather than the stack for
2589 recursion during matching, it was not correctly preserving the data for
2590 the other case of a UTF-8 character when checking ahead for a match
2591 while processing a minimizing repeat. If the check also involved
2592 matching a wide character, but failed, corruption could cause an
2593 erroneous result when trying to check for a repeat of the original
2594 character.
2595
2596 13. Some tidying changes to the testing mechanism:
2597
2598 (a) The RunTest script now detects the internal link size and whether there
2599 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2600 values substituted by "configure". (The RunGrepTest script already did
2601 this for UTF-8.) The configure.ac script no longer substitutes the
2602 relevant variables.
2603
2604 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2605 with length and offset values. This means that the output is different
2606 for different internal link sizes. Test 2 is skipped for link sizes
2607 other than 2 because of this, bypassing the problem. Unfortunately,
2608 there was also a test in test 3 (the locale tests) that used /B and
2609 failed for link sizes other than 2. Rather than cut the whole test out,
2610 I have added a new /Z option to pcretest that replaces the length and
2611 offset values with spaces. This is now used to make test 3 independent
2612 of link size. (Test 2 will be tidied up later.)
2613
2614 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2615 segmentation fault instead of returning the appropriate error message.
2616
2617 15. In multiline mode when the newline sequence was set to "any", the pattern
2618 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2619 This doesn't seem right; it now treats the CRLF combination as the line
2620 ending, and so does not match in that case. It's only a pattern such as ^$
2621 that would hit this one: something like ^ABC$ would have failed after \r
2622 and then tried again after \r\n.
2623
2624 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2625 in an attempt to make files that differ only in their line terminators
2626 compare equal. This works on Linux.
2627
2628 17. Under certain error circumstances pcregrep might try to free random memory
2629 as it exited. This is now fixed, thanks to valgrind.
2630
2631 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2632 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2633 was because its rules for how to advance for /g after matching an empty
2634 string at the end of a line did not allow for this case. They now check for
2635 it specially.
2636
2637 20. pcretest is supposed to handle patterns and data of any length, by
2638 extending its buffers when necessary. It was getting this wrong when the
2639 buffer for a data line had to be extended.
2640
2641 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2642 CRLF as a newline sequence.
2643
2644 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2645 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2646 I have nevertheless tidied it up.
2647
2648 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2649
2650 24. Added a man page for pcre-config.
2651
2652
2653 Version 7.0 19-Dec-06
2654 ---------------------
2655
2656 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2657 moving to gcc 4.1.1.
2658
2659 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2660 sys/time.h, which is documented as needed for this function. It doesn't
2661 seem to matter on Linux, but it showed up on some releases of OS X.
2662
2663 3. It seems that there are systems where bytes whose values are greater than
2664 127 match isprint() in the "C" locale. The "C" locale should be the
2665 default when a C program starts up. In most systems, only ASCII printing
2666 characters match isprint(). This difference caused the output from pcretest
2667 to vary, making some of the tests fail. I have changed pcretest so that:
2668
2669 (a) When it is outputting text in the compiled version of a pattern, bytes
2670 other than 32-126 are always shown as hex escapes.
2671
2672 (b) When it is outputting text that is a matched part of a subject string,
2673 it does the same, unless a different locale has been set for the match
2674 (using the /L modifier). In this case, it uses isprint() to decide.
2675
2676 4. Fixed a major bug that caused incorrect computation of the amount of memory
2677 required for a compiled pattern when options that changed within the
2678 pattern affected the logic of the preliminary scan that determines the
2679 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2680 that the computed length was too small. The symptoms of this bug were
2681 either the PCRE error "internal error: code overflow" from pcre_compile(),
2682 or a glibc crash with a message such as "pcretest: free(): invalid next
2683 size (fast)". Examples of patterns that provoked this bug (shown in
2684 pcretest format) are:
2685
2686 /(?-x: )/x
2687 /(?x)(?-x: \s*#\s*)/
2688 /((?i)[\x{c0}])/8
2689 /(?i:[\x{c0}])/8
2690
2691 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2692 is now done differently.
2693
2694 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2695 wrapper classes; (b) implement a new function in the C++ scanner that is
2696 more efficient than the old way of doing things because it avoids levels of
2697 recursion in the regex matching; (c) add a paragraph to the documentation
2698 for the FullMatch() function.
2699
2700 6. The escape sequence \n was being treated as whatever was defined as
2701 "newline". Not only was this contrary to the documentation, which states
2702 that \n is character 10 (hex 0A), but it also went horribly wrong when
2703 "newline" was defined as CRLF. This has been fixed.
2704
2705 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2706 was being set to -1 for the "end of line" case (supposedly a value that no
2707 character can have). Though this value is never used (the check for end of
2708 line is "zero bytes in current character"), it caused compiler complaints.
2709 I've changed it to 0xffffffff.
2710
2711 8. In pcre_version.c, the version string was being built by a sequence of
2712 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2713 string (as it is for production releases) called a macro with an empty
2714 argument. The C standard says the result of this is undefined. The gcc
2715 compiler treats it as an empty string (which was what was wanted) but it is
2716 reported that Visual C gives an error. The source has been hacked around to
2717 avoid this problem.
2718
2719 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2720 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2721 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2722 of them did).
2723
2724 10. Originally, pcretest opened its input and output without "b"; then I was
2725 told that "b" was needed in some environments, so it was added for release
2726 5.0 to both the input and output. (It makes no difference on Unix-like
2727 systems.) Later I was told that it is wrong for the input on Windows. I've
2728 now abstracted the modes into two macros, to make it easier to fiddle with
2729 them, and removed "b" from the input mode under Windows.
2730
2731 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2732
2733 12. Added -help and --help to pcretest as an official way of being reminded
2734 of the options.
2735
2736 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2737 and pcrecpp.cc because they annoy compilers at high warning levels.
2738
2739 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2740
2741 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2742 scripts are not C programs :-) and which was not noticed because it works
2743 on Linux.
2744
2745 16. pcretest is supposed to handle any length of pattern and data line (as one
2746 line or as a continued sequence of lines) by extending its input buffer if
2747 necessary. This feature was broken for very long pattern lines, leading to
2748 a string of junk being passed to pcre_compile() if the pattern was longer
2749 than about 50K.
2750
2751 17. I have done a major re-factoring of the way pcre_compile() computes the
2752 amount of memory needed for a compiled pattern. Previously, there was code
2753 that made a preliminary scan of the pattern in order to do this. That was
2754 OK when PCRE was new, but as the facilities have expanded, it has become
2755 harder and harder to keep it in step with the real compile phase, and there
2756 have been a number of bugs (see for example, 4 above). I have now found a
2757 cunning way of running the real compile function in a "fake" mode that
2758 enables it to compute how much memory it would need, while actually only
2759 ever using a few hundred bytes of working memory and without too many
2760 tests of the mode. This should make future maintenance and development
2761 easier. A side effect of this work is that the limit of 200 on the nesting
2762 depth of parentheses has been removed (though this was never a serious
2763 limitation, I suspect). However, there is a downside: pcre_compile() now
2764 runs more slowly than before (30% or more, depending on the pattern). I
2765 hope this isn't a big issue. There is no effect on runtime performance.
2766
2767 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2768 newline (only possible for the last line of a file) and it was a
2769 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2770
2771 19. Added additional timing features to pcretest. (1) The -tm option now times
2772 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2773 separate command line item, by a number that specifies the number of
2774 repeats to use when timing. The default is 50000; this gives better
2775 precision, but takes uncomfortably long for very large patterns.
2776
2777 20. Extended pcre_study() to be more clever in cases where a branch of a
2778 subpattern has no definite first character. For example, (a*|b*)[cd] would
2779 previously give no result from pcre_study(). Now it recognizes that the
2780 first character must be a, b, c, or d.
2781
2782 21. There was an incorrect error "recursive call could loop indefinitely" if
2783 a subpattern (or the entire pattern) that was being tested for matching an
2784 empty string contained only one non-empty item after a nested subpattern.
2785 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2786 incorrectly, because the \d was being skipped in the check.
2787
2788 22. The pcretest program now has a new pattern option /B and a command line
2789 option -b, which is equivalent to adding /B to every pattern. This causes
2790 it to show the compiled bytecode, without the additional information that
2791 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2792 is the same as /B/I).
2793
2794 23. A new optimization is now able automatically to treat some sequences such
2795 as a*b as a*+b. More specifically, if something simple (such as a character
2796 or a simple class like \d) has an unlimited quantifier, and is followed by
2797 something that cannot possibly match the quantified thing, the quantifier
2798 is automatically "possessified".
2799
2800 24. A recursive reference to a subpattern whose number was greater than 39
2801 went wrong under certain circumstances in UTF-8 mode. This bug could also
2802 have affected the operation of pcre_study().
2803
2804 25. Realized that a little bit of performance could be had by replacing
2805 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2806
2807 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2808
2809 27. Possessive quantifiers such as a++ were previously implemented by turning
2810 them into atomic groups such as ($>a+). Now they have their own opcodes,
2811 which improves performance. This includes the automatically created ones
2812 from 23 above.
2813
2814 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2815 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2816 the first matched character to be a colon. This applied both to named and
2817 numbered groups.
2818
2819 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2820
2821 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2822 building PCRE on Windows easier, so I have included it in the distribution.
2823
2824 31. There is now a check in pcretest against a ridiculously large number being
2825 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2826 loop, the loop is abandoned.
2827
2828 32. Forward references to subpatterns in conditions such as (?(2)...) where
2829 subpattern 2 is defined later cause pcre_compile() to search forwards in
2830 the pattern for the relevant set of parentheses. This search went wrong
2831 when there were unescaped parentheses in a character class, parentheses
2832 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2833
2834 33. "Subroutine" calls and backreferences were previously restricted to
2835 referencing subpatterns earlier in the regex. This restriction has now
2836 been removed.
2837
2838 34. Added a number of extra features that are going to be in Perl 5.10. On the
2839 whole, these are just syntactic alternatives for features that PCRE had
2840 previously implemented using the Python syntax or my own invention. The
2841 other formats are all retained for compatibility.
2842
2843 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2844 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2845 also .NET compatible.
2846
2847 (b) A recursion or subroutine call to a named group can now be defined as
2848 (?&name) as well as (?P>name).
2849
2850 (c) A backreference to a named group can now be defined as \k<name> or
2851 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2852 5.10, are also .NET compatible.
2853
2854 (d) A conditional reference to a named group can now use the syntax
2855 (?(<name>) or (?('name') as well as (?(name).
2856
2857 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2858 groups (named and numbered) that are never evaluated inline, but can be
2859 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2860 is always false. There may be only one alternative in such a group.
2861
2862 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2863 as the simple (?(R). The condition is true only if the most recent
2864 recursion is that of the given number or name. It does not search out
2865 through the entire recursion stack.
2866
2867 (g) The escape \gN or \g{N} has been added, where N is a positive or
2868 negative number, specifying an absolute or relative reference.
2869
2870 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2871 some "unreachable code" warnings.
2872
2873 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2874 things, this adds five new scripts.
2875
2876 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2877 There were also incompatibilities regarding the handling of \Q..\E inside
2878 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2879 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2880
2881 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2882 matches an empty string, and forcibly breaks the loop. There were bugs in
2883 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2884 against aaaa the result was just "a" rather than "aaaa", for example. Two
2885 separate and independent bugs (that affected different cases) have been
2886 fixed.
