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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1353 - (show annotations)
Mon Aug 5 16:24:02 2013 UTC (6 years, 3 months ago) by ph10
File size: 250628 byte(s)
Fix pcregrep looping bug for multiline empty string match.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.34 xx-xxxx-201x
5 --------------------------
6
7 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8 executable memory. Patch inspired by Carsten Klein.
9
10 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11 this macro is never tested and has no effect, because the work to support
12 coverage involves only compiling and linking options and special targets in
13 the Makefile. The comment in config.h implied that defining the macro would
14 enable coverage support, which is totally false. There was also support for
15 setting this macro in the CMake files (my fault, I just copied it from
16 configure). SUPPORT_GCOV has now been removed.
17
18 3. Make a small performance improvement in strlen16() and strlen32() in
19 pcretest.
20
21 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23 these cases). There was also one in pcretest.c.
24
25 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
26
27 6. In UTF mode, the code for checking whether a group could match an empty
28 string (which is used for indefinitely repeated groups to allow for
29 breaking an infinite loop) was broken when the group contained a repeated
30 negated single-character class with a character that occupied more than one
31 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
32 UTF-8 mode). The effect was undefined: the group might or might not be
33 deemed as matching an empty string, or the program might have crashed.
34
35 7. The code for checking whether a group could match an empty string was not
36 recognizing that \h, \H, \v, \V, and \R must match a character.
37
38 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
39 an empty string. If it can, pcretest shows this in its information output.
40
41 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
42 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
43 matched by pcre_exec() without using JIT:
44
45 (a) If the rest of the pattern did not match after a maximal run of
46 grapheme clusters, the code for backing up to try with fewer of them
47 did not always back up over a full grapheme when characters that do not
48 have the modifier quality were involved, e.g. Hangul syllables.
49
50 (b) If the match point in a subject started with modifier character, and
51 there was no match, the code could incorrectly back up beyond the match
52 point, and potentially beyond the first character in the subject,
53 leading to a segfault or an incorrect match result.
54
55 10. A conditional group with an assertion condition could lead to PCRE
56 recording an incorrect first data item for a match if no other firse data
57 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58 first data item, and therefore matched "ca" after "c" instead of at the
59 start.
60
61 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
62 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
63
64
65 Version 8.33 28-May-2013
66 --------------------------
67
68 1. Added 'U' to some constants that are compared to unsigned integers, to
69 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
70 variables that are added to signed variables, to ensure the result is
71 signed and can be negated.
72
73 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
74 CMake config files.
75
76 3. Revise the creation of config.h.generic so that all boolean macros are
77 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
78 overriding via -D on the command line possible.
79
80 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
81 to unsigned int is reported to make a quite noticeable speed difference in
82 a specific Windows environment. Testing on Linux did also appear to show
83 some benefit (and it is clearly not harmful). Also fixed the definition of
84 Xop which should be unsigned.
85
86 5. Related to (4), changing the definition of the intermediate variable cc
87 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
88 improvements.
89
90 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
91 unnecessary spaces.
92
93 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
94 and later.
95
96 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
97
98 9. Optimizing fast_forward_start_bits in JIT.
99
100 10. Adding support for callouts in JIT, and fixing some issues revealed
101 during this work. Namely:
102
103 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
104
105 (b) Minimum length was not checked before the matching is started.
106
107 11. The value of capture_last that is passed to callouts was incorrect in some
108 cases when there was a capture on one path that was subsequently abandoned
109 after a backtrack. Also, the capture_last value is now reset after a
110 recursion, since all captures are also reset in this case.
111
112 12. The interpreter no longer returns the "too many substrings" error in the
113 case when an overflowing capture is in a branch that is subsequently
114 abandoned after a backtrack.
115
116 13. In the pathological case when an offset vector of size 2 is used, pcretest
117 now prints out the matched string after a yield of 0 or 1.
118
119 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
120 Only supported by the JIT compiler at the moment.
121
122 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
123
124 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
125 offset of the starting point of the matching process, provided the offsets
126 vector is large enough.
127
128 17. The \A escape now records a lookbehind value of 1, though its execution
129 does not actually inspect the previous character. This is to ensure that,
130 in partial multi-segment matching, at least one character from the old
131 segment is retained when a new segment is processed. Otherwise, if there
132 are no lookbehinds in the pattern, \A might match incorrectly at the start
133 of a new segment.
134
135 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
136
137 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
138 gives some modest performance improvement in 8-bit mode.
139
140 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
141 be expressed in certain programming languages using Universal Character
142 Names.
143
144 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
145 which points out that "non characters" are not "characters that may not
146 appear in Unicode strings" but rather "characters that are reserved for
147 internal use and have only local meaning".
148
149 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
150 there was a conditional group that depended on an assertion, if the
151 assertion was false, the callout that immediately followed the alternation
152 in the condition was skipped when pcre_exec() was used for matching.
153
154 23. Allow an explicit callout to be inserted before an assertion that is the
155 condition for a conditional group, for compatibility with automatic
156 callouts, which always insert a callout at this point.
157
158 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
159 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
160
161 25. (*PRUNE) is now supported by the JIT compiler.
162
163 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
164
165 27. Fix the case where there are two or more SKIPs with arguments that may be
166 ignored.
167
168 28. (*SKIP) is now supported by the JIT compiler.
169
170 29. (*THEN) is now supported by the JIT compiler.
171
172 30. Update RunTest with additional test selector options.
173
174 31. The way PCRE handles backtracking verbs has been changed in two ways.
175
176 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
177 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
178 backtracking. In some cases this makes it more Perl-compatible, but Perl's
179 rather obscure rules do not always do the same thing.
180
181 (2) Previously, backtracking verbs were confined within assertions. This is
182 no longer the case for positive assertions, except for (*ACCEPT). Again,
183 this sometimes improves Perl compatibility, and sometimes does not.
184
185 32. A number of tests that were in test 2 because Perl did things differently
186 have been moved to test 1, because either Perl or PCRE has changed, and
187 these tests are now compatible.
188
189 32. Backtracking control verbs are now handled in the same way in JIT and
190 interpreter.
191
192 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
193 contained a forward subroutine reference caused a compile error.
194
195 34. Auto-detect and optimize limited repetitions in JIT.
196
197 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
198 blocking (*UTF) etc.
199
200 36. In the interpreter, maximizing pattern repetitions for characters and
201 character types now use tail recursion, which reduces stack usage.
202
203 37. The value of the max lookbehind was not correctly preserved if a compiled
204 and saved regex was reloaded on a host of different endianness.
205
206 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
207 of the compiled pattern block, expand the flags field from 16 to 32 bits
208 because it was almost full.
209
210 39. Try madvise first before posix_madvise.
211
212 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
213 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
214 original change locked out all matches of empty strings. This has been
215 changed so that one match of an empty string per line is recognized.
216 Subsequent searches on the same line (for colouring or for --only-matching,
217 for example) do not recognize empty strings.
218
219 41. Applied a user patch to fix a number of spelling mistakes in comments.
220
221 42. Data lines longer than 65536 caused pcretest to crash.
222
223 43. Clarified the data type for length and startoffset arguments for pcre_exec
224 and pcre_dfa_exec in the function-specific man pages, where they were
225 explicitly stated to be in bytes, never having been updated. I also added
226 some clarification to the pcreapi man page.
227
228 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
229 a segmentation fault.
230
231
232 Version 8.32 30-November-2012
233 -----------------------------
234
235 1. Improved JIT compiler optimizations for first character search and single
236 character iterators.
237
238 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
239 Patch by Daniel Richard G.
240
241 3. Single character iterator optimizations in the JIT compiler.
242
243 4. Improved JIT compiler optimizations for character ranges.
244
245 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
246 Reported by Giuseppe D'Angelo.
247
248 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
249 of a line, was being set incorrectly in cases where .* appeared inside
250 atomic brackets at the start of a pattern, or where there was a subsequent
251 *PRUNE or *SKIP.
252
253 7. Improved instruction cache flush for POWER/PowerPC.
254 Patch by Daniel Richard G.
255
256 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
257 grep:
258
259 (a) There is now no limit to the number of patterns to be matched.
260
261 (b) An error is given if a pattern is too long.
262
263 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
264 are now supported.
265
266 (d) --exclude-from and --include-from (multiple use) have been added.
267
268 (e) Exclusions and inclusions now apply to all files and directories, not
269 just to those obtained from scanning a directory recursively.
270
271 (f) Multiple uses of -f and --file-list are now supported.
272
273 (g) In a Windows environment, the default for -d has been changed from
274 "read" (the GNU grep default) to "skip", because otherwise the presence
275 of a directory in the file list provokes an error.
276
277 (h) The documentation has been revised and clarified in places.
278
279 9. Improve the matching speed of capturing brackets.
280
281 10. Changed the meaning of \X so that it now matches a Unicode extended
282 grapheme cluster.
283
284 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
285 out POSIX threads when JIT support is configured.