2887
2888 39. Refactored the code to abolish the use of different opcodes for small
2889 capturing bracket numbers. This is a tidy that I avoided doing when I
2890 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2891 The new approach is not only tidier, it makes it possible to reduce the
2892 memory needed to fix the previous bug (38).
2893
2894 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2895 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2896 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2897 mode.
2898
2899 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2900 report.
2901
2902 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2903 copy construction and assignment in the C++ wrapper.
2904
2905 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2906 couple of bugs that could have given wrong results in the "--newline=crlf"
2907 case.
2908
2909 44. Added a number of casts and did some reorganization of signed/unsigned int
2910 variables following suggestions from Dair Grant. Also renamed the variable
2911 "this" as "item" because it is a C++ keyword.
2912
2913 45. Arranged for dftables to add
2914
2915 #include "pcre_internal.h"
2916
2917 to pcre_chartables.c because without it, gcc 4.x may remove the array
2918 definition from the final binary if PCRE is built into a static library and
2919 dead code stripping is activated.
2920
2921 46. For an unanchored pattern, if a match attempt fails at the start of a
2922 newline sequence, and the newline setting is CRLF or ANY, and the next two
2923 characters are CRLF, advance by two characters instead of one.
2924
2925
2926 Version 6.7 04-Jul-06
2927 ---------------------
2928
2929 1. In order to handle tests when input lines are enormously long, pcretest has
2930 been re-factored so that it automatically extends its buffers when
2931 necessary. The code is crude, but this _is_ just a test program. The
2932 default size has been increased from 32K to 50K.
2933
2934 2. The code in pcre_study() was using the value of the re argument before
2935 testing it for NULL. (Of course, in any sensible call of the function, it
2936 won't be NULL.)
2937
2938 3. The memmove() emulation function in pcre_internal.h, which is used on
2939 systems that lack both memmove() and bcopy() - that is, hardly ever -
2940 was missing a "static" storage class specifier.
2941
2942 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2943 containing an extended class (one that cannot be represented by a bitmap
2944 because it contains high-valued characters or Unicode property items, e.g.
2945 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2946 pattern, but PCRE should not loop if you do not (it no longer does).
2947 [Detail: two cases were found: (a) a repeated subpattern containing an
2948 extended class; (b) a recursive reference to a subpattern that followed a
2949 previous extended class. It wasn't skipping over the extended class
2950 correctly when UTF-8 mode was not set.]
2951
2952 5. A negated single-character class was not being recognized as fixed-length
2953 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2954 compile error "lookbehind assertion is not fixed length".
2955
2956 6. The RunPerlTest auxiliary script was showing an unexpected difference
2957 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2958 write a Perl script that can interpret lines of an input file either as
2959 byte characters or as UTF-8, which is what "perltest" was being required to
2960 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2961 can't do is switch easily at run time between having the "use utf8;" pragma
2962 or not. In the end, I fudged it by using the RunPerlTest script to insert
2963 "use utf8;" explicitly for the UTF-8 tests.
2964
2965 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2966 the end of the subject string, contrary to the documentation and to what
2967 Perl does. This was true of both matching functions. Now it matches only at
2968 the start of the subject and immediately after *internal* newlines.
2969
2970 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2971 a pointer to an int instead of a pointer to an unsigned long int. This
2972 caused problems on 64-bit systems.
2973
2974 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2975 instance of the 'standard' template library not being so standard".
2976
2977 10. There was no check on the number of named subpatterns nor the maximum
2978 length of a subpattern name. The product of these values is used to compute
2979 the size of the memory block for a compiled pattern. By supplying a very
2980 long subpattern name and a large number of named subpatterns, the size
2981 computation could be caused to overflow. This is now prevented by limiting
2982 the length of names to 32 characters, and the number of named subpatterns
2983 to 10,000.
2984
2985 11. Subpatterns that are repeated with specific counts have to be replicated in
2986 the compiled pattern. The size of memory for this was computed from the
2987 length of the subpattern and the repeat count. The latter is limited to
2988 65535, but there was no limit on the former, meaning that integer overflow
2989 could in principle occur. The compiled length of a repeated subpattern is
2990 now limited to 30,000 bytes in order to prevent this.
2991
2992 12. Added the optional facility to have named substrings with the same name.
2993
2994 13. Added the ability to use a named substring as a condition, using the
2995 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2996 are numbers (not recommended). Forward references are permitted.
2997
2998 14. Added forward references in named backreferences (if you see what I mean).
2999
3000 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3001 pattern could run off the end of the subject. For example, the pattern
3002 "(?s)(.{1,5})"8 did this with the subject "ab".
3003
3004 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3005 PCRE_CASELESS was set when matching characters that were quantified with ?
3006 or *.
3007
3008 17. A character class other than a single negated character that had a minimum
3009 but no maximum quantifier - for example [ab]{6,} - was not handled
3010 correctly by pce_dfa_exec(). It would match only one character.
3011
3012 18. A valid (though odd) pattern that looked like a POSIX character
3013 class but used an invalid character after [ (for example [[,abc,]]) caused
3014 pcre_compile() to give the error "Failed: internal error: code overflow" or
3015 in some cases to crash with a glibc free() error. This could even happen if
3016 the pattern terminated after [[ but there just happened to be a sequence of
3017 letters, a binary zero, and a closing ] in the memory that followed.
3018
3019 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3020 over the years. Originally (before any Unicode support), just the bottom 8
3021 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3022 output from "man perlunicode" includes this:
3023
3024 The regular expression compiler produces polymorphic opcodes. That
3025 is, the pattern adapts to the data and automatically switches to
3026 the Unicode character scheme when presented with Unicode data--or
3027 instead uses a traditional byte scheme when presented with byte
3028 data.
3029
3030 Sadly, a wide octal escape does not cause a switch, and in a string with
3031 no other multibyte characters, these octal escapes are treated as before.
3032 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3033 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3034 Unicode string.
3035
3036 I have not perpetrated such confusion in PCRE. Up till now, it took just
3037 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3038 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3039 translate to the appropriate multibyte character.
3040
3041 29. Applied some refactoring to reduce the number of warnings from Microsoft
3042 and Borland compilers. This has included removing the fudge introduced
3043 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3044 a warning about an unused variable.
3045
3046 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3047 characters since release 4.0, because Perl (from release 5.004) does not.
3048 [Or at least, is documented not to: some releases seem to be in conflict
3049 with the documentation.] However, when a pattern was studied with
3050 pcre_study() and all its branches started with \s, PCRE still included VT
3051 as a possible starting character. Of course, this did no harm; it just
3052 caused an unnecessary match attempt.
3053
3054 22. Removed a now-redundant internal flag bit that recorded the fact that case
3055 dependency changed within the pattern. This was once needed for "required
3056 byte" processing, but is no longer used. This recovers a now-scarce options
3057 bit. Also moved the least significant internal flag bit to the most-
3058 significant bit of the word, which was not previously used (hangover from
3059 the days when it was an int rather than a uint) to free up another bit for
3060 the future.
3061
3062 23. Added support for CRLF line endings as well as CR and LF. As well as the
3063 default being selectable at build time, it can now be changed at runtime
3064 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3065 specify that it is scanning data with non-default line endings.
3066
3067 24. Changed the definition of CXXLINK to make it agree with the definition of
3068 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3069
3070 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3071 recursions. This makes a big different to stack usage for some patterns.
3072
3073 26. If a subpattern containing a named recursion or subroutine reference such
3074 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3075 the space required for the compiled pattern went wrong and gave too small a
3076 value. Depending on the environment, this could lead to "Failed: internal
3077 error: code overflow at offset 49" or "glibc detected double free or
3078 corruption" errors.
3079
3080 27. Applied patches from Google (a) to support the new newline modes and (b) to
3081 advance over multibyte UTF-8 characters in GlobalReplace.
3082
3083 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3084 difference for some implementation of PCRE in some Windows version.
3085
3086 29. Added some extra testing facilities to pcretest:
3087
3088 \q<number> in a data line sets the "match limit" value
3089 \Q<number> in a data line sets the "match recursion limt" value
3090 -S <number> sets the stack size, where <number> is in megabytes
3091
3092 The -S option isn't available for Windows.
3093
3094
3095 Version 6.6 06-Feb-06
3096 ---------------------
3097
3098 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3099 in pcreposix.h. I have copied the definition from pcre.h.
3100
3101 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3102 because pcre.h is no longer a built file.
3103
3104 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3105 not normally included in the compiled code.
3106
3107
3108 Version 6.5 01-Feb-06
3109 ---------------------
3110
3111 1. When using the partial match feature with pcre_dfa_exec(), it was not
3112 anchoring the second and subsequent partial matches at the new starting
3113 point. This could lead to incorrect results. For example, with the pattern
3114 /1234/, partially matching against "123" and then "a4" gave a match.
3115
3116 2. Changes to pcregrep:
3117
3118 (a) All non-match returns from pcre_exec() were being treated as failures
3119 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3120 error message is output. Some extra information is given for the
3121 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3122 probably the only errors that are likely to be caused by users (by
3123 specifying a regex that has nested indefinite repeats, for instance).
3124 If there are more than 20 of these errors, pcregrep is abandoned.
3125
3126 (b) A binary zero was treated as data while matching, but terminated the
3127 output line if it was written out. This has been fixed: binary zeroes
3128 are now no different to any other data bytes.
3129
3130 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3131 used to set a locale for matching. The --locale=xxxx long option has
3132 been added (no short equivalent) to specify a locale explicitly on the
3133 pcregrep command, overriding the environment variables.
3134
3135 (d) When -B was used with -n, some line numbers in the output were one less
3136 than they should have been.
3137
3138 (e) Added the -o (--only-matching) option.
3139
3140 (f) If -A or -C was used with -c (count only), some lines of context were
3141 accidentally printed for the final match.
3142
3143 (g) Added the -H (--with-filename) option.
3144
3145 (h) The combination of options -rh failed to suppress file names for files
3146 that were found from directory arguments.
3147
3148 (i) Added the -D (--devices) and -d (--directories) options.
3149
3150 (j) Added the -F (--fixed-strings) option.
3151
3152 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3153
3154 (l) Added the --colo(u)r option.
3155
3156 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3157 is not present by default.
3158
3159 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3160 items such as (?R) or (?1), when the recursion could match a number of
3161 alternatives. If it matched one of the alternatives, but subsequently,
3162 outside the recursion, there was a failure, the code tried to back up into
3163 the recursion. However, because of the way PCRE is implemented, this is not
3164 possible, and the result was an incorrect result from the match.
3165
3166 In order to prevent this happening, the specification of recursion has
3167 been changed so that all such subpatterns are automatically treated as
3168 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3169
3170 4. I had overlooked the fact that, in some locales, there are characters for
3171 which isalpha() is true but neither isupper() nor islower() are true. In
3172 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3173 and ordfeminine) are like this. This affected the treatment of \w and \W
3174 when they appeared in character classes, but not when they appeared outside
3175 a character class. The bit map for "word" characters is now created
3176 separately from the results of isalnum() instead of just taking it from the
3177 upper, lower, and digit maps. (Plus the underscore character, of course.)