286
287 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
288
289 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
290 structure could go wrong in environments where size_t is not the same size
291 as int.
292
293 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
294 set.
295
296 15. The EBCDIC support had decayed; later updates to the code had included
297 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
298 general tidy up of EBCDIC-related issues, and the documentation was also
299 not quite right. There is now a test that can be run on ASCII systems to
300 check some of the EBCDIC-related things (but is it not a full test).
301
302 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
303 in a small tidy to the code.
304
305 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
306
307 18. If the --only-matching (-o) option in pcregrep is specified multiple
308 times, each one causes appropriate output. For example, -o1 -o2 outputs the
309 substrings matched by the 1st and 2nd capturing parentheses. A separating
310 string can be specified by --om-separator (default empty).
311
312 19. Improving the first n character searches.
313
314 20. Turn case lists for horizontal and vertical white space into macros so that
315 they are defined only once.
316
317 21. This set of changes together give more compatible Unicode case-folding
318 behaviour for characters that have more than one other case when UCP
319 support is available.
320
321 (a) The Unicode property table now has offsets into a new table of sets of
322 three or more characters that are case-equivalent. The MultiStage2.py
323 script that generates these tables (the pcre_ucd.c file) now scans
324 CaseFolding.txt instead of UnicodeData.txt for character case
325 information.
326
327 (b) The code for adding characters or ranges of characters to a character
328 class has been abstracted into a generalized function that also handles
329 case-independence. In UTF-mode with UCP support, this uses the new data
330 to handle characters with more than one other case.
331
332 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
333 whose other case is greater than 256 are now correctly matched
334 caselessly. Previously, the high codepoint matched the low one, but not
335 vice versa.
336
337 (d) The processing of \h, \H, \v, and \ in character classes now makes use
338 of the new class addition function, using character lists defined as
339 macros alongside the case definitions of 20 above.
340
341 (e) Caseless back references now work with characters that have more than
342 one other case.
343
344 (f) General caseless matching of characters with more than one other case
345 is supported.
346
347 22. Unicode character properties were updated from Unicode 6.2.0
348
349 23. Improved CMake support under Windows. Patch by Daniel Richard G.
350
351 24. Add support for 32-bit character strings, and UTF-32
352
353 25. Major JIT compiler update (code refactoring and bugfixing).
354 Experimental Sparc 32 support is added.
355
356 26. Applied a modified version of Daniel Richard G's patch to create
357 pcre.h.generic and config.h.generic by "make" instead of in the
358 PrepareRelease script.
359
360 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
361 pcre_compile.c when checking for a zero character.
362
363 28. Introducing a native interface for JIT. Through this interface, the compiled
364 machine code can be directly executed. The purpose of this interface is to
365 provide fast pattern matching, so several sanity checks are not performed.
366 However, feature tests are still performed. The new interface provides
367 1.4x speedup compared to the old one.
368
369 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
370 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
371 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
372
373 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
374 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
375 the "old" RFC 2279). Instead, it ended up passing a negative length to
376 pcre_exec().
377
378 31. Add support for GCC's visibility feature to hide internal functions.
379
380 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
381 "unknown -C option" after outputting 0 or 1.
382
383 33. There is now support for generating a code coverage report for the test
384 suite in environments where gcc is the compiler and lcov is installed. This
385 is mainly for the benefit of the developers.
386
387 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
388 unaddressable using valgrind annotations, allowing valgrind to detect
389 invalid memory accesses. This is mainly for the benefit of the developers.
390
391 25. (*UTF) can now be used to start a pattern in any of the three libraries.
392
393 26. Give configure error if --enable-cpp but no C++ compiler found.
394
395
396 Version 8.31 06-July-2012
397 -------------------------
398
399 1. Fixing a wrong JIT test case and some compiler warnings.
400
401 2. Removed a bashism from the RunTest script.
402
403 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
404 to unsigned type, result still unsigned" that was given by an MS compiler
405 on encountering the code "-sizeof(xxx)".
406
407 4. Partial matching support is added to the JIT compiler.
408
409 5. Fixed several bugs concerned with partial matching of items that consist
410 of more than one character:
411
412 (a) /^(..)\1/ did not partially match "aba" because checking references was
413 done on an "all or nothing" basis. This also applied to repeated
414 references.
415
416 (b) \R did not give a hard partial match if \r was found at the end of the
417 subject.
418
419 (c) \X did not give a hard partial match after matching one or more
420 characters at the end of the subject.
421
422 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
423 a partial match for the string "\r".
424
425 (e) When newline was set to CRLF, the metacharacter "." did not recognize
426 a partial match for a CR character at the end of the subject string.
427
428 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
429 running pcretest, the text "(JIT)" added to the output whenever JIT is
430 actually used to run the match.
431
432 7. Individual JIT compile options can be set in pcretest by following -s+[+]
433 or /S+[+] with a digit between 1 and 7.
434
435 8. OP_NOT now supports any UTF character not just single-byte ones.
436
437 9. (*MARK) control verb is now supported by the JIT compiler.
438
439 10. The command "./RunTest list" lists the available tests without actually
440 running any of them. (Because I keep forgetting what they all are.)
441
442 11. Add PCRE_INFO_MAXLOOKBEHIND.
443
444 12. Applied a (slightly modified) user-supplied patch that improves performance
445 when the heap is used for recursion (compiled with --disable-stack-for-
446 recursion). Instead of malloc and free for each heap frame each time a
447 logical recursion happens, frames are retained on a chain and re-used where
448 possible. This sometimes gives as much as 30% improvement.
449
450 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
451 call.
452
453 14. As documented, (*COMMIT) is now confined to within a positive assertion.
454
455 15. It is now possible to link pcretest with libedit as an alternative to
456 libreadline.
457
458 16. (*COMMIT) control verb is now supported by the JIT compiler.
459
460 17. The Unicode data tables have been updated to Unicode 6.1.0.
461
462 18. Added --file-list option to pcregrep.
463
464 19. Added binary file support to pcregrep, including the -a, --binary-files,
465 -I, and --text options.
466
467 20. The madvise function is renamed for posix_madvise for QNX compatibility
468 reasons. Fixed by Giuseppe D'Angelo.
469
470 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
471
472 22. Changed the option for creating symbolic links for 16-bit man pages from
473 -s to -sf so that re-installing does not cause issues.
474
475 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
476
477 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
478 matches in certain environments (the workspace was not being correctly
479 retained). Also added to pcre_dfa_exec() a simple plausibility check on
480 some of the workspace data at the beginning of a restart.
481
482 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
483 was not doing so when it should - probably a typo introduced by SVN 528
484 (change 8.10/14).
485
486 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
487 \w+ when the character tables indicated that \x{c4} was a word character.
488 There were several related cases, all because the tests for doing a table
489 lookup were testing for characters less than 127 instead of 255.
490
491 27. If a pattern contains capturing parentheses that are not used in a match,
492 their slots in the ovector are set to -1. For those that are higher than
493 any matched groups, this happens at the end of processing. In the case when
494 there were back references that the ovector was too small to contain
495 (causing temporary malloc'd memory to be used during matching), and the
496 highest capturing number was not used, memory off the end of the ovector
497 was incorrectly being set to -1. (It was using the size of the temporary
498 memory instead of the true size.)
499
500 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
501 ovector size, it uses memory at the end of the block that it has got.
502
503 29. Check for an overlong MARK name and give an error at compile time. The
504 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
505
506 30. JIT compiler update.
507
508 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
509 Rill for the patch.
510
511 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
512
513 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
514
515 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
516 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
517 was enabled.
518
519 35. Improve JIT code generation for greedy plus quantifier.
520
521 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
522 1 to "aa" instead of to an empty string. The bug affected repeated groups
523 that could potentially match an empty string.
524
525 37. Optimizing single character iterators in JIT.
526
527 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
528 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
529 codepoints that are too big for the mode are faulted, and in a UTF mode,
530 disallowed codepoints are also faulted.
531
532 39. If PCRE was compiled with UTF support, in three places in the DFA
533 matcher there was code that should only have been obeyed in UTF mode, but
534 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
535 processing when bytes with values greater than 127 were present. In 16-bit
536 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
537 both cases the values are those that cannot be the first data item in a UTF
538 character. The three items that might have provoked this were recursions,
539 possessively repeated groups, and atomic groups.
540
541 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
542 and pcregrep, because some OS require shared objects to be explicitly
543 passed to ld, causing the link step to fail if they are not.
544
545 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
546 mode, patterns that started with \h* or \R* might be incorrectly matched.
547
548
549 Version 8.30 04-February-2012
550 -----------------------------
551
552 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
553 name is defined in ctype.h.
554
555 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
556 only in quite long subpatterns.
557
558 3. Removed the function pcre_info(), which has been obsolete and deprecated
559 since it was replaced by pcre_fullinfo() in February 2000.
560
561 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
562 match a (*MARK), and the match failed at the start of the subject, a
563 reference to memory before the start of the subject could occur. This bug
564 was introduced by fix 17 of release 8.21.
565
566 5. A reference to an unset group with zero minimum repetition was giving
567 totally wrong answers (in non-JavaScript-compatibility mode). For example,
568 /(another)?(\1?)test/ matched against "hello world test". This bug was
569 introduced in release 8.13.
570
571 6. Add support for 16-bit character strings (a large amount of work involving
572 many changes and refactorings).
573
574 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
575 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
576 from a file.
577
578 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
579 rounding is not applied in this particular case).
580
581 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
582 if they appear, or are escaped, in patterns.
583
584 10. Get rid of a number of -Wunused-but-set-variable warnings.
585
586 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
587 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
588 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
589 also returns the mark "x". This bug applied to capturing parentheses,
590 non-capturing parentheses, and atomic parentheses. It also applied to some
591 assertions.
592
593 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
594 information out of configure.ac instead of relying on pcre.h.generic, which
595 is not stored in the repository.
596
597 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
598 -lreadline.
599
600 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
601
602 15. Applied Graycode's patch to put the top-level frame on the stack rather
603 than the heap when not using the stack for recursion. This gives a
604 performance improvement in many cases when recursion is not deep.
605
606 16. Experimental code added to "pcretest -C" to output the stack frame size.
607
608
609 Version 8.21 12-Dec-2011
610 ------------------------
611
612 1. Updating the JIT compiler.
613
614 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
615 are added as well.
616
617 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
618 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
619 calling _pcre_jit_exec. Some extra comments are added.
620
621 4. (*MARK) settings inside atomic groups that do not contain any capturing
622 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
623 was introduced by change 18 for 8.20.
624
625 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
626 ECMA-262 standard.
627
628 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
629 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
630 This bug was probably introduced by change 9 of 8.13.
631
632 7. While fixing 6 above, I noticed that a number of other items were being
633 incorrectly rejected as "not fixed length". This arose partly because newer
634 opcodes had not been added to the fixed-length checking code. I have (a)
635 corrected the bug and added tests for these items, and (b) arranged for an
636 error to occur if an unknown opcode is encountered while checking for fixed
637 length instead of just assuming "not fixed length". The items that were
638 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
639 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
640 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
641
642 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
643 being incorrectly compiled and would have given unpredicatble results.
644
645 9. A possessively repeated subpattern with minimum repeat count greater than
646 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
647 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
648 the first (A) could occur when it should not.
649
650 10. Add a cast and remove a redundant test from the code.
651
652 11. JIT should use pcre_malloc/pcre_free for allocation.
653
654 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
655 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
656 is anything other than /usr, -L is still shown).
657
658 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
659
660 14. Perl does not support \N without a following name in a [] class; PCRE now
661 also gives an error.
662
663 15. If a forward reference was repeated with an upper limit of around 2000,
664 it caused the error "internal error: overran compiling workspace". The
665 maximum number of forward references (including repeats) was limited by the
666 internal workspace, and dependent on the LINK_SIZE. The code has been
667 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
668 the default depends on LINK_SIZE. There is a new upper limit (for safety)
669 of around 200,000 forward references. While doing this, I also speeded up
670 the filling in of repeated forward references.
671
672 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
673 incorrectly expecting the subject to contain another "a" after the start.
674
675 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
676 in the match, the SKIP should be ignored. This was not happening; instead
677 the SKIP was being treated as NOMATCH. For patterns such as
678 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
679 tested.
680
681 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
682 now much more compatible with Perl, in particular in cases where the result
683 is a non-match for a non-anchored pattern. For example, if
684 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
685 "m", where previously it did not return a name. A side effect of this
686 change is that for partial matches, the last encountered mark name is
687 returned, as for non matches. A number of tests that were previously not
688 Perl-compatible have been moved into the Perl-compatible test files. The
689 refactoring has had the pleasing side effect of removing one argument from
690 the match() function, thus reducing its stack requirements.
691
692 19. If the /S+ option was used in pcretest to study a pattern using JIT,
693 subsequent uses of /S (without +) incorrectly behaved like /S+.
694
695 21. Retrieve executable code size support for the JIT compiler and fixing
696 some warnings.
697
698 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
699 not work when the shorter character appeared right at the end of the
700 subject string.
701
702 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
703 systems.
704
705 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
706 output it when the /M option is used in pcretest.
707
708 25. The CheckMan script was not being included in the distribution. Also, added
709 an explicit "perl" to run Perl scripts from the PrepareRelease script
710 because this is reportedly needed in Windows.
711
712 26. If study data was being save in a file and studying had not found a set of
713 "starts with" bytes for the pattern, the data written to the file (though
714 never used) was taken from uninitialized memory and so caused valgrind to
715 complain.
716
717 27. Updated RunTest.bat as provided by Sheri Pierce.
718
719 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
720
721 29. Computation of memory usage for the table of capturing group names was
722 giving an unnecessarily large value.
723
724
725 Version 8.20 21-Oct-2011
726 ------------------------
727
728 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
729 a POSIX class. After further experiments with Perl, which convinced me that
730 Perl has bugs and confusions, a closing square bracket is no longer allowed
731 in a POSIX name. This bug also affected patterns with classes that started
732 with full stops.
733
734 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
735 captured substring, but while checking the failing first alternative,
736 substring 1 is temporarily captured. If the output vector supplied to
737 pcre_exec() was not big enough for this capture, the yield of the function
738 was still zero ("insufficient space for captured substrings"). This cannot
739 be totally fixed without adding another stack variable, which seems a lot
740 of expense for a edge case. However, I have improved the situation in cases
741 such as /(a)(b)x|abc/ matched against "abc", where the return code
742 indicates that fewer than the maximum number of slots in the ovector have
743 been set.
744
745 3. Related to (2) above: when there are more back references in a pattern than
746 slots in the output vector, pcre_exec() uses temporary memory during
747 matching, and copies in the captures as far as possible afterwards. It was
748 using the entire output vector, but this conflicts with the specification
749 that only 2/3 is used for passing back captured substrings. Now it uses
750 only the first 2/3, for compatibility. This is, of course, another edge
751 case.
752
753 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
754 main code base, and can be used by building with --enable-jit. When this is
755 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
756 runtime --no-jit option is given.
757
758 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
759 ovector, the return from the function was zero, implying that there were
760 other matches that did not fit. The correct "exactly full" value is now
761 returned.
762
763 6. If a subpattern that was called recursively or as a subroutine contained
764 (*PRUNE) or any other control that caused it to give a non-standard return,
765 invalid errors such as "Error -26 (nested recursion at the same subject
766 position)" or even infinite loops could occur.
767
768 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
769 computing the minimum length on reaching *ACCEPT, and so ended up with the
770 wrong value of 1 rather than 0. Further investigation indicates that
771 computing a minimum subject length in the presence of *ACCEPT is difficult
772 (think back references, subroutine calls), and so I have changed the code
773 so that no minimum is registered for a pattern that contains *ACCEPT.
774
775 8. If (*THEN) was present in the first (true) branch of a conditional group,
776 it was not handled as intended. [But see 16 below.]
777
778 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
779 Sheri Pierce.
780
781 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
782 the first byte in a match must be "a".
783
784 11. Change 17 for 8.13 increased the recursion depth for patterns like
785 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
786 pattern contains any instances of (*THEN). If it does not, the old
787 optimizations are restored. It would be nice to do this on a per-group
788 basis, but at the moment that is not feasible.
789
790 12. In some environments, the output of pcretest -C is CRLF terminated. This
791 broke RunTest's code that checks for the link size. A single white space
792 character after the value is now allowed for.
793
794 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
795 For "fr", it uses the Windows-specific input and output files.
796
797 14. If (*THEN) appeared in a group that was called recursively or as a
798 subroutine, it did not work as intended. [But see next item.]
799
800 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
801 pattern fragments (but not containing any | characters). If A and B are
802 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
803 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
804 D. In other words, Perl considers parentheses that do not contain any |
805 characters to be part of a surrounding alternative, whereas PCRE was
806 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
807 differently. PCRE now behaves in the same way as Perl, except in the case
808 of subroutine/recursion calls such as (?1) which have in any case always
809 been different (but PCRE had them first :-).
810
811 16. Related to 15 above: Perl does not treat the | in a conditional group as
812 creating alternatives. Such a group is treated in the same way as an
813 ordinary group without any | characters when processing (*THEN). PCRE has
814 been changed to match Perl's behaviour.
815
816 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
817 RunGrepTest script failed.
818
819 18. Change 22 for version 13 caused atomic groups to use more stack. This is
820 inevitable for groups that contain captures, but it can lead to a lot of
821 stack use in large patterns. The old behaviour has been restored for atomic
822 groups that do not contain any capturing parentheses.
823
824 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
825 suppress the check for a minimum subject length at run time. (If it was
826 given to pcre_exec() or pcre_dfa_exec() it did work.)
827
828 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
829 fail to work when decoding hex characters in data strings in EBCDIC
830 environments.