3178
3179 5. The above bug also affected the handling of POSIX character classes such as
3180 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3181 permanent tables. Instead, the bit maps for such a class were previously
3182 created as the appropriate unions of the upper, lower, and digit bitmaps.
3183 Now they are created by subtraction from the [[:word:]] class, which has
3184 its own bitmap.
3185
3186 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3187 It is created by subtracting the vertical space characters (\x09, \x0a,
3188 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3189 subtraction was done in the overall bitmap for a character class, meaning
3190 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3191 be recognized. This bug has been fixed.
3192
3193 7. Patches from the folks at Google:
3194
3195 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3196 real life, but is still worth protecting against".
3197
3198 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3199 regular expressions".
3200
3201 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3202 have it.
3203
3204 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3205 "configure" and the latter not, in order to fix a problem somebody had
3206 with compiling the Arg class on HP-UX.
3207
3208 (e) Improve the error-handling of the C++ wrapper a little bit.
3209
3210 (f) New tests for checking recursion limiting.
3211
3212 8. The pcre_memmove() function, which is used only if the environment does not
3213 have a standard memmove() function (and is therefore rarely compiled),
3214 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3215 returning a result (though PCRE never actually uses the result).
3216
3217 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3218 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3219 returned instead of calling malloc() with an overflowing number that would
3220 most likely cause subsequent chaos.
3221
3222 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3223
3224 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3225 with this option is matched, the nmatch and pmatch options of regexec() are
3226 ignored.
3227
3228 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3229 provided in case anyone wants to the the POSIX interface with UTF-8
3230 strings.
3231
3232 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3233 C++ linking (needed for some HP-UX environments).
3234
3235 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3236 (unused parameter) and in the pcre_printint() function (omitted "default"
3237 switch label when the default is to do nothing).
3238
3239 15. Added some code to make it possible, when PCRE is compiled as a C++
3240 library, to replace subject pointers for pcre_exec() with a smart pointer
3241 class, thus making it possible to process discontinuous strings.
3242
3243 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3244 much the same function. They were added by different people who were trying
3245 to make PCRE easy to compile on non-Unix systems. It has been suggested
3246 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3247 for compiling on Windows systems. I have therefore replaced it with
3248 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3249 defaults to "extern" for C or "extern C" for C++, which works fine on
3250 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3251 SCOPE with something explicit in config.h. In addition:
3252
3253 (a) pcreposix.h still had just "extern" instead of either of these macros;
3254 I have replaced it with PCRE_DATA_SCOPE.
3255
3256 (b) Functions such as _pcre_xclass(), which are internal to the library,
3257 but external in the C sense, all had PCRE_EXPORT in their definitions.
3258 This is apparently wrong for the Windows case, so I have removed it.
3259 (It makes no difference on Unix-like systems.)
3260
3261 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3262 of recursive calls to match(). This is different to MATCH_LIMIT because
3263 that limits the total number of calls to match(), not all of which increase
3264 the depth of recursion. Limiting the recursion depth limits the amount of
3265 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3266 when PCRE is compiled, and changed at run time. A patch from Google adds
3267 this functionality to the C++ interface.
3268
3269 18. Changes to the handling of Unicode character properties:
3270
3271 (a) Updated the table to Unicode 4.1.0.
3272
3273 (b) Recognize characters that are not in the table as "Cn" (undefined).
3274
3275 (c) I revised the way the table is implemented to a much improved format
3276 which includes recognition of ranges. It now supports the ranges that
3277 are defined in UnicodeData.txt, and it also amalgamates other
3278 characters into ranges. This has reduced the number of entries in the
3279 table from around 16,000 to around 3,000, thus reducing its size
3280 considerably. I realized I did not need to use a tree structure after
3281 all - a binary chop search is just as efficient. Having reduced the
3282 number of entries, I extended their size from 6 bytes to 8 bytes to
3283 allow for more data.
3284
3285 (d) Added support for Unicode script names via properties such as \p{Han}.
3286
3287 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3288 matching that character.
3289
3290 20. When matching a repeated Unicode property with a minimum greater than zero,
3291 (for example \pL{2,}), PCRE could look past the end of the subject if it
3292 reached it while seeking the minimum number of characters. This could
3293 happen only if some of the characters were more than one byte long, because
3294 there is a check for at least the minimum number of bytes.
3295
3296 21. Refactored the implementation of \p and \P so as to be more general, to
3297 allow for more different types of property in future. This has changed the
3298 compiled form incompatibly. Anybody with saved compiled patterns that use
3299 \p or \P will have to recompile them.
3300
3301 22. Added "Any" and "L&" to the supported property types.
3302
3303 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3304 but give a compile time error if the value is greater than 0xff.
3305
3306 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3307 accidentally not being installed or uninstalled.
3308
3309 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3310 made were to insert the current release number. This seemed silly, because
3311 it made things harder for people building PCRE on systems that don't run
3312 "configure". I have turned pcre.h into a distributed file, no longer built
3313 by "configure", with the version identification directly included. There is
3314 no longer a pcre.h.in file.
3315
3316 However, this change necessitated a change to the pcre-config script as
3317 well. It is built from pcre-config.in, and one of the substitutions was the
3318 release number. I have updated configure.ac so that ./configure now finds
3319 the release number by grepping pcre.h.
3320
3321 26. Added the ability to run the tests under valgrind.
3322
3323
3324 Version 6.4 05-Sep-05
3325 ---------------------
3326
3327 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3328 "--" to be printed when multiple files were scanned, even when none of the
3329 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3330 consider it to be a bug, and have restored the previous behaviour.
3331
3332 2. A couple of code tidies to get rid of compiler warnings.
3333
3334 3. The pcretest program used to cheat by referring to symbols in the library
3335 whose names begin with _pcre_. These are internal symbols that are not
3336 really supposed to be visible externally, and in some environments it is
3337 possible to suppress them. The cheating is now confined to including
3338 certain files from the library's source, which is a bit cleaner.
3339
3340 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3341 file's purpose clearer.
3342
3343 5. Reorganized pcre_ucp_findchar().
3344
3345
3346 Version 6.3 15-Aug-05
3347 ---------------------
3348
3349 1. The file libpcre.pc.in did not have general read permission in the tarball.
3350
3351 2. There were some problems when building without C++ support:
3352
3353 (a) If C++ support was not built, "make install" and "make test" still
3354 tried to test it.
3355
3356 (b) There were problems when the value of CXX was explicitly set. Some
3357 changes have been made to try to fix these, and ...
3358
3359 (c) --disable-cpp can now be used to explicitly disable C++ support.
3360
3361 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3362 backslash in a target when C++ was disabled. This confuses some
3363 versions of "make", apparently. Using an intermediate variable solves
3364 this. (Same for CPP_LOBJ.)
3365
3366 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3367 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3368 necessary on certain architectures.
3369
3370 4. Added a setting of -export-symbols-regex to the link command to remove
3371 those symbols that are exported in the C sense, but actually are local
3372 within the library, and not documented. Their names all begin with
3373 "_pcre_". This is not a perfect job, because (a) we have to except some
3374 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3375 available (and never for static libraries). I have made a note to try to
3376 find a way round (a) in the future.
3377
3378
3379 Version 6.2 01-Aug-05
3380 ---------------------
3381
3382 1. There was no test for integer overflow of quantifier values. A construction
3383 such as {1111111111111111} would give undefined results. What is worse, if
3384 a minimum quantifier for a parenthesized subpattern overflowed and became
3385 negative, the calculation of the memory size went wrong. This could have
3386 led to memory overwriting.
3387
3388 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3389
3390 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3391 operating environments where this matters.
3392
3393 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3394 PCRE options from within the C++ wrapper.
3395
3396 5. Named capturing subpatterns were not being correctly counted when a pattern
3397 was compiled. This caused two problems: (a) If there were more than 100
3398 such subpatterns, the calculation of the memory needed for the whole
3399 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3400 back references of the form \12, where the number was greater than 9, were
3401 not recognized as back references, even though there were sufficient
3402 previous subpatterns.
3403
3404 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3405 versions of gcc, e.g. 2.95.4.
3406
3407
3408 Version 6.1 21-Jun-05
3409 ---------------------
3410
3411 1. There was one reference to the variable "posix" in pcretest.c that was not
3412 surrounded by "#if !defined NOPOSIX".
3413
3414 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3415 the cross-check on the old pcre_info() function, for the benefit of the
3416 cut-down version of PCRE that is currently imported into Exim.
3417
3418 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3419 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3420 patterns that start (?i) but I don't think that matters. The use of (?i) is
3421 just an example; this all applies to the other options as well.
3422
3423 4. Since libtool seems to echo the compile commands it is issuing, the output
3424 from "make" can be reduced a bit by putting "@" in front of each libtool
3425 compile command.
3426
3427 5. Patch from the folks at Google for configure.in to be a bit more thorough
3428 in checking for a suitable C++ installation before trying to compile the
3429 C++ stuff. This should fix a reported problem when a compiler was present,
3430 but no suitable headers.
3431
3432 6. The man pages all had just "PCRE" as their title. I have changed them to
3433 be the relevant file name. I have also arranged that these names are
3434 retained in the file doc/pcre.txt, which is a concatenation in text format
3435 of all the man pages except the little individual ones for each function.
3436
3437 7. The NON-UNIX-USE file had not been updated for the different set of source
3438 files that come with release 6. I also added a few comments about the C++
3439 wrapper.
3440
3441
3442 Version 6.0 07-Jun-05
3443 ---------------------
3444
3445 1. Some minor internal re-organization to help with my DFA experiments.
3446
3447 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3448 didn't matter for the library itself when fully configured, but did matter
3449 when compiling without UCP support, or within Exim, where the ucp files are
3450 not imported.
3451
3452 3. Refactoring of the library code to split up the various functions into
3453 different source modules. The addition of the new DFA matching code (see
3454 below) to a single monolithic source would have made it really too
3455 unwieldy, quite apart from causing all the code to be include in a
3456 statically linked application, when only some functions are used. This is
3457 relevant even without the DFA addition now that patterns can be compiled in
3458 one application and matched in another.
3459
3460 The downside of splitting up is that there have to be some external
3461 functions and data tables that are used internally in different modules of
3462 the library but which are not part of the API. These have all had their
3463 names changed to start with "_pcre_" so that they are unlikely to clash
3464 with other external names.
3465
3466 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3467 a different (DFA) algorithm. Although it is slower than the original
3468 function, it does have some advantages for certain types of matching
3469 problem.
3470
3471 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3472 including restarting after a partial match.
3473
3474 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3475 defined when compiling for Windows was sent to me. I have put it into the
3476 code, though I have no means of testing or verifying it.
3477
3478 7. Added the pcre_refcount() auxiliary function.
3479
3480 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3481 match before or at the first newline in the subject string. In pcretest,
3482 the /f option on a pattern can be used to set this.
3483
3484 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3485 would behave wrongly. This has been present in PCRE since release 4.0.
3486
3487 10. A number of changes to the pcregrep command:
3488
3489 (a) Refactored how -x works; insert ^(...)$ instead of setting
3490 PCRE_ANCHORED and checking the length, in preparation for adding
3491 something similar for -w.
3492
3493 (b) Added the -w (match as a word) option.
3494
3495 (c) Refactored the way lines are read and buffered so as to have more
3496 than one at a time available.
3497
3498 (d) Implemented a pcregrep test script.
3499
3500 (e) Added the -M (multiline match) option. This allows patterns to match
3501 over several lines of the subject. The buffering ensures that at least
3502 8K, or the rest of the document (whichever is the shorter) is available
3503 for matching (and similarly the previous 8K for lookbehind assertions).
3504
3505 (f) Changed the --help output so that it now says
3506
3507 -w, --word-regex(p)
3508
3509 instead of two lines, one with "regex" and the other with "regexp"
3510 because that confused at least one person since the short forms are the
3511 same. (This required a bit of code, as the output is generated
3512 automatically from a table. It wasn't just a text change.)
3513
3514 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3515 option but starts with a hyphen. Could be a pattern or a path name
3516 starting with a hyphen, for instance.
3517
3518 (h) "-" can be given as a file name to represent stdin.
3519
3520 (i) When file names are being printed, "(standard input)" is used for
3521 the standard input, for compatibility with GNU grep. Previously
3522 "<stdin>" was used.
3523
3524 (j) The option --label=xxx can be used to supply a name to be used for
3525 stdin when file names are being printed. There is no short form.
3526
3527 (k) Re-factored the options decoding logic because we are going to add
3528 two more options that take data. Such options can now be given in four
3529 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3530
3531 (l) Added the -A, -B, and -C options for requesting that lines of context
3532 around matches be printed.
3533
3534 (m) Added the -L option to print the names of files that do not contain
3535 any matching lines, that is, the complement of -l.
3536
3537 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3538 continue to scan other files.
3539
3540 (o) The -s option was incorrectly implemented. For compatibility with other
3541 greps, it now suppresses the error message for a non-existent or non-
3542 accessible file (but not the return code). There is a new option called
3543 -q that suppresses the output of matching lines, which was what -s was
3544 previously doing.
3545
3546 (p) Added --include and --exclude options to specify files for inclusion
3547 and exclusion when recursing.
3548
3549 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3550 Hopefully, it now does.
3551
3552 12. Missing cast in pcre_study().
3553
3554 13. Added an "uninstall" target to the makefile.
3555
3556 14. Replaced "extern" in the function prototypes in Makefile.in with
3557 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3558 world, but is set differently for Windows.
3559
3560 15. Added a second compiling function called pcre_compile2(). The only
3561 difference is that it has an extra argument, which is a pointer to an
3562 integer error code. When there is a compile-time failure, this is set
3563 non-zero, in addition to the error test pointer being set to point to an
3564 error message. The new argument may be NULL if no error number is required
3565 (but then you may as well call pcre_compile(), which is now just a
3566 wrapper). This facility is provided because some applications need a
3567 numeric error indication, but it has also enabled me to tidy up the way
3568 compile-time errors are handled in the POSIX wrapper.
3569
3570 16. Added VPATH=.libs to the makefile; this should help when building with one
3571 prefix path and installing with another. (Or so I'm told by someone who
3572 knows more about this stuff than I do.)
3573
3574 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3575 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3576 match everything, including newlines. This is not POSIX-compatible, but
3577 somebody wanted the feature. From pcretest it can be activated by using
3578 both the P and the s flags.
3579
3580 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3581
3582 19. libpcre.pc was being incorrectly installed as executable.
3583
3584 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3585 it now also looks for '\r' so that it will work unmodified on Windows.
3586
3587 21. Added Google's contributed C++ wrapper to the distribution.
3588
3589 22. Added some untidy missing memory free() calls in pcretest, to keep
3590 Electric Fence happy when testing.
3591
3592
3593
3594 Version 5.0 13-Sep-04
3595 ---------------------
3596
3597 1. Internal change: literal characters are no longer packed up into items
3598 containing multiple characters in a single byte-string. Each character
3599 is now matched using a separate opcode. However, there may be more than one
3600 byte in the character in UTF-8 mode.
3601
3602 2. The pcre_callout_block structure has two new fields: pattern_position and
3603 next_item_length. These contain the offset in the pattern to the next match
3604 item, and its length, respectively.
3605
3606 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3607 insertion of callouts before each pattern item. Added the /C option to
3608 pcretest to make use of this.
3609
3610 4. On the advice of a Windows user, the lines
3611
3612 #if defined(_WIN32) || defined(WIN32)
3613 _setmode( _fileno( stdout ), 0x8000 );
3614 #endif /* defined(_WIN32) || defined(WIN32) */
3615
3616 have been added to the source of pcretest. This apparently does useful
3617 magic in relation to line terminators.
3618
3619 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3620 for the benefit of those environments where the "b" makes a difference.
3621
3622 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3623 to know about it. I have put a hack into configure.in that adds in code
3624 to set GCC=yes if CC=icc. This seems to end up at a point in the
3625 generated configure script that is early enough to affect the setting of
3626 compiler options, which is what is needed, but I have no means of testing
3627 whether it really works. (The user who reported this had patched the
3628 generated configure script, which of course I cannot do.)
3629
3630 LATER: After change 22 below (new libtool files), the configure script
3631 seems to know about icc (and also ecc). Therefore, I have commented out
3632 this hack in configure.in.
3633
3634 7. Added support for pkg-config (2 patches were sent in).
3635
3636 8. Negated POSIX character classes that used a combination of internal tables
3637 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3638 [[:^ascii]]. Typically, they would match almost any characters. The other
3639 POSIX classes were not broken in this way.
3640
3641 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3642 to find the match, as PCRE was deluded into thinking that the match had to
3643 start at the start point or following a newline. The same bug applied to
3644 patterns with negative forward assertions or any backward assertions
3645 preceding ".*" at the start, unless the pattern required a fixed first
3646 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3647
3648 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3649 starting at the last subject character, bytes beyond the end of the subject
3650 string were read.
3651
3652 11. Renamed the variable "class" as "classbits" to make life easier for C++
3653 users. (Previously there was a macro definition, but it apparently wasn't
3654 enough.)
3655
3656 12. Added the new field "tables" to the extra data so that tables can be passed
3657 in at exec time, or the internal tables can be re-selected. This allows
3658 a compiled regex to be saved and re-used at a later time by a different
3659 program that might have everything at different addresses.
3660
3661 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3662 -R library as well as a -L library.
3663
3664 14. The debugging options of pcretest (-d on the command line or D on a
3665 pattern) showed incorrect output for anything following an extended class
3666 that contained multibyte characters and which was followed by a quantifier.
3667
3668 15. Added optional support for general category Unicode character properties
3669 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3670 support. It adds about 90K to the size of the library. The meanings of the
3671 inbuilt class escapes such as \d and \s have NOT been changed.
3672
3673 16. Updated pcredemo.c to include calls to free() to release the memory for the
3674 compiled pattern.
3675
3676 17. The generated file chartables.c was being created in the source directory
3677 instead of in the building directory. This caused the build to fail if the
3678 source directory was different from the building directory, and was
3679 read-only.
3680
3681 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3682 file. No doubt somebody will tell me if they don't make sense... Also added
3683 Dan Mooney's comments about building on OpenVMS.
3684
3685 19. Added support for partial matching via the PCRE_PARTIAL option for
3686 pcre_exec() and the \P data escape in pcretest.
3687
3688 20. Extended pcretest with 3 new pattern features:
3689
3690 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3691 write the compiled pattern to the file whose name is "rest-of-line".
3692 This is a straight binary dump of the data, with the saved pointer to
3693 the character tables forced to be NULL. The study data, if any, is
3694 written too. After writing, pcretest reads a new pattern.
3695
3696 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3697 compiled pattern from the given file. There must not be any
3698 occurrences of "<" in the file name (pretty unlikely); if there are,
3699 pcretest will instead treat the initial "<" as a pattern delimiter.
3700 After reading in the pattern, pcretest goes on to read data lines as
3701 usual.
3702
3703 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3704 and 16-bit fields in a compiled pattern, to simulate a pattern that
3705 was compiled on a host of opposite endianness.
3706
3707 21. The pcre-exec() function can now cope with patterns that were compiled on
3708 hosts of opposite endianness, with this restriction:
3709
3710 As for any compiled expression that is saved and used later, the tables
3711 pointer field cannot be preserved; the extra_data field in the arguments
3712 to pcre_exec() should be used to pass in a tables address if a value
3713 other than the default internal tables were used at compile time.
3714
3715 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3716 now diagnosed as an error. Previously, most of the time, a negative number
3717 would have been treated as zero, but if in addition "ovector" was passed as
3718 NULL, a crash could occur.
3719
3720 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3721 new versions from the libtool 1.5 distribution (the last one is a copy of
3722 a file called libtool.m4). This seems to have fixed the need to patch
3723 "configure" to support Darwin 1.3 (which I used to do). However, I still
3724 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3725 workstation).
3726
3727 24. Changed the PCRE licence to be the more standard "BSD" licence.
3728
3729
3730 Version 4.5 01-Dec-03
3731 ---------------------
3732
3733 1. There has been some re-arrangement of the code for the match() function so
3734 that it can be compiled in a version that does not call itself recursively.
3735 Instead, it keeps those local variables that need separate instances for
3736 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3737 needs to "recurse". Keeping track of where control must go is done by means
3738 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3739 hide most of the details from the main code, and operates only if
3740 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3741 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3742 operating.
3743
3744 To make it easier for callers to provide specially tailored get/free
3745 functions for this usage, two new functions, pcre_stack_malloc, and
3746 pcre_stack_free, are used. They are always called in strict stacking order,
3747 and the size of block requested is always the same.
3748
3749 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3750 PCRE has been compiled to use the stack or the heap for recursion. The
3751 -C option of pcretest uses this to show which version is compiled.
3752
3753 A new data escape \S, is added to pcretest; it causes the amounts of store
3754 obtained and freed by both kinds of malloc/free at match time to be added
3755 to the output.
3756
3757 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3758 what's available on my current Linux desktop machine.
3759
3760 3. When matching a UTF-8 string, the test for a valid string at the start has
3761 been extended. If start_offset is not zero, PCRE now checks that it points
3762 to a byte that is the start of a UTF-8 character. If not, it returns
3763 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3764 this is necessary because there may be backward assertions in the pattern.
3765 When matching the same subject several times, it may save resources to use
3766 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3767
3768 4. The code for checking the validity of UTF-8 strings has been tightened so
3769 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3770 containing "overlong sequences".
3771
3772 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3773 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3774 should have been "&", but it just so happened that all the cases this let
3775 through by mistake were picked up later in the function.
3776
3777 6. I had used a variable called "isblank" - this is a C99 function, causing
3778 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3779
3780 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3781 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3782 so the version is shown; (c) stop "make test" echoing "./RunTest".
3783
3784 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3785
3786 9. The prototype for memmove() for systems that don't have it was using
3787 size_t, but the inclusion of the header that defines size_t was later. I've
3788 moved the #includes for the C headers earlier to avoid this.
3789
3790 10. Added some adjustments to the code to make it easier to compiler on certain
3791 special systems:
3792
3793 (a) Some "const" qualifiers were missing.
3794 (b) Added the macro EXPORT before all exported functions; by default this
3795 is defined to be empty.
3796 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3797 that it reads its output file name as an argument instead of writing
3798 to the standard output and assuming this can be redirected.
3799
3800 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3801 class containing characters with values greater than 255, PCRE compilation
3802 went into a loop.