831
832 21. It appears that in at least one Mac OS environment, the isxdigit() function
833 is implemented as a macro that evaluates to its argument more than once,
834 contravening the C 90 Standard (I haven't checked a later standard). There
835 was an instance in pcretest which caused it to go wrong when processing
836 \x{...} escapes in subject strings. The has been rewritten to avoid using
837 things like p++ in the argument of isxdigit().
838
839
840 Version 8.13 16-Aug-2011
841 ------------------------
842
843 1. The Unicode data tables have been updated to Unicode 6.0.0.
844
845 2. Two minor typos in pcre_internal.h have been fixed.
846
847 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
848 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
849 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
850
851 4. There were a number of related bugs in the code for matching backrefences
852 caselessly in UTF-8 mode when codes for the characters concerned were
853 different numbers of bytes. For example, U+023A and U+2C65 are an upper
854 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
855 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
856 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
857 2-byte code at the end of the subject (it thought there wasn't enough data
858 left).
859
860 5. Comprehensive information about what went wrong is now returned by
861 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
862 as the output vector has at least 2 elements. The offset of the start of
863 the failing character and a reason code are placed in the vector.
864
865 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
866 now returned is for the first byte of the failing character, instead of the
867 last byte inspected. This is an incompatible change, but I hope it is small
868 enough not to be a problem. It makes the returned offset consistent with
869 pcre_exec() and pcre_dfa_exec().
870
871 7. pcretest now gives a text phrase as well as the error number when
872 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
873 failure, the offset and reason code are output.
874
875 8. When \R was used with a maximizing quantifier it failed to skip backwards
876 over a \r\n pair if the subsequent match failed. Instead, it just skipped
877 back over a single character (\n). This seems wrong (because it treated the
878 two characters as a single entity when going forwards), conflicts with the
879 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
880 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
881 has been changed.
882
883 9. Some internal refactoring has changed the processing so that the handling
884 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
885 time (the PCRE_DOTALL option was changed this way some time ago: version
886 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
887 which was always a bit of a fudge. It also means that there is one less
888 argument for the match() function, which reduces its stack requirements
889 slightly. This change also fixes an incompatibility with Perl: the pattern
890 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
891
892 10. More internal refactoring has drastically reduced the number of recursive
893 calls to match() for possessively repeated groups such as (abc)++ when
894 using pcre_exec().
895
896 11. While implementing 10, a number of bugs in the handling of groups were
897 discovered and fixed:
898
899 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
900 (a|)*(?1) gave a compile-time internal error.
901 ((a|)+)+ did not notice that the outer group could match an empty string.
902 (^a|^)+ was not marked as anchored.
903 (.*a|.*)+ was not marked as matching at start or after a newline.
904
905 12. Yet more internal refactoring has removed another argument from the match()
906 function. Special calls to this function are now indicated by setting a
907 value in a variable in the "match data" data block.
908
909 13. Be more explicit in pcre_study() instead of relying on "default" for
910 opcodes that mean there is no starting character; this means that when new
911 ones are added and accidentally left out of pcre_study(), testing should
912 pick them up.
913
914 14. The -s option of pcretest has been documented for ages as being an old
915 synonym of -m (show memory usage). I have changed it to mean "force study
916 for every regex", that is, assume /S for every regex. This is similar to -i
917 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
918 using it. It makes it easier to run collections of tests with and without
919 study enabled, and thereby test pcre_study() more easily. All the standard
920 tests are now run with and without -s (but some patterns can be marked as
921 "never study" - see 20 below).
922
923 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
924 restoration of the capturing data to the outer values was not happening
925 correctly.
926
927 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
928 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
929 pattern had matched an empty string, and so incorrectly returned a no
930 match.
931
932 17. There was optimizing code for the last branch of non-capturing parentheses,
933 and also for the obeyed branch of a conditional subexpression, which used
934 tail recursion to cut down on stack usage. Unfortunately, now that there is
935 the possibility of (*THEN) occurring in these branches, tail recursion is
936 no longer possible because the return has to be checked for (*THEN). These
937 two optimizations have therefore been removed. [But see 8.20/11 above.]
938
939 18. If a pattern containing \R was studied, it was assumed that \R always
940 matched two bytes, thus causing the minimum subject length to be
941 incorrectly computed because \R can also match just one byte.
942
943 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
944 was incorrectly computed.
945
946 20. If /S is present twice on a test pattern in pcretest input, it now
947 *disables* studying, thereby overriding the use of -s on the command line
948 (see 14 above). This is necessary for one or two tests to keep the output
949 identical in both cases.
950
951 21. When (*ACCEPT) was used in an assertion that matched an empty string and
952 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
953
954 22. When an atomic group that contained a capturing parenthesis was
955 successfully matched, but the branch in which it appeared failed, the
956 capturing was not being forgotten if a higher numbered group was later
957 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
958 group 1 to "a", when in fact it should be unset. This applied to multi-
959 branched capturing and non-capturing groups, repeated or not, and also to
960 positive assertions (capturing in negative assertions does not happen
961 in PCRE) and also to nested atomic groups.
962
963 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
964 subject after a captured substring, to make it easier to tell which of a
965 number of identical substrings has been captured.
966
967 24. The way atomic groups are processed by pcre_exec() has been changed so that
968 if they are repeated, backtracking one repetition now resets captured
969 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
970 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
971 "aaa". Previously, it would have been "a". As part of this code
972 refactoring, the way recursive calls are handled has also been changed.
973
974 25. If an assertion condition captured any substrings, they were not passed
975 back unless some other capturing happened later. For example, if
976 (?(?=(a))a) was matched against "a", no capturing was returned.
977
978 26. When studying a pattern that contained subroutine calls or assertions,
979 the code for finding the minimum length of a possible match was handling
980 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
981 group 1 called group 2 while simultaneously a separate group 2 called group
982 1). A stack overflow occurred in this case. I have fixed this by limiting
983 the recursion depth to 10.
984
985 27. Updated RunTest.bat in the distribution to the version supplied by Tom
986 Fortmann. This supports explicit test numbers on the command line, and has
987 argument validation and error reporting.
988
989 28. An instance of \X with an unlimited repeat could fail if at any point the
990 first character it looked at was a mark character.
991
992 29. Some minor code refactoring concerning Unicode properties and scripts
993 should reduce the stack requirement of match() slightly.
994
995 30. Added the '=' option to pcretest to check the setting of unused capturing
996 slots at the end of the pattern, which are documented as being -1, but are
997 not included in the return count.
998
999 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1000 compiled something random. Now it gives a compile-time error (as does
1001 Perl).
1002
1003 32. A *MARK encountered during the processing of a positive assertion is now
1004 recorded and passed back (compatible with Perl).
1005
1006 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1007 had alternative anchored branches, the search for a second match in a line
1008 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1009 matched the line "0102" twice. The same bug affected patterns that started
1010 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1011 twice.
1012
1013 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1014 does, and because of capturing effects, quantifying parenthesized
1015 assertions may at times be useful. Quantifiers are now allowed for
1016 parenthesized assertions.
1017
1018 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1019
1020 36. \g was being checked for fancy things in a character class, when it should
1021 just be a literal "g".
1022
1023 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1024 appearance of a nested POSIX class supersedes an apparent external class.
1025 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1026 unescaped square brackets may also appear as part of class names. For
1027 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1028 more like Perl. (But see 8.20/1 above.)
1029
1030 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1031 was because it thought it was \N{name}, which is not supported).
1032
1033 39. Add minix to OS list not supporting the -S option in pcretest.
1034
1035 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1036 cannot analyze patterns in sufficient detail to catch mutual recursions
1037 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1038 subgroup is called recursively as a subpattern for a second time at the
1039 same position in the subject string. In previous releases this might have
1040 been caught by the recursion limit, or it might have run out of stack.
1041
1042 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1043 happen only once. PCRE was, however incorrectly giving a compile time error
1044 "recursive call could loop indefinitely" because it cannot analyze the
1045 pattern in sufficient detail. The compile time test no longer happens when
1046 PCRE is compiling a conditional subpattern, but actual runaway loops are
1047 now caught at runtime (see 40 above).
1048
1049 42. It seems that Perl allows any characters other than a closing parenthesis
1050 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1051 has been changed to be the same.
1052
1053 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1054 as not to get warnings when autogen.sh is called. Also changed
1055 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1056
1057 44. To help people who use pcregrep to scan files containing exceedingly long
1058 lines, the following changes have been made:
1059
1060 (a) The default value of the buffer size parameter has been increased from
1061 8K to 20K. (The actual buffer used is three times this size.)
1062
1063 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1064 PCRE is built.
1065
1066 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1067 to be set at run time.
1068
1069 (d) Numerical values in pcregrep options can be followed by K or M, for
1070 example --buffer-size=50K.
1071
1072 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1073 given and the return code is set to 2.
1074
1075 45. Add a pointer to the latest mark to the callout data block.
1076
1077 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1078 partial match of an empty string instead of no match. This was specific to
1079 the use of ".".
1080
1081 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1082 complete match instead of a partial match. This bug was dependent on both
1083 the PCRE_UTF8 and PCRE_DOTALL options being set.
1084
1085 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1086 starting byte set, because \b was not being ignored.
1087
1088
1089 Version 8.12 15-Jan-2011
1090 ------------------------
1091
1092 1. Fixed some typos in the markup of the man pages, and wrote a script that
1093 checks for such things as part of the documentation building process.
1094
1095 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1096 --match-limit and --recursion-limit options (added for 8.11). In
1097 particular, this made one of the standard tests fail. (The integer value
1098 went into the wrong half of a long int.)
1099
1100 3. If the --colour option was given to pcregrep with -v (invert match), it
1101 did strange things, either producing crazy output, or crashing. It should,
1102 of course, ignore a request for colour when reporting lines that do not
1103 match.
1104
1105 4. Another pcregrep bug caused similar problems if --colour was specified with
1106 -M (multiline) and the pattern match finished with a line ending.
1107
1108 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1109 matched in multiline mode, the following line was shown as part of the
1110 match. This seems wrong, so I have changed it.
1111
1112 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1113 the check for further matches in the same line (so they could be coloured)
1114 to overrun the end of the current line. If another match was found, it was
1115 incorrectly shown (and then shown again when found in the next line).
1116
1117 7. If pcregrep was compiled under Windows, there was a reference to the
1118 function pcregrep_exit() before it was defined. I am assuming this was
1119 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1120 reported by a user. I've moved the definition above the reference.
1121
1122
1123 Version 8.11 10-Dec-2010
1124 ------------------------
1125
1126 1. (*THEN) was not working properly if there were untried alternatives prior
1127 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1128 backtracked to try for "b" instead of moving to the next alternative branch
1129 at the same level (in this case, to look for "c"). The Perl documentation
1130 is clear that when (*THEN) is backtracked onto, it goes to the "next
1131 alternative in the innermost enclosing group".
1132
1133 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1134 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1135 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1136 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1137 (*THEN).
1138
1139 3. If \s appeared in a character class, it removed the VT character from
1140 the class, even if it had been included by some previous item, for example
1141 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1142 of \s, but is part of the POSIX "space" class.)
1143
1144 4. A partial match never returns an empty string (because you can always
1145 match an empty string at the end of the subject); however the checking for
1146 an empty string was starting at the "start of match" point. This has been
1147 changed to the "earliest inspected character" point, because the returned
1148 data for a partial match starts at this character. This means that, for
1149 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1150 (previously it gave "no match").
1151
1152 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1153 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1154 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1155 has an implication that the given string is incomplete (because a partial
1156 match is preferred over a full match). For this reason, these items now
1157 give a partial match in this situation. [Aside: previously, the one case
1158 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1159 match rather than a full match, which was wrong by the old rules, but is
1160 now correct.]
1161
1162 6. There was a bug in the handling of #-introduced comments, recognized when
1163 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1164 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1165 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1166 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1167 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1168 places in pcre_compile().
1169
1170 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1171 comments when looking ahead for named forward references to subpatterns,
1172 the only newline sequence it recognized was NL. It now handles newlines
1173 according to the set newline convention.
1174
1175 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1176 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1177 cater for a lack of strerror(). These oversights have been fixed.
1178
1179 9. Added --match-limit and --recursion-limit to pcregrep.
1180
1181 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1182
1183 11. When the -o option was used, pcregrep was setting a return code of 1, even
1184 when matches were found, and --line-buffered was not being honoured.
1185
1186 12. Added an optional parentheses number to the -o and --only-matching options
1187 of pcregrep.
1188
1189 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1190 can match an empty string. The code to do it in pcretest and pcredemo
1191 needed fixing:
1192
1193 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1194 only one byte after an empty string match just before CRLF (this case
1195 just got forgotten; "any" and "anycrlf" were OK).
1196
1197 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1198 mode when an empty string match preceded an ASCII character followed by
1199 a non-ASCII character. (The code for advancing by one character rather
1200 than one byte was nonsense.)
1201
1202 (c) The pcredemo.c sample program did not have any code at all to handle
1203 the cases when CRLF is a valid newline sequence.
1204
1205 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1206 as a starting offset was within the subject string. There is now a new
1207 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1208 negative or greater than the length of the string. In order to test this,
1209 pcretest is extended to allow the setting of negative starting offsets.
1210
1211 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1212 starting offset points to the beginning of a UTF-8 character was
1213 unnecessarily clumsy. I tidied it up.
1214
1215 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1216 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1217
1218 17. Nobody had reported that the --include_dir option, which was added in
1219 release 7.7 should have been called --include-dir (hyphen, not underscore)
1220 for compatibility with GNU grep. I have changed it to --include-dir, but
1221 left --include_dir as an undocumented synonym, and the same for
1222 --exclude-dir, though that is not available in GNU grep, at least as of
1223 release 2.5.4.
1224
1225 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1226 characters from a string of bytes) have been redefined so as not to use
1227 loops, in order to improve performance in some environments. At the same
1228 time, I abstracted some of the common code into auxiliary macros to save
1229 repetition (this should not affect the compiled code).
1230
1231 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1232 compile-time error is now given if \c is not followed by an ASCII
1233 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1234 different, and any byte value is allowed.)
1235
1236 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1237 START_OPTIMIZE option, which is now allowed at compile time - but just
1238 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1239 to pcregrep and other applications that have no direct access to PCRE
1240 options. The new /Y option in pcretest sets this option when calling
1241 pcre_compile().
1242
1243 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1244 back references. Groups containing recursive back references were forced to
1245 be atomic by that change, but in the case of named groups, the amount of
1246 memory required was incorrectly computed, leading to "Failed: internal
1247 error: code overflow". This has been fixed.
1248
1249 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1250 pcretest.c, to avoid build problems in some Borland environments.
1251
1252
1253 Version 8.10 25-Jun-2010
1254 ------------------------
1255
1256 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1257 THEN.
1258
1259 2. (*ACCEPT) was not working when inside an atomic group.
1260
1261 3. Inside a character class, \B is treated as a literal by default, but
1262 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1263 causes the error). The code is unchanged, but I tidied the documentation.
1264
1265 4. Inside a character class, PCRE always treated \R and \X as literals,
1266 whereas Perl faults them if its -w option is set. I have changed PCRE so
1267 that it faults them when PCRE_EXTRA is set.
1268
1269 5. Added support for \N, which always matches any character other than
1270 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1271
1272 6. When compiling pcregrep with newer versions of gcc which may have
1273 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1274 declared with attribute warn_unused_result" were given. Just casting the
1275 result to (void) does not stop the warnings; a more elaborate fudge is
1276 needed. I've used a macro to implement this.
1277
1278 7. Minor change to pcretest.c to avoid a compiler warning.
1279
1280 8. Added four artifical Unicode properties to help with an option to make
1281 \s etc use properties (see next item). The new properties are: Xan
1282 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1283
1284 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1285 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1286 this option. Modified pcretest to add /W to test this facility. Added
1287 REG_UCP to make it available via the POSIX interface.
1288
1289 10. Added --line-buffered to pcregrep.
1290
1291 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1292 studied, and the match started with a letter with a code point greater than
1293 127 whose first byte was different to the first byte of the other case of
1294 the letter, the other case of this starting letter was not recognized
1295 (#976).
1296
1297 12. If a pattern that was studied started with a repeated Unicode property
1298 test, for example, \p{Nd}+, there was the theoretical possibility of
1299 setting up an incorrect bitmap of starting bytes, but fortunately it could
1300 not have actually happened in practice until change 8 above was made (it
1301 added property types that matched character-matching opcodes).
1302
1303 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1304 possible starting bytes for non-anchored patterns.
1305
1306 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1307 \R, and also a number of cases that involve Unicode properties, both
1308 explicit and implicit when PCRE_UCP is set.
1309
1310 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1311 input, it could crash or give wrong results if characters with values
1312 greater than 0xc0 were present in the subject string. (Detail: it assumed
1313 UTF-8 input when processing these items.)
1314
1315 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1316 size_t is 64-bit (#991).
1317
1318 17. Added a check for running out of memory when PCRE is compiled with
1319 --disable-stack-for-recursion (#990).
1320
1321 18. If the last data line in a file for pcretest does not have a newline on
1322 the end, a newline was missing in the output.
1323
1324 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1325 less than 128) in its various bitmaps. However, there is a facility for
1326 generating tables according to the current locale when PCRE is compiled. It
1327 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1328 characters, are recognized by isspace() and therefore were getting set in
1329 these tables, and indeed these tables seem to approximate to ISO 8859. This
1330 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1331 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1332 which of course cannot start UTF-8 characters. I have changed the code so
1333 that only real ASCII characters (less than 128) and the correct starting
1334 bytes for UTF-8 encodings are set for characters greater than 127 when in
1335 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1336 altogether.)