3803
3804 12. A recursive reference to a subpattern that was within another subpattern
3805 that had a minimum quantifier of zero caused PCRE to crash. For example,
3806 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3807 recursion. If the recursively-called subpattern itself had a zero repeat,
3808 that was OK.
3809
3810 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3811 buffer into which it was copied (for escape processing) was still set at
3812 1024, so long lines caused crashes.
3813
3814 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3815 "internal error: code overflow...". This applied to any character class
3816 that was followed by a possessive quantifier.
3817
3818 15. Modified the Makefile to add libpcre.la as a prerequisite for
3819 libpcreposix.la because I was told this is needed for a parallel build to
3820 work.
3821
3822 16. If a pattern that contained .* following optional items at the start was
3823 studied, the wrong optimizing data was generated, leading to matching
3824 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3825 matching string must start with a or b or c. The correct conclusion for
3826 this pattern is that a match can start with any character.
3827
3828
3829 Version 4.4 13-Aug-03
3830 ---------------------
3831
3832 1. In UTF-8 mode, a character class containing characters with values between
3833 127 and 255 was not handled correctly if the compiled pattern was studied.
3834 In fixing this, I have also improved the studying algorithm for such
3835 classes (slightly).
3836
3837 2. Three internal functions had redundant arguments passed to them. Removal
3838 might give a very teeny performance improvement.
3839
3840 3. Documentation bug: the value of the capture_top field in a callout is *one
3841 more than* the number of the hightest numbered captured substring.
3842
3843 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3844 in incorrectly linking with a previously installed version. They now link
3845 explicitly with libpcre.la.
3846
3847 5. configure.in no longer needs to recognize Cygwin specially.
3848
3849 6. A problem in pcre.in for Windows platforms is fixed.
3850
3851 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3852 pcretest, it used to include the size of the study block as part of its
3853 output. Unfortunately, the structure contains a field that has a different
3854 size on different hardware architectures. This meant that the tests that
3855 showed this size failed. As the block is currently always of a fixed size,
3856 this information isn't actually particularly useful in pcretest output, so
3857 I have just removed it.
3858
3859 8. Three pre-processor statements accidentally did not start in column 1.
3860 Sadly, there are *still* compilers around that complain, even though
3861 standard C has not required this for well over a decade. Sigh.
3862
3863 9. In pcretest, the code for checking callouts passed small integers in the
3864 callout_data field, which is a void * field. However, some picky compilers
3865 complained about the casts involved for this on 64-bit systems. Now
3866 pcretest passes the address of the small integer instead, which should get
3867 rid of the warnings.
3868
3869 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3870 both compile and run time, and gives an error if an invalid UTF-8 sequence
3871 is found. There is a option for disabling this check in cases where the
3872 string is known to be correct and/or the maximum performance is wanted.
3873
3874 11. In response to a bug report, I changed one line in Makefile.in from
3875
3876 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3877 to
3878 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3879
3880 to look similar to other lines, but I have no way of telling whether this
3881 is the right thing to do, as I do not use Windows. No doubt I'll get told
3882 if it's wrong...
3883
3884
3885 Version 4.3 21-May-03
3886 ---------------------
3887
3888 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3889 Makefile.
3890
3891 2. Some refactoring to improve the quality of the code:
3892
3893 (i) The utf8_table... variables are now declared "const".
3894
3895 (ii) The code for \cx, which used the "case flipping" table to upper case
3896 lower case letters, now just substracts 32. This is ASCII-specific,
3897 but the whole concept of \cx is ASCII-specific, so it seems
3898 reasonable.
3899
3900 (iii) PCRE was using its character types table to recognize decimal and
3901 hexadecimal digits in the pattern. This is silly, because it handles
3902 only 0-9, a-f, and A-F, but the character types table is locale-
3903 specific, which means strange things might happen. A private
3904 table is now used for this - though it costs 256 bytes, a table is
3905 much faster than multiple explicit tests. Of course, the standard
3906 character types table is still used for matching digits in subject
3907 strings against \d.
3908
3909 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3910 ending in _t are). So I've renamed it as ESC_tee.
3911
3912 3. The first argument for regexec() in the POSIX wrapper should have been
3913 defined as "const".
3914
3915 4. Changed pcretest to use malloc() for its buffers so that they can be
3916 Electric Fenced for debugging.
3917
3918 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3919 to read one or more bytes before the start of the subject string. Often this
3920 had no effect on PCRE's behaviour, but in some circumstances it could
3921 provoke a segmentation fault.
3922
3923 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3924 to try to read one or more bytes before the start of the subject string.
3925
3926 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3927 UTF-8 support could misbehave in various ways if the subject string
3928 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3929 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3930 back over UTF-8 characters.)
3931
3932
3933 Version 4.2 14-Apr-03
3934 ---------------------
3935
3936 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3937
3938 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3939 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3940 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3941 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3942 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3943 and BUILD_EXEEXT
3944 Note: automatic setting of the BUILD variables is not yet working
3945 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3946 compile-time but not at link-time
3947 [LINK]: use for linking executables only
3948 make different versions for Windows and non-Windows
3949 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3950 libraries
3951 [LINK_FOR_BUILD]: new variable
3952 [OBJEXT]: use throughout
3953 [EXEEXT]: use throughout
3954 <winshared>: new target
3955 <wininstall>: new target
3956 <dftables.o>: use native compiler
3957 <dftables>: use native linker
3958 <install>: handle Windows platform correctly
3959 <clean>: ditto
3960 <check>: ditto
3961 copy DLL to top builddir before testing
3962
3963 As part of these changes, -no-undefined was removed again. This was reported
3964 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3965 in any case.
3966
3967 3. Some tidies to get rid of compiler warnings:
3968
3969 . In the match_data structure, match_limit was an unsigned long int, whereas
3970 match_call_count was an int. I've made them both unsigned long ints.
3971
3972 . In pcretest the fact that a const uschar * doesn't automatically cast to
3973 a void * provoked a warning.
3974
3975 . Turning on some more compiler warnings threw up some "shadow" variables
3976 and a few more missing casts.
3977
3978 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3979 option, a class that contained a single character with a value between 128
3980 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3981
3982 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3983 option, a class that contained several characters, but with at least one
3984 whose value was between 128 and 255 caused PCRE to crash.
3985
3986
3987 Version 4.1 12-Mar-03
3988 ---------------------
3989
3990 1. Compiling with gcc -pedantic found a couple of places where casts were
3991 needed, and a string in dftables.c that was longer than standard compilers are
3992 required to support.
3993
3994 2. Compiling with Sun's compiler found a few more places where the code could
3995 be tidied up in order to avoid warnings.
3996
3997 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3998 first of these names is deprecated in the latest Autoconf in favour of the name
3999 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4000 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4001 analogy I have changed it to CFLAGS_FOR_BUILD.
4002
4003 4. Added -no-undefined to the linking command in the Makefile, because this is
4004 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4005 linking step for the pcreposix library.
4006
4007 5. PCRE was failing to diagnose the case of two named groups with the same
4008 name.
4009
4010 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4011 literal character that is needed in the subject for a match, and scans along to
4012 ensure that it is present before embarking on the full matching process. This
4013 saves time in cases of nested unlimited repeats that are never going to match.
4014 Problem: the scan can take a lot of time if the subject is very long (e.g.
4015 megabytes), thus penalizing straightforward matches. It is now done only if the
4016 amount of subject to be scanned is less than 1000 bytes.
4017
4018 7. A lesser problem with the same optimization is that it was recording the
4019 first character of an anchored pattern as "needed", thus provoking a search
4020 right along the subject, even when the first match of the pattern was going to
4021 fail. The "needed" character is now not set for anchored patterns, unless it
4022 follows something in the pattern that is of non-fixed length. Thus, it still
4023 fulfils its original purpose of finding quick non-matches in cases of nested
4024 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4025
4026
4027 Version 4.0 17-Feb-03
4028 ---------------------
4029
4030 1. If a comment in an extended regex that started immediately after a meta-item
4031 extended to the end of string, PCRE compiled incorrect data. This could lead to
4032 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4033
4034 2. Moved to autoconf 2.53 and libtool 1.4.2.
4035
4036 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4037 the special perltest8 script is no longer needed - all the tests can be run
4038 from a single perltest script.
4039
4040 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4041 by \s. It has now been removed in PCRE. This means it isn't recognized as
4042 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4043 class [:space:] *does* include VT, thereby creating a mess.
4044
4045 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4046 space and tab.
4047
4048 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4049 its new features into the main test script, reducing the number of scripts.
4050
4051 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4052 were backward compatible, and made the (?i) apply to the whole pattern, as if
4053 /i were given. Now it behaves more logically, and applies the option setting
4054 only to what follows. PCRE has been changed to follow suit. However, if it
4055 finds options settings right at the start of the pattern, it extracts them into
4056 the global options, as before. Thus, they show up in the info data.
4057
4058 8. Added support for the \Q...\E escape sequence. Characters in between are
4059 treated as literals. This is slightly different from Perl in that $ and @ are
4060 also handled as literals inside the quotes. In Perl, they will cause variable
4061 interpolation. Note the following examples:
4062
4063 Pattern PCRE matches Perl matches
4064
4065 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4066 \Qabc\$xyz\E abc\$xyz abc\$xyz
4067 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4068
4069 For compatibility with Perl, \Q...\E sequences are recognized inside character
4070 classes as well as outside them.
4071
4072 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4073 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4074 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4075 signed/unsigned warnings.
4076
4077 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4078 option for pcretest, so I've replaced it by a simple function that does just
4079 that job.
4080
4081 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4082 "pcregrep -".
4083
4084 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4085 Java package. This provides some syntactic sugar for simple cases of what my
4086 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4087 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4088 item, you can use this simplified notation. Note that only makes sense with
4089 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4090 greediness, whatever the setting of the PCRE_UNGREEDY option.
4091
4092 13. A change of greediness default within a pattern was not taking effect at
4093 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4094 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4095 was abstracted outside.
4096
4097 14. PCRE now supports the \G assertion. It is true when the current matching
4098 position is at the start point of the match. This differs from \A when the
4099 starting offset is non-zero. Used with the /g option of pcretest (or similar
4100 code), it works in the same way as it does for Perl's /g option. If all
4101 alternatives of a regex begin with \G, the expression is anchored to the start
4102 match position, and the "anchored" flag is set in the compiled expression.
4103
4104 15. Some bugs concerning the handling of certain option changes within patterns
4105 have been fixed. These applied to options other than (?ims). For example,
4106 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4107 been the other way round. Some of this was related to change 7 above.
4108
4109 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4110 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4111 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4112 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4113
4114 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4115 mode. Unlike ".", it always matches newline, whatever the setting of
4116 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4117 assertions. Perl allows it, but it doesn't (in general) work because it can't
4118 calculate the length of the lookbehind. At least, that's the case for Perl
4119 5.8.0 - I've been told they are going to document that it doesn't work in
4120 future.
4121
4122 18. Added an error diagnosis for escapes that PCRE does not support: these are
4123 \L, \l, \N, \P, \p, \U, \u, and \X.
4124
4125 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4126 reading past the end of the pattern in cases such as /[abcd/.
4127
4128 20. PCRE was getting more memory than necessary for patterns with classes that
4129 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4130
4131 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4132 compiling PCRE for use with Virtual Pascal.
4133
4134 22. Small fix to the Makefile to make it work properly if the build is done
4135 outside the source tree.