1337
1338 20. Added the /T option to pcretest so as to be able to run tests with non-
1339 standard character tables, thus making it possible to include the tests
1340 used for 19 above in the standard set of tests.
1341
1342 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1343 reference to a subpattern the other side of a comment that contains an
1344 opening parenthesis caused either an internal compiling error, or a
1345 reference to the wrong subpattern.
1346
1347
1348 Version 8.02 19-Mar-2010
1349 ------------------------
1350
1351 1. The Unicode data tables have been updated to Unicode 5.2.0.
1352
1353 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1354 configured.
1355
1356 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1357 original author of that file, following a query about its status.
1358
1359 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1360 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1361
1362 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1363 quantifier applied to a forward-referencing subroutine call, could compile
1364 incorrect code or give the error "internal error: previously-checked
1365 referenced subpattern not found".
1366
1367 6. Both MS Visual Studio and Symbian OS have problems with initializing
1368 variables to point to external functions. For these systems, therefore,
1369 pcre_malloc etc. are now initialized to local functions that call the
1370 relevant global functions.
1371
1372 7. There were two entries missing in the vectors called coptable and poptable
1373 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1374 I've fixed the data, and added a kludgy way of testing at compile time that
1375 the lengths are correct (equal to the number of opcodes).
1376
1377 8. Following on from 7, I added a similar kludge to check the length of the
1378 eint vector in pcreposix.c.
1379
1380 9. Error texts for pcre_compile() are held as one long string to avoid too
1381 much relocation at load time. To find a text, the string is searched,
1382 counting zeros. There was no check for running off the end of the string,
1383 which could happen if a new error number was added without updating the
1384 string.
1385
1386 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1387
1388 11. \K was not working if it appeared in an atomic group or in a group that
1389 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1390 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1391 the assertion is positive, but not if it is negative.
1392
1393 12. Change 11 fortuitously reduced the size of the stack frame used in the
1394 "match()" function of pcre_exec.c by one pointer. Forthcoming
1395 implementation of support for (*MARK) will need an extra pointer on the
1396 stack; I have reserved it now, so that the stack frame size does not
1397 decrease.
1398
1399 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1400 item in branch that calls a recursion is a subroutine call - as in the
1401 second branch in the above example - was incorrectly given the compile-
1402 time error "recursive call could loop indefinitely" because pcre_compile()
1403 was not correctly checking the subroutine for matching a non-empty string.
1404
1405 14. The checks for overrunning compiling workspace could trigger after an
1406 overrun had occurred. This is a "should never occur" error, but it can be
1407 triggered by pathological patterns such as hundreds of nested parentheses.
1408 The checks now trigger 100 bytes before the end of the workspace.
1409
1410 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1411
1412
1413 Version 8.01 19-Jan-2010
1414 ------------------------
1415
1416 1. If a pattern contained a conditional subpattern with only one branch (in
1417 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1418 computed the wrong minimum data length (which is of course zero for such
1419 subpatterns). This could cause incorrect "no match" results.
1420
1421 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1422 the pattern is reset in the first branch, pcre_compile() failed with
1423 "internal error: code overflow at offset...". This happened only when
1424 the reset was to the original external option setting. (An optimization
1425 abstracts leading options settings into an external setting, which was the
1426 cause of this.)
1427
1428 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1429 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1430 assertion pattern did not match (meaning that the assertion was true), it
1431 was incorrectly treated as false if the SKIP had been reached during the
1432 matching. This also applied to assertions used as conditions.
1433
1434 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1435 assertion subpattern, including such a pattern used as a condition,
1436 unpredictable results occurred, instead of the error return
1437 PCRE_ERROR_DFA_UITEM.
1438
1439 5. The C++ GlobalReplace function was not working like Perl for the special
1440 situation when an empty string is matched. It now does the fancy magic
1441 stuff that is necessary.
1442
1443 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1444 removed. (These were left over from very, very early versions of PCRE.)
1445
1446 7. Some cosmetic changes to the code to make life easier when compiling it
1447 as part of something else:
1448
1449 (a) Change DEBUG to PCRE_DEBUG.
1450
1451 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1452 called "current" as "current_branch", to prevent a collision with the
1453 Linux macro when compiled as a kernel module.
1454
1455 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1456 prevent a collision with the Linux macro when compiled as a kernel
1457 module.
1458
1459 8. In pcre_compile() there are some checks for integer overflows that used to
1460 cast potentially large values to (double). This has been changed to that
1461 when building, a check for int64_t is made, and if it is found, it is used
1462 instead, thus avoiding the use of floating point arithmetic. (There is no
1463 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1464 double.
1465
1466 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1467 2005 (difference between two addresses compared to an unsigned value).
1468
1469 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1470 custom one, because of the following reported problem in Windows:
1471
1472 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1473 under Win32.
1474 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1475 therefore missing the function definition.
1476 - The compiler thus generates a "C" signature for the test function.
1477 - The linker fails to find the "C" function.
1478 - PCRE fails to configure if asked to do so against libbz2.
1479
1480 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1481 messages were output:
1482
1483 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1484 rerunning libtoolize, to keep the correct libtool macros in-tree.
1485 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1486
1487 I have done both of these things.
1488
1489 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1490 most of the time, it *can* run out if it is given a pattern that contains a
1491 runaway infinite recursion. I updated the discussion in the pcrestack man
1492 page.
1493
1494 13. Now that we have gone to the x.xx style of version numbers, the minor
1495 version may start with zero. Using 08 or 09 is a bad idea because users
1496 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1497 interpreted as invalid octal numbers. I've updated the previous comment in
1498 configure.ac, and also added a check that gives an error if 08 or 09 are
1499 used.
1500
1501 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1502 causing partial matching to fail when the end of the subject matched \W
1503 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1504
1505 15. There were some discrepancies between the declarations in pcre_internal.h
1506 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1507 their definitions. The declarations used "const uschar *" and the
1508 definitions used USPTR. Even though USPTR is normally defined as "const
1509 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1510 reported that: "This difference in casting confuses some C++ compilers, for
1511 example, SunCC recognizes above declarations as different functions and
1512 generates broken code for hbpcre." I have changed the declarations to use
1513 USPTR.
1514
1515 16. GNU libtool is named differently on some systems. The autogen.sh script now
1516 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1517 (FreeBSD).
1518
1519 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1520 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1521 comment: "Figure out how to create a longlong from a string: strtoll and
1522 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1523 instance, but it only takes 2 args instead of 3!"
1524
1525 18. A subtle bug concerned with back references has been fixed by a change of
1526 specification, with a corresponding code fix. A pattern such as
1527 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1528 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1529 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1530 same bug. Such groups have to be quantified to be useful, or contained
1531 inside another quantified group. (If there's no repetition, the reference
1532 can never match.) The problem arises because, having left the group and
1533 moved on to the rest of the pattern, a later failure that backtracks into
1534 the group uses the captured value from the final iteration of the group
1535 rather than the correct earlier one. I have fixed this in PCRE by forcing
1536 any group that contains a reference to itself to be an atomic group; that
1537 is, there cannot be any backtracking into it once it has completed. This is
1538 similar to recursive and subroutine calls.
1539
1540
1541 Version 8.00 19-Oct-09
1542 ----------------------
1543
1544 1. The table for translating pcre_compile() error codes into POSIX error codes
1545 was out-of-date, and there was no check on the pcre_compile() error code
1546 being within the table. This could lead to an OK return being given in
1547 error.
1548
1549 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1550 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1551 in a Windows environment.
1552
1553 3. The pcregrep --count option prints the count for each file even when it is
1554 zero, as does GNU grep. However, pcregrep was also printing all files when
1555 --files-with-matches was added. Now, when both options are given, it prints
1556 counts only for those files that have at least one match. (GNU grep just
1557 prints the file name in this circumstance, but including the count seems
1558 more useful - otherwise, why use --count?) Also ensured that the
1559 combination -clh just lists non-zero counts, with no names.
1560
1561 4. The long form of the pcregrep -F option was incorrectly implemented as
1562 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1563 but it seems right to fix it, and I didn't think it was worth preserving
1564 the old behaviour.
1565
1566 5. The command line items --regex=pattern and --regexp=pattern were not
1567 recognized by pcregrep, which required --regex pattern or --regexp pattern
1568 (with a space rather than an '='). The man page documented the '=' forms,
1569 which are compatible with GNU grep; these now work.
1570
1571 6. No libpcreposix.pc file was created for pkg-config; there was just
1572 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1573
1574 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1575 when UCP support is not needed, by modifying the Python script that
1576 generates it from Unicode data files. This should not matter if the module
1577 is correctly used as a library, but I received one complaint about 50K of
1578 unwanted data. My guess is that the person linked everything into his
1579 program rather than using a library. Anyway, it does no harm.
1580
1581 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1582 was a minimum greater than 1 for a wide character in a possessive
1583 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1584 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1585 character. Chaos in the form of incorrect output or a compiling loop could
1586 result.
1587
1588 9. The restrictions on what a pattern can contain when partial matching is
1589 requested for pcre_exec() have been removed. All patterns can now be
1590 partially matched by this function. In addition, if there are at least two
1591 slots in the offset vector, the offset of the earliest inspected character
1592 for the match and the offset of the end of the subject are set in them when
1593 PCRE_ERROR_PARTIAL is returned.
1594
1595 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1596 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1597 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1598 and may be more useful for multi-segment matching.
1599
1600 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1601 used to be given if ever the end of the subject was reached; now it is
1602 given only if matching could not proceed because another character was
1603 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1604 string "Z", which now yields "no match" instead of "partial match". In the
1605 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1606 final character ended with (*FAIL).
1607
1608 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1609 if the pattern had a "must contain" character that was already found in the
1610 earlier partial match, unless partial matching was again requested. For
1611 example, with the pattern /dog.(body)?/, the "must contain" character is
1612 "g". If the first part-match was for the string "dog", restarting with
1613 "sbody" failed. This bug has been fixed.
1614
1615 13. The string returned by pcre_dfa_exec() after a partial match has been
1616 changed so that it starts at the first inspected character rather than the
1617 first character of the match. This makes a difference only if the pattern
1618 starts with a lookbehind assertion or \b or \B (\K is not supported by
1619 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1620 matching functions compatible, and I think it's the right thing to do.
1621
1622 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1623 so that the demonstration program is easily available in environments where
1624 PCRE has not been installed from source.
1625
1626 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1627 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1628 library.
1629
1630 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1631 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1632 is not the first non-POSIX option to be added. Clearly some people find
1633 these options useful.
1634
1635 17. If a caller to the POSIX matching function regexec() passes a non-zero
1636 value for nmatch with a NULL value for pmatch, the value of
1637 nmatch is forced to zero.
1638
1639 18. RunGrepTest did not have a test for the availability of the -u option of
1640 the diff command, as RunTest does. It now checks in the same way as
1641 RunTest, and also checks for the -b option.
1642
1643 19. If an odd number of negated classes containing just a single character
1644 interposed, within parentheses, between a forward reference to a named
1645 subpattern and the definition of the subpattern, compilation crashed with
1646 an internal error, complaining that it could not find the referenced
1647 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1648 [The bug was that it was starting one character too far in when skipping
1649 over the character class, thus treating the ] as data rather than
1650 terminating the class. This meant it could skip too much.]
1651
1652 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1653 /g option in pcretest when the pattern contains \K, which makes it possible
1654 to have an empty string match not at the start, even when the pattern is
1655 anchored. Updated pcretest and pcredemo to use this option.
1656
1657 21. If the maximum number of capturing subpatterns in a recursion was greater
1658 than the maximum at the outer level, the higher number was returned, but
1659 with unset values at the outer level. The correct (outer level) value is
1660 now given.
1661
1662 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1663 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1664 make it do so. The string so far is captured, making this feature
1665 compatible with Perl.
1666
1667 23. The tests have been re-organized, adding tests 11 and 12, to make it
1668 possible to check the Perl 5.10 features against Perl 5.10.
1669
1670 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1671 pattern matches a fixed length string. PCRE did not allow this; now it
1672 does. Neither allows recursion.
1673
1674 25. I finally figured out how to implement a request to provide the minimum
1675 length of subject string that was needed in order to match a given pattern.
1676 (It was back references and recursion that I had previously got hung up
1677 on.) This code has now been added to pcre_study(); it finds a lower bound
1678 to the length of subject needed. It is not necessarily the greatest lower
1679 bound, but using it to avoid searching strings that are too short does give
1680 some useful speed-ups. The value is available to calling programs via
1681 pcre_fullinfo().
1682
1683 26. While implementing 25, I discovered to my embarrassment that pcretest had
1684 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1685 study optimizations had never been tested with that matching function.
1686 Oops. What is worse, even when it was passed study data, there was a bug in
1687 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1688 were also very few tests of studied patterns with pcre_dfa_exec().
1689
1690 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1691 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1692 on the other side of the coin, they are no longer allowed to have different
1693 names, because these cannot be distinguished in PCRE, and this has caused
1694 confusion. (This is a difference from Perl.)
1695
1696 28. When duplicate subpattern names are present (necessarily with different
1697 numbers, as required by 27 above), and a test is made by name in a
1698 conditional pattern, either for a subpattern having been matched, or for
1699 recursion in such a pattern, all the associated numbered subpatterns are
1700 tested, and the overall condition is true if the condition is true for any
1701 one of them. This is the way Perl works, and is also more like the way
1702 testing by number works.
1703
1704
1705 Version 7.9 11-Apr-09
1706 ---------------------
1707
1708 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1709 (pcretest), all targets were linked against these libraries. This included
1710 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1711 libraries. This caused unwanted dependencies to be created. This problem
1712 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1713 pcretest is linked with readline.
1714
1715 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1716 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1717 moved outside it again, because FALSE and TRUE are already defined in AIX,
1718 but BOOL is not.
1719
1720 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1721 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1722
1723 4. The pcregrep documentation said spaces were inserted as well as colons (or
1724 hyphens) following file names and line numbers when outputting matching
1725 lines. This is not true; no spaces are inserted. I have also clarified the
1726 wording for the --colour (or --color) option.
1727
1728 5. In pcregrep, when --colour was used with -o, the list of matching strings
1729 was not coloured; this is different to GNU grep, so I have changed it to be
1730 the same.
1731
1732 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1733 each matching line was coloured. Now it goes on to look for further matches
1734 of any of the test patterns, which is the same behaviour as GNU grep.
1735
1736 7. A pattern that could match an empty string could cause pcregrep to loop; it
1737 doesn't make sense to accept an empty string match in pcregrep, so I have
1738 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1739 seems to be how GNU grep behaves. [But see later change 40 for release
1740 8.33.]
1741
1742 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1743 start or after a newline", because the conditional assertion was not being
1744 correctly handled. The rule now is that both the assertion and what follows
1745 in the first alternative must satisfy the test.
1746
1747 9. If auto-callout was enabled in a pattern with a conditional group whose
1748 condition was an assertion, PCRE could crash during matching, both with
1749 pcre_exec() and pcre_dfa_exec().
1750
1751 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1752 used for matching.
1753
1754 11. Unicode property support in character classes was not working for
1755 characters (bytes) greater than 127 when not in UTF-8 mode.
1756
1757 12. Added the -M command line option to pcretest.
1758
1759 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1760
1761 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1762
1763 16. Added comments and documentation about mis-use of no_arg in the C++
1764 wrapper.
1765
1766 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1767 from Martin Jerabek that uses macro names for all relevant character and
1768 string constants.
1769
1770 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1771 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1772 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1773 these, but not everybody uses configure.
1774
1775 19. A conditional group that had only one branch was not being correctly
1776 recognized as an item that could match an empty string. This meant that an
1777 enclosing group might also not be so recognized, causing infinite looping
1778 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1779 with the subject "ab", where knowledge that the repeated group can match
1780 nothing is needed in order to break the loop.
1781
1782 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1783 exec(), but without supplying a callout function, matching went wrong.
1784
1785 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1786 leak if the size of the offset vector was greater than 30. When the vector
1787 is smaller, the saved offsets during recursion go onto a local stack
1788 vector, but for larger vectors malloc() is used. It was failing to free
1789 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1790 error, in fact).
1791
1792 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1793 heapframe that is used only when UTF-8 support is enabled. This caused no
1794 problem, but was untidy.
1795
1796 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1797 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1798 included within another project.
1799
1800 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1801 slightly modified by me:
1802
1803 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1804 not building pcregrep.
1805
1806 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1807 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1808
1809 25. Forward references, both numeric and by name, in patterns that made use of
1810 duplicate group numbers, could behave incorrectly or give incorrect errors,
1811 because when scanning forward to find the reference group, PCRE was not
1812 taking into account the duplicate group numbers. A pattern such as
1813 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1814
1815 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1816 the feature of a custom pointer more persuasive (as requested by a user).
1817
1818 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1819 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1820 pre-defined.
1821
1822 28. Added support for (*UTF8) at the start of a pattern.
1823
1824 29. Arrange for flags added by the "release type" setting in CMake to be shown
1825 in the configuration summary.
1826
1827
1828 Version 7.8 05-Sep-08
1829 ---------------------
1830
1831 1. Replaced UCP searching code with optimized version as implemented for Ad
1832 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1833 stage table and inline lookup instead of a function, giving speed ups of 2
1834 to 5 times on some simple patterns that I tested. Permission was given to
1835 distribute the MultiStage2.py script that generates the tables (it's not in
1836 the tarball, but is in the Subversion repository).
1837
1838 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1839 scripts.
1840
1841 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1842 a group with a zero qualifier. The result of the study could be incorrect,
1843 or the function might crash, depending on the pattern.
1844
1845 4. Caseless matching was not working for non-ASCII characters in back
1846 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1847 It now works when Unicode Property Support is available.