4136
4137 23. Added a new extension: a condition to go with recursion. If a conditional
4138 subpattern starts with (?(R) the "true" branch is used if recursion has
4139 happened, whereas the "false" branch is used only at the top level.
4140
4141 24. When there was a very long string of literal characters (over 255 bytes
4142 without UTF support, over 250 bytes with UTF support), the computation of how
4143 much memory was required could be incorrect, leading to segfaults or other
4144 strange effects.
4145
4146 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4147 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4148 there was a subsequent back reference to those brackets. This meant that, for
4149 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4150 possible to check for precisely this case. All we can do is abandon the
4151 optimization if .* occurs inside capturing brackets when there are any back
4152 references whatsoever. (See below for a better fix that came later.)
4153
4154 26. The handling of the optimization for finding the first character of a
4155 non-anchored pattern, and for finding a character that is required later in the
4156 match were failing in some cases. This didn't break the matching; it just
4157 failed to optimize when it could. The way this is done has been re-implemented.
4158
4159 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4160
4161 28. Added a new feature that provides some of the functionality that Perl
4162 provides with (?{...}). The facility is termed a "callout". The way it is done
4163 in PCRE is for the caller to provide an optional function, by setting
4164 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4165 global variable. By default it is unset, which disables all calling out. To get
4166 the function called, the regex must include (?C) at appropriate points. This
4167 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4168 This provides a means of identifying different callout points. When PCRE
4169 reaches such a point in the regex, if pcre_callout has been set, the external
4170 function is called. It is provided with data in a structure called
4171 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4172 matching continues; if it returns a non-zero value, the match at the current
4173 point fails. However, backtracking will occur if possible. [This was changed
4174 later and other features added - see item 49 below.]
4175
4176 29. pcretest is upgraded to test the callout functionality. It provides a
4177 callout function that displays information. By default, it shows the start of
4178 the match and the current position in the text. There are some new data escapes
4179 to vary what happens:
4180
4181 \C+ in addition, show current contents of captured substrings
4182 \C- do not supply a callout function
4183 \C!n return 1 when callout number n is reached
4184 \C!n!m return 1 when callout number n is reached for the mth time
4185
4186 30. If pcregrep was called with the -l option and just a single file name, it
4187 output "<stdin>" if a match was found, instead of the file name.
4188
4189 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4190 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4191 pcre_exec(). This saves a malloc/free per call. The default value of
4192 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4193 when configuring.
4194
4195 32. The default maximum size of a compiled pattern is 64K. There have been a
4196 few cases of people hitting this limit. The code now uses macros to handle the
4197 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4198 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4199 configuring. Tests 2 and 5 work only with 2-byte links because they output
4200 debugging information about compiled patterns.
4201
4202 33. Internal code re-arrangements:
4203
4204 (a) Moved the debugging function for printing out a compiled regex into
4205 its own source file (printint.c) and used #include to pull it into
4206 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4207 separate copies.
4208
4209 (b) Defined the list of op-code names for debugging as a macro in
4210 internal.h so that it is next to the definition of the opcodes.
4211
4212 (c) Defined a table of op-code lengths for simpler skipping along compiled
4213 code. This is again a macro in internal.h so that it is next to the
4214 definition of the opcodes.
4215
4216 34. Added support for recursive calls to individual subpatterns, along the
4217 lines of Robin Houston's patch (but implemented somewhat differently).
4218
4219 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4220 allow it to read and process whole directories in Win32. This code was
4221 contributed by Lionel Fourquaux; it has not been tested by me.
4222
4223 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4224 used to name a group. Names consist of alphanumerics and underscores, and must
4225 be unique. Back references use the syntax (?P=name) and recursive calls use
4226 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4227 numbers. The function pcre_fullinfo() can be used after compilation to extract
4228 a name/number map. There are three relevant calls:
4229
4230 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4231 PCRE_INFO_NAMECOUNT yields the number of entries
4232 PCRE_INFO_NAMETABLE yields a pointer to the map.
4233
4234 The map is a vector of fixed-size entries. The size of each entry depends on
4235 the length of the longest name used. The first two bytes of each entry are the
4236 group number, most significant byte first. There follows the corresponding
4237 name, zero terminated. The names are in alphabetical order.
4238
4239 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4240 case instead of 255. Making it the same both with and without UTF-8 support
4241 means that the same test output works with both.
4242
4243 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4244 calling malloc() with a zero argument.
4245
4246 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4247 optimization. I've improved things by keeping a bitmap of backreferences with
4248 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4249 fact referenced, the optimization can be applied. It is unlikely that a
4250 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4251 the match to follow \n) will appear inside brackets with a number greater than
4252 31, but if it does, any back reference > 31 suppresses the optimization.
4253
4254 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4255 of disabling numbered capturing parentheses. Any opening parenthesis that is
4256 not followed by ? behaves as if it were followed by ?: but named parentheses
4257 can still be used for capturing (and they will acquire numbers in the usual
4258 way).
4259
4260 41. Redesigned the return codes from the match() function into yes/no/error so
4261 that errors can be passed back from deep inside the nested calls. A malloc
4262 failure while inside a recursive subpattern call now causes the
4263 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4264
4265 42. It is now possible to set a limit on the number of times the match()
4266 function is called in a call to pcre_exec(). This facility makes it possible to
4267 limit the amount of recursion and backtracking, though not in a directly
4268 obvious way, because the match() function is used in a number of different
4269 circumstances. The count starts from zero for each position in the subject
4270 string (for non-anchored patterns). The default limit is, for compatibility, a
4271 large number, namely 10 000 000. You can change this in two ways:
4272
4273 (a) When configuring PCRE before making, you can use --with-match-limit=n
4274 to set a default value for the compiled library.
4275
4276 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4277 a different value is set. See 45 below.
4278
4279 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4280
4281 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4282 of things that can be changed at compile time. The first argument specifies
4283 what is wanted and the second points to where the information is to be placed.
4284 The current list of available information is:
4285
4286 PCRE_CONFIG_UTF8
4287
4288 The output is an integer that is set to one if UTF-8 support is available;
4289 otherwise it is set to zero.
4290
4291 PCRE_CONFIG_NEWLINE
4292
4293 The output is an integer that it set to the value of the code that is used for
4294 newline. It is either LF (10) or CR (13).
4295
4296 PCRE_CONFIG_LINK_SIZE
4297
4298 The output is an integer that contains the number of bytes used for internal
4299 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4300
4301 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4302
4303 The output is an integer that contains the threshold above which the POSIX
4304 interface uses malloc() for output vectors. See item 31 above.
4305
4306 PCRE_CONFIG_MATCH_LIMIT
4307
4308 The output is an unsigned integer that contains the default limit of the number
4309 of match() calls in a pcre_exec() execution. See 42 above.
4310
4311 44. pcretest has been upgraded by the addition of the -C option. This causes it
4312 to extract all the available output from the new pcre_config() function, and to
4313 output it. The program then exits immediately.
4314
4315 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4316 order to support additional features. One way would have been to define
4317 pcre_exec2() (for example) with extra arguments, but this would not have been
4318 extensible, and would also have required all calls to the original function to
4319 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4320 is used for passing in "extra" data from pcre_study().
4321
4322 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4323 contains the following fields:
4324
4325 flags a bitmap indicating which of the following fields are set
4326 study_data opaque data from pcre_study()
4327 match_limit a way of specifying a limit on match() calls for a specific
4328 call to pcre_exec()
4329 callout_data data for callouts (see 49 below)
4330
4331 The flag bits are also defined in pcre.h, and are
4332
4333 PCRE_EXTRA_STUDY_DATA
4334 PCRE_EXTRA_MATCH_LIMIT
4335 PCRE_EXTRA_CALLOUT_DATA
4336
4337 The pcre_study() function now returns one of these new pcre_extra blocks, with
4338 the actual study data pointed to by the study_data field, and the
4339 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4340 before. That is, this change is entirely upwards-compatible and requires no
4341 change to existing code.
4342
4343 If you want to pass in additional data to pcre_exec(), you can either place it
4344 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4345 block.
4346
4347 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4348 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4349 times with different match limits, until it finds the minimum value needed for
4350 pcre_exec() to complete. The value is then output. This can be instructive; for
4351 most simple matches the number is quite small, but for pathological cases it
4352 gets very large very quickly.
4353
4354 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4355 returns the size of the data block pointed to by the study_data field in a
4356 pcre_extra block, that is, the value that was passed as the argument to
4357 pcre_malloc() when PCRE was getting memory in which to place the information
4358 created by pcre_study(). The fourth argument should point to a size_t variable.
4359 pcretest has been extended so that this information is shown after a successful
4360 pcre_study() call when information about the compiled regex is being displayed.
4361
4362 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4363 because what follows is always an absolute path. (Later: it turns out that this
4364 is more than cosmetic for MinGW, because it doesn't like empty path
4365 components.)
4366
4367 49. Some changes have been made to the callout feature (see 28 above):
4368
4369 (i) A callout function now has three choices for what it returns:
4370
4371 0 => success, carry on matching
4372 > 0 => failure at this point, but backtrack if possible
4373 < 0 => serious error, return this value from pcre_exec()
4374
4375 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4376 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4377 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4378 use by callout functions. It will never be used by PCRE itself.
4379
4380 (ii) The pcre_extra structure (see 45 above) has a void * field called
4381 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4382 pcre_callout_block structure has a field of the same name. The contents of
4383 the field passed in the pcre_extra structure are passed to the callout
4384 function in the corresponding field in the callout block. This makes it
4385 easier to use the same callout-containing regex from multiple threads. For
4386 testing, the pcretest program has a new data escape
4387
4388 \C*n pass the number n (may be negative) as callout_data
4389
4390 If the callout function in pcretest receives a non-zero value as
4391 callout_data, it returns that value.
4392
4393 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4394 there were some redundant $(CFLAGS) in commands that are now specified as
4395 $(LINK), which already includes $(CFLAGS).
4396
4397 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4398 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4399 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4400 one-byte characters throughout. Note that case-insensitive matching applies
4401 only to characters whose values are less than 256. PCRE doesn't support the
4402 notion of cases for higher-valued characters.
4403
4404 (i) A character class whose characters are all within 0-255 is handled as
4405 a bit map, and the map is inverted for negative classes. Previously, a
4406 character > 255 always failed to match such a class; however it should
4407 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4408
4409 (ii) A negated character class with a single character < 255 is coded as
4410 "not this character" (OP_NOT). This wasn't working properly when the test
4411 character was multibyte, either singly or repeated.
4412
4413 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4414 mode, for example: \x{100}{2,3}.
4415
4416 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4417 singly or repeated) now correctly test multibyte characters. However,
4418 PCRE doesn't recognize any characters with values greater than 255 as
4419 digits, spaces, or word characters. Such characters always match \D, \S,
4420 and \W, and never match \d, \s, or \w.
4421
4422 (v) Classes may now contain characters and character ranges with values
4423 greater than 255. For example: [ab\x{100}-\x{400}].
4424
4425 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4426 PCRE in UTF-8 mode.
4427
4428 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4429 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4430 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4431 value.)
4432
4433 53. The single man page has become too large. I have therefore split it up into
4434 a number of separate man pages. These also give rise to individual HTML pages;
4435 these are now put in a separate directory, and there is an index.html page that
4436 lists them all. Some hyperlinking between the pages has been installed.
4437
4438 54. Added convenience functions for handling named capturing parentheses.
4439
4440 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4441 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4442 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4443 are faulted.
4444
4445 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4446 calling configure. These values are used when compiling the dftables.c program
4447 which is run to generate the source of the default character tables. They
4448 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4449 you will need to set these values.
4450
4451 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4452
4453
4454 Version 3.9 02-Jan-02
4455 ---------------------
4456
4457 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4458
4459 2. If --disable-static was given, the building process failed when trying to
4460 build pcretest and pcregrep. (For some reason it was using libtool to compile
4461 them, which is not right, as they aren't part of the library.)
4462
4463
4464 Version 3.8 18-Dec-01
4465 ---------------------
4466
4467 1. The experimental UTF-8 code was completely screwed up. It was packing the
4468 bytes in the wrong order. How dumb can you get?
4469
4470
4471 Version 3.7 29-Oct-01
4472 ---------------------
4473
4474 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4475 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4476 this didn't happen under Solaris 8, where I normally test things.
4477
4478 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4479 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4480 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4481 this fix an hour or so after the initial 3.7 release.)
4482
4483
4484 Version 3.6 23-Oct-01
4485 ---------------------
4486
4487 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4488 offsets passed as NULL with zero offset count.
4489
4490 2. The config.guess and config.sub files had not been updated when I moved to
4491 the latest autoconf.
4492
4493
4494 Version 3.5 15-Aug-01
4495 ---------------------
4496
4497 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4498 had been forgotten.
4499
4500 2. By using declared but undefined structures, we can avoid using "void"
4501 definitions in pcre.h while keeping the internal definitions of the structures
4502 private.
4503
4504 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4505 user point of view, this means that both static and shared libraries are built
4506 by default, but this can be individually controlled. More of the work of
4507 handling this static/shared cases is now inside libtool instead of PCRE's make
4508 file.
4509
4510 4. The pcretest utility is now installed along with pcregrep because it is
4511 useful for users (to test regexs) and by doing this, it automatically gets
4512 relinked by libtool. The documentation has been turned into a man page, so
4513 there are now .1, .txt, and .html versions in /doc.
4514
4515 5. Upgrades to pcregrep:
4516 (i) Added long-form option names like gnu grep.
4517 (ii) Added --help to list all options with an explanatory phrase.
4518 (iii) Added -r, --recursive to recurse into sub-directories.
4519 (iv) Added -f, --file to read patterns from a file.
4520
4521 6. pcre_exec() was referring to its "code" argument before testing that
4522 argument for NULL (and giving an error if it was NULL).
4523
4524 7. Upgraded Makefile.in to allow for compiling in a different directory from
4525 the source directory.
4526
4527 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4528 options bits, the pointer it was passed was to an int instead of to an unsigned
4529 long int. This mattered only on 64-bit systems.
4530
4531 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4532 generated) instead of pcre.in, which it its source. Also made the same change
4533 in several of the .c files.
4534
4535 10. A new release of gcc defines printf() as a macro, which broke pcretest
4536 because it had an ifdef in the middle of a string argument for printf(). Fixed
4537 by using separate calls to printf().
4538
4539 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4540 script, to force use of CR or LF instead of \n in the source. On non-Unix
4541 systems, the value can be set in config.h.
4542
4543 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4544 absolute limit. Changed the text of the error message to make this clear, and
4545 likewise updated the man page.
4546
4547 13. The limit of 99 on the number of capturing subpatterns has been removed.
4548 The new limit is 65535, which I hope will not be a "real" limit.
4549
4550
4551 Version 3.4 22-Aug-00
4552 ---------------------
4553
4554 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4555
4556 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4557
4558
4559 Version 3.3 01-Aug-00
4560 ---------------------
4561
4562 1. If an octal character was given, but the value was greater than \377, it
4563 was not getting masked to the least significant bits, as documented. This could
4564 lead to crashes in some systems.
4565
4566 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4567 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4568
4569 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4570 These just pass their arguments on to (pcre_free)(), but they are provided
4571 because some uses of PCRE bind it to non-C systems that can call its functions,
4572 but cannot call free() or pcre_free() directly.
4573
4574 4. Add "make test" as a synonym for "make check". Corrected some comments in
4575 the Makefile.
4576
4577 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4578 Makefile.
4579
4580 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4581 command called pgrep for grepping around the active processes.
4582
4583 7. Added the beginnings of support for UTF-8 character strings.
4584
4585 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4586 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4587 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4588 out for the ar command.)
4589
4590
4591 Version 3.2 12-May-00
4592 ---------------------
4593
4594 This is purely a bug fixing release.
4595
4596 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4597 of ZA. This was just one example of several cases that could provoke this bug,
4598 which was introduced by change 9 of version 2.00. The code for breaking
4599 infinite loops after an iteration that matches an empty string was't working
4600 correctly.
4601
4602 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4603 when matched against abbab (for example). After matching an empty string, it
4604 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4605 caused it to match further down the string than it should.
4606
4607 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4608 was there because it doesn't seem to be needed, and it causes trouble on some
4609 systems, as it is not a Standard C header. It has been removed.
4610
4611 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4612 were reported on the Macintosh. The changes were from
4613
4614 while ((c = *(++ptr)) != 0 && c != '\n');
4615 to
4616 while ((c = *(++ptr)) != 0 && c != '\n') ;
4617
4618 Totally extraordinary, but if that's what it takes...
4619
4620 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4621 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4622 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4623 assumes the way PCRE uses memmove() (always moving upwards).
4624
4625 6. PCRE is being used in one environment where strchr() is not available. There
4626 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4627 faster code anyway.
4628
4629
4630 Version 3.1 09-Feb-00
4631 ---------------------
4632
4633 The only change in this release is the fixing of some bugs in Makefile.in for
4634 the "install" target:
4635
4636 (1) It was failing to install pcreposix.h.
4637
4638 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4639
4640
4641 Version 3.0 01-Feb-00
4642 ---------------------
4643
4644 1. Add support for the /+ modifier to perltest (to output $` like it does in
4645 pcretest).
4646
4647 2. Add support for the /g modifier to perltest.
4648
4649 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4650 matches null strings.
4651
4652 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4653 pattern. Perl treats empty patterns specially - it reuses the most recent
4654 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4655 effect.
4656
4657 5. The POSIX interface was broken in that it was just handing over the POSIX
4658 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4659 required a bigger vector, with some working space on the end. This means that
4660 the POSIX wrapper now has to get and free some memory, and copy the results.
4661
4662 6. Added some simple autoconf support, placing the test data and the
4663 documentation in separate directories, re-organizing some of the
4664 information files, and making it build pcre-config (a GNU standard). Also added
4665 libtool support for building PCRE as a shared library, which is now the
4666 default.
4667
4668 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4669 09 are not valid octal constants. Single digits will be used for minor values
4670 less than 10.
4671
4672 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4673 existing programs that set these in the POSIX interface can use PCRE without
4674 modification.
4675
4676 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4677 return all that pcre_info() returns, plus additional data. The pcre_info()
4678 function is retained for compatibility, but is considered to be obsolete.
4679
4680 10. Added experimental recursion feature (?R) to handle one common case that
4681 Perl 5.6 will be able to do with (?p{...}).
4682
4683 11. Added support for POSIX character classes like [:alpha:], which Perl is
4684 adopting.
4685
4686
4687 Version 2.08 31-Aug-99
4688 ----------------------
4689
4690 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4691 trying to match at the startoffset position, but instead was moving forward to
4692 the next newline as if a previous match had failed.
4693
4694 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4695 and could get into a loop if a null string was matched other than at the start
4696 of the subject.
4697
4698 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4699 be distinguished at compile time, and for completeness also added PCRE_DATE.
4700
4701 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4702 in GnuWin32 environments.
4703
4704
4705 Version 2.07 29-Jul-99
4706 ----------------------
4707
4708 1. The documentation is now supplied in plain text form and HTML as well as in
4709 the form of man page sources.
4710
4711 2. C++ compilers don't like assigning (void *) values to other pointer types.
4712 In particular this affects malloc(). Although there is no problem in Standard
4713 C, I've put in casts to keep C++ compilers happy.
4714
4715 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4716 should be (const char *).
4717
4718 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4719 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4720 However, I haven't made this a standard facility. The documentation doesn't
4721 mention it, and the Makefile doesn't support it.
4722
4723 5. The Makefile now contains an "install" target, with editable destinations at
4724 the top of the file. The pcretest program is not installed.
4725
4726 6. pgrep -V now gives the PCRE version number and date.
4727
4728 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4729 causing the entire string to be ignored, instead of just the last character.
4730
4731 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4732 non-matching string, it can take a very, very long time, even for strings of
4733 quite modest length, because of the nested recursion. PCRE now does better in
4734 some of these cases. It does this by remembering the last required literal
4735 character in the pattern, and pre-searching the subject to ensure it is present
4736 before running the real match. In other words, it applies a heuristic to detect
4737 some types of certain failure quickly, and in the above example, if presented
4738 with a string that has no trailing " it gives "no match" very quickly.
4739
4740 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4741 other alternatives are tried instead.
4742
4743
4744 Version 2.06 09-Jun-99
4745 ----------------------
4746
4747 1. Change pcretest's output for amount of store used to show just the code
4748 space, because the remainder (the data block) varies in size between 32-bit and
4749 64-bit systems.
4750
4751 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4752 start matching at. This allows lookbehinds to work when searching for multiple
4753 occurrences in a string.
4754
4755 3. Added additional options to pcretest for testing multiple occurrences:
4756
4757 /+ outputs the rest of the string that follows a match
4758 /g loops for multiple occurrences, using the new startoffset argument
4759 /G loops for multiple occurrences by passing an incremented pointer
4760
4761 4. PCRE wasn't doing the "first character" optimization for patterns starting
4762 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4763 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4764 the letter 'x'. On long subject strings, this gives a significant speed-up.
4765
4766
4767 Version 2.05 21-Apr-99
4768 ----------------------
4769
4770 1. Changed the type of magic_number from int to long int so that it works
4771 properly on 16-bit systems.
4772
4773 2. Fixed a bug which caused patterns starting with .* not to work correctly
4774 when the subject string contained newline characters. PCRE was assuming
4775 anchoring for such patterns in all cases, which is not correct because .* will
4776 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4777 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4778 must be retried after every newline in the subject.
4779
4780
4781 Version 2.04 18-Feb-99
4782 ----------------------
4783
4784 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4785 computation of the store needed to hold the pattern was incorrect (too large).
4786 If such patterns were nested a few deep, this could multiply and become a real
4787 problem.
4788
4789 2. Added /M option to pcretest to show the memory requirement of a specific
4790 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4791
4792 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4793 compiled in such a way that the backtracking after subsequent failure was
4794 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4795 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4796
4797
4798 Version 2.03 02-Feb-99
4799 ----------------------
4800
4801 1. Fixed typo and small mistake in man page.
4802
4803 2. Added 4th condition (GPL supersedes if conflict) and created separate
4804 LICENCE file containing the conditions.
4805
4806 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4807 Perl, that is the internal \ allows the delimiter to be included in the
4808 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4809 the final delimiter, add \ to the end of the pattern (to test the error).
4810
4811 4. Added the convenience functions for extracting substrings after a successful
4812 match. Updated pcretest to make it able to test these functions.
4813
4814
4815 Version 2.02 14-Jan-99
4816 ----------------------
4817
4818 1. Initialized the working variables associated with each extraction so that
4819 their saving and restoring doesn't refer to uninitialized store.
4820
4821 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4822 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4823 fix the problem.
4824
4825 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4826 calls, and wasn't printing the correct value for compiling calls. Increased the
4827 default value of LOOPREPEAT, and the number of significant figures in the
4828 times.
4829
4830 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4831
4832 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4833 a building problem on Windows NT with a FAT file system.
4834
4835
4836 Version 2.01 21-Oct-98
4837 ----------------------
4838
4839 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4840 to character tables built by pcre_maketables() in the current locale. If NULL
4841 is passed, the default tables are used.
4842
4843
4844 Version 2.00 24-Sep-98
4845 ----------------------
4846
4847 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4848 it any more.
4849
4850 2. Allow quantification of (?>) groups, and make it work correctly.
4851
4852 3. The first character computation wasn't working for (?>) groups.
4853
4854 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4855 end of the subject) and add 5.005's \z, which really does match only at the
4856 very end of the subject.
4857
4858 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4859
4860 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4861 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4862 localized options. All options to pcre_study() were also removed.
4863
4864 7. Add other new features from 5.005:
4865
4866 $(?<= positive lookbehind
4867 $(?<! negative lookbehind
4868 (?imsx-imsx) added the unsetting capability
4869 such a setting is global if at outer level; local otherwise
4870 (?imsx-imsx:) non-capturing groups with option setting
4871 (?(cond)re|re) conditional pattern matching
4872
4873 A backreference to itself in a repeated group matches the previous
4874 captured string.
4875
4876 8. General tidying up of studying (both automatic and via "study")
4877 consequential on the addition of new assertions.
4878
4879 9. As in 5.005, unlimited repeated groups that could match an empty substring
4880 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4881 runtime if any iteration does actually match an empty substring.
4882
4883 10. Include the RunTest script in the distribution.
4884
4885 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4886 discrepancies, some of which were old and were also with respect to 5.004. They
4887 have now been fixed.
4888
4889
4890 Version 1.09 28-Apr-98
4891 ----------------------
4892
4893 1. A negated single character class followed by a quantifier with a minimum
4894 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4895 program crashes, or just wrong answers. This did not apply to negated classes
4896 containing more than one character, or to minima other than one.
4897
4898
4899 Version 1.08 27-Mar-98
4900 ----------------------
4901
4902 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4903
4904 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4905 latter must appear before anything that relies on it in the pattern.
4906
4907
4908 Version 1.07 16-Feb-98
4909 ----------------------
4910
4911 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4912 repeat of a potentially empty string).
4913
4914
4915 Version 1.06 23-Jan-98
4916 ----------------------
4917
4918 1. Added Markus Oberhumer's little patches for C++.
4919
4920 2. Literal strings longer than 255 characters were broken.
4921
4922
4923 Version 1.05 23-Dec-97
4924 ----------------------
4925
4926 1. Negated character classes containing more than one character were failing if
4927 PCRE_CASELESS was set at run time.
4928
4929
4930 Version 1.04 19-Dec-97
4931 ----------------------
4932
4933 1. Corrected the man page, where some "const" qualifiers had been omitted.
4934
4935 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4936 input syntax.
4937
4938 3. Fixed memory leak which occurred when a regex with back references was
4939 matched with an offsets vector that wasn't big enough. The temporary memory
4940 that is used in this case wasn't being freed if the match failed.
4941
4942 4. Tidied pcretest to ensure it frees memory that it gets.
4943
4944 5. Temporary memory was being obtained in the case where the passed offsets
4945 vector was exactly big enough.
4946
4947 6. Corrected definition of offsetof() from change 5 below.
4948
4949 7. I had screwed up change 6 below and broken the rules for the use of
4950 setjmp(). Now fixed.
4951
4952
4953 Version 1.03 18-Dec-97
4954 ----------------------
4955
4956 1. A erroneous regex with a missing opening parenthesis was correctly
4957 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4958 on some systems.
4959
4960 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4961 it was reported that one broken compiler failed on the former because "code" is
4962 also an independent variable.
4963
4964 3. The erroneous regex a[]b caused an array overrun reference.
4965
4966 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4967 fail on data ending with that character. (It was going on too far, and checking
4968 the next character, typically a binary zero.) This was specific to the
4969 optimized code for single-character negative classes.
4970
4971 5. Added a contributed patch from the TIN world which does the following:
4972
4973 + Add an undef for memmove, in case the the system defines a macro for it.
4974
4975 + Add a definition of offsetof(), in case there isn't one. (I don't know
4976 the reason behind this - offsetof() is part of the ANSI standard - but
4977 it does no harm).
4978
4979 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4980 most of the places where whitespace preceded '#'. I have given up and
4981 allowed the remaining 2 cases to be at the margin.
4982
4983 + Rename some variables in pcre to eliminate shadowing. This seems very
4984 pedantic, but does no harm, of course.
4985
4986 6. Moved the call to setjmp() into its own function, to get rid of warnings
4987 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4988
4989 7. Constructs such as \d{8,} were compiling into the equivalent of
4990 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4991 outcome, but in this particular case used more store than had been allocated,
4992 which caused the bug to be discovered because it threw up an internal error.
4993
4994 8. The debugging code in both pcre and pcretest for outputting the compiled
4995 form of a regex was going wrong in the case of back references followed by
4996 curly-bracketed repeats.
4997
4998
4999 Version 1.02 12-Dec-97
5000 ----------------------
5001
5002 1. Typos in pcre.3 and comments in the source fixed.
5003
5004 2. Applied a contributed patch to get rid of places where it used to remove
5005 'const' from variables, and fixed some signed/unsigned and uninitialized
5006 variable warnings.
5007
5008 3. Added the "runtest" target to Makefile.
5009
5010 4. Set default compiler flag to -O2 rather than just -O.
5011
5012
5013 Version 1.01 19-Nov-97
5014 ----------------------
5015
5016 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5017 like /([ab]*)*/, that is, for classes with more than one character in them.
5018
5019 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5020 as /((?>a*))*/ (a PCRE_EXTRA facility).
5021
5022
5023 Version 1.00 18-Nov-97
5024 ----------------------
5025
5026 1. Added compile-time macros to support systems such as SunOS4 which don't have
5027 memmove() or strerror() but have other things that can be used instead.
5028
5029 2. Arranged that "make clean" removes the executables.
5030
5031
5032 Version 0.99 27-Oct-97
5033 ----------------------
5034
5035 1. Fixed bug in code for optimizing classes with only one character. It was
5036 initializing a 32-byte map regardless, which could cause it to run off the end
5037 of the memory it had got.
5038
5039 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5040
5041
5042 Version 0.98 22-Oct-97
5043 ----------------------
5044
5045 1. Fixed bug in code for handling temporary memory usage when there are more
5046 back references than supplied space in the ovector. This could cause segfaults.
5047
5048
5049 Version 0.97 21-Oct-97
5050 ----------------------
5051
5052 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5053
5054 2. Optimized negated single characters not to use a bit map.
5055
5056 3. Brought error texts together as macro definitions; clarified some of them;
5057 fixed one that was wrong - it said "range out of order" when it meant "invalid
5058 escape sequence".
5059
5060 4. Changed some char * arguments to const char *.
5061
5062 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5063
5064 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5065 pcretest.
5066
5067
5068 Version 0.96 16-Oct-97
5069 ----------------------
5070
5071 1. Added a simple "pgrep" utility to the distribution.
5072
5073 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5074 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5075 where "ddd" means "one or more decimal digits".
5076
5077 3. Fixed serious bug. If a pattern had a back reference, but the call to
5078 pcre_exec() didn't supply a large enough ovector to record the related
5079 identifying subpattern, the match always failed. PCRE now remembers the number
5080 of the largest back reference, and gets some temporary memory in which to save
5081 the offsets during matching if necessary, in order to ensure that
5082 backreferences always work.
5083
5084 4. Increased the compatibility with Perl in a number of ways:
5085
5086 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5087 to request this handling. The option can be set at compile or exec time.
5088
5089 (b) $ matches before a terminating newline by default; an option
5090 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5091 mode). The option can be set at compile or exec time.
5092
5093 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5094 the same as Perl's. If the decimal number it represents is less than 10
5095 or there aren't that many previous left capturing parentheses, an octal
5096 escape is read. Inside a character class, it's always an octal escape,
5097 even if it is a single digit.
5098
5099 (d) An escaped but undefined alphabetic character is taken as a literal,
5100 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5101 escapes.
5102
5103 (e) {0} is now permitted. (The previous item is removed from the compiled
5104 pattern).
5105
5106 5. Changed all the names of code files so that the basic parts are no longer
5107 than 10 characters, and abolished the teeny "globals.c" file.
5108
5109 6. Changed the handling of character classes; they are now done with a 32-byte
5110 bit map always.
5111
5112 7. Added the -d and /D options to pcretest to make it possible to look at the
5113 internals of compilation without having to recompile pcre.
5114
5115
5116 Version 0.95 23-Sep-97
5117 ----------------------
5118
5119 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5120 \x20 at the start of a run of normal characters. These were being treated as
5121 real characters, instead of the source characters being re-checked.
5122
5123
5124 Version 0.94 18-Sep-97
5125 ----------------------
5126
5127 1. The functions are now thread-safe, with the caveat that the global variables
5128 containing pointers to malloc() and free() or alternative functions are the
5129 same for all threads.
5130
5131 2. Get pcre_study() to generate a bitmap of initial characters for non-
5132 anchored patterns when this is possible, and use it if passed to pcre_exec().
5133
5134
5135 Version 0.93 15-Sep-97
5136 ----------------------
5137
5138 1. /(b)|(:+)/ was computing an incorrect first character.
5139
5140 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5141 but not actually doing anything yet.
5142
5143 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5144 as Perl does (e.g. [-az] or [az-]).
5145
5146 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5147 all possible positions.
5148
5149 5. Split up into different modules to avoid including unneeded functions in a
5150 compiled binary. However, compile and exec are still in one module. The "study"
5151 function is split off.
5152
5153 6. The character tables are now in a separate module whose source is generated
5154 by an auxiliary program - but can then be edited by hand if required. There are
5155 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5156 toupper() in the code.
5157
5158 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5159 make them global. Abolish the function for setting them, as the caller can now
5160 set them directly.
5161
5162
5163 Version 0.92 11-Sep-97
5164 ----------------------
5165
5166 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5167 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5168
5169 2. Caseless matching was not working in character classes if the characters in
5170 the pattern were in upper case.
5171
5172 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5173
5174 4. Make PCRE_ANCHORED public and accept as a compile option.
5175
5176 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5177 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5178 pass them.
5179
5180 6. Give an error if bad option bits passed at compile or run time.
5181
5182 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5183 pcretest to cause it to pass that flag.
5184
5185 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5186 options, and the first character, if set.
5187
5188 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5189
5190
5191 Version 0.91 10-Sep-97
5192 ----------------------
5193
5194 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5195 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5196
5197 2. PCRE was looping on encountering an indefinitely repeated back reference to
5198 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5199 Perl does - treats the match as successful.
5200
5201 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5