1848
1849 5. In pcretest, an escape such as \x{de} in the data was always generating
1850 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1851 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1852 truncation.
1853
1854 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1855
1856 7. Added two (int) casts to pcregrep when printing the difference of two
1857 pointers, in case they are 64-bit values.
1858
1859 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1860 test 2 if it fails.
1861
1862 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1863 and a #define of that name to empty if it is not externally set. This is to
1864 allow users of MSVC to set it if necessary.
1865
1866 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1867 the convenience functions in the pcre_get.c source file.
1868
1869 11. An option change at the start of a pattern that had top-level alternatives
1870 could cause overwriting and/or a crash. This command provoked a crash in
1871 some environments:
1872
1873 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1874
1875 This potential security problem was recorded as CVE-2008-2371.
1876
1877 12. For a pattern where the match had to start at the beginning or immediately
1878 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1879 pcre_dfa_exec() could read past the end of the passed subject if there was
1880 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1881 pcretest so that it places the subject at the end of its malloc-ed buffer.
1882
1883 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1884 exec() might read past the end of the data buffer in UTF-8 mode.
1885
1886 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1887 the data contained the byte 0x85 as part of a UTF-8 character within its
1888 first line. This applied both to normal and DFA matching.
1889
1890 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1891 /^[^d]*?$/8 failed to match "abc".
1892
1893 16. Added a missing copyright notice to pcrecpp_internal.h.
1894
1895 17. Make it more clear in the documentation that values returned from
1896 pcre_exec() in ovector are byte offsets, not character counts.
1897
1898 18. Tidied a few places to stop certain compilers from issuing warnings.
1899
1900 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1901 supplied by Stefan Weber. I made a further small update for 7.8 because
1902 there is a change of source arrangements: the pcre_searchfuncs.c module is
1903 replaced by pcre_ucd.c.
1904
1905
1906 Version 7.7 07-May-08
1907 ---------------------
1908
1909 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1910 a string to a long long, pretend we don't even have a long long." This is
1911 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1912
1913 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1914 pre-7.6 versions, which defined a global no_arg variable instead of putting
1915 it in the RE class. (See also #8 below.)
1916
1917 3. Remove a line of dead code, identified by coverity and reported by Nuno
1918 Lopes.
1919
1920 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1921
1922 (1) The include/exclude patterns were being applied to the whole pathnames
1923 of files, instead of just to the final components.
1924
1925 (2) If there was more than one level of directory, the subdirectories were
1926 skipped unless they satisfied the include/exclude conditions. This is
1927 inconsistent with GNU grep (and could even be seen as contrary to the
1928 pcregrep specification - which I improved to make it absolutely clear).
1929 The action now is always to scan all levels of directory, and just
1930 apply the include/exclude patterns to regular files.
1931
1932 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1933 --exclude_dir in the tests to avoid scanning .svn directories.
1934
1935 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1936 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1937 doesn't support NULs in patterns.
1938
1939 7. Added some missing "const"s to declarations of static tables in
1940 pcre_compile.c and pcre_dfa_exec.c.
1941
1942 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1943 caused by fix #2 above. (Subsequently also a second patch to fix the
1944 first patch. And a third patch - this was a messy problem.)
1945
1946 9. Applied Craig's patch to remove the use of push_back().
1947
1948 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1949 matching function regexec().
1950
1951 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1952 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1953 references. PCRE supports relative numbers with this syntax (I don't think
1954 Oniguruma does).
1955
1956 12. Previously, a group with a zero repeat such as (...){0} was completely
1957 omitted from the compiled regex. However, this means that if the group
1958 was called as a subroutine from elsewhere in the pattern, things went wrong
1959 (an internal error was given). Such groups are now left in the compiled
1960 pattern, with a new opcode that causes them to be skipped at execution
1961 time.
1962
1963 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1964 to the way PCRE behaves:
1965
1966 (a) A lone ] character is dis-allowed (Perl treats it as data).
1967
1968 (b) A back reference to an unmatched subpattern matches an empty string
1969 (Perl fails the current match path).
1970
1971 (c) A data ] in a character class must be notated as \] because if the
1972 first data character in a class is ], it defines an empty class. (In
1973 Perl it is not possible to have an empty class.) The empty class []
1974 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1975 The negative empty class [^] matches any one character, independently
1976 of the DOTALL setting.
1977
1978 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1979 non-existent subpattern following a character class starting with ']' and
1980 containing () gave an internal compiling error instead of "reference to
1981 non-existent subpattern". Fortunately, when the pattern did exist, the
1982 compiled code was correct. (When scanning forwards to check for the
1983 existence of the subpattern, it was treating the data ']' as terminating
1984 the class, so got the count wrong. When actually compiling, the reference
1985 was subsequently set up correctly.)
1986
1987 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1988 it was being rejected as not supported by pcre_dfa_exec(), even though
1989 other assertions are supported. I have made pcre_dfa_exec() support
1990 (*FAIL).
1991
1992 16. The implementation of 13c above involved the invention of a new opcode,
1993 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1994 cannot be changed at match time, I realized I could make a small
1995 improvement to matching performance by compiling OP_ALLANY instead of
1996 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1997 on the OP_ANY path.
1998
1999 17. Compiling pcretest on Windows with readline support failed without the
2000 following two fixes: (1) Make the unistd.h include conditional on
2001 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2002
2003 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2004 ncurses library to be included for pcretest when ReadLine support is
2005 requested, but also to allow for it to be overridden. This patch came from
2006 Daniel Bergström.
2007
2008 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2009 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2010 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2011 spotting this.
2012
2013
2014 Version 7.6 28-Jan-08
2015 ---------------------
2016
2017 1. A character class containing a very large number of characters with
2018 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2019 overflow.
2020
2021 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2022 HAVE_LONG_LONG is not defined.
2023
2024 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2025 bring them up to date and include new features. This patch includes:
2026
2027 - Fixed PH's badly added libz and libbz2 support.
2028 - Fixed a problem with static linking.
2029 - Added pcredemo. [But later removed - see 7 below.]
2030 - Fixed dftables problem and added an option.
2031 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2032 HAVE_LONG_LONG.
2033 - Added readline support for pcretest.
2034 - Added an listing of the option settings after cmake has run.
2035
2036 4. A user submitted a patch to Makefile that makes it easy to create
2037 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2038 Makefile.am that cause it to include this special target, without
2039 affecting anything else. Note that the same mingw target plus all
2040 the other distribution libraries and programs are now supported
2041 when configuring with CMake (see 6 below) instead of with
2042 Configure/Make.
2043
2044 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2045 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2046 exported in the Windows port". It has not yet been confirmed that the patch
2047 solves the problem, but it does no harm.
2048
2049 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2050 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2051 with CMake, and also correct the comment about stack recursion.
2052
2053 7. Remove the automatic building of pcredemo from the ./configure system and
2054 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2055 of a program that users should build themselves after PCRE is installed, so
2056 building it automatically is not really right. What is more, it gave
2057 trouble in some build environments.
2058
2059 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2060
2061
2062 Version 7.5 10-Jan-08
2063 ---------------------
2064
2065 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2066 values in parens when parsing an RE using the C++ wrapper."
2067
2068 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2069 Characters greater than 255 were excluded from the class instead of being
2070 included.
2071
2072 3. The same bug as (2) above applied to negated POSIX classes such as
2073 [:^space:].
2074
2075 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2076 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2077 I have changed it.
2078
2079 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2080 first named subpattern) and a construct such as (?&a) would reference the
2081 first named subpattern whose name started with "a" (in other words, the
2082 length check was missing). Both these problems are fixed. "Subpattern name
2083 expected" is now given for (?&) (a zero-length name), and this patch also
2084 makes it give the same error for \k'' (previously it complained that that
2085 was a reference to a non-existent subpattern).
2086
2087 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2088 this is right because (?- can be followed by option settings as well as by
2089 digits. I have, however, made the messages clearer.
2090
2091 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2092 than the number used in the conditional) now cause a compile-time error.
2093 This is actually not compatible with Perl, which accepts such patterns, but
2094 treats the conditional as always being FALSE (as PCRE used to), but it
2095 seems to me that giving a diagnostic is better.
2096
2097 8. Change "alphameric" to the more common word "alphanumeric" in comments
2098 and messages.
2099
2100 9. Fix two occurrences of "backslash" in comments that should have been
2101 "backspace".
2102
2103 10. Remove two redundant lines of code that can never be obeyed (their function
2104 was moved elsewhere).
2105
2106 11. The program that makes PCRE's Unicode character property table had a bug
2107 which caused it to generate incorrect table entries for sequences of
2108 characters that have the same character type, but are in different scripts.
2109 It amalgamated them into a single range, with the script of the first of
2110 them. In other words, some characters were in the wrong script. There were
2111 thirteen such cases, affecting characters in the following ranges:
2112
2113 U+002b0 - U+002c1
2114 U+0060c - U+0060d
2115 U+0061e - U+00612
2116 U+0064b - U+0065e
2117 U+0074d - U+0076d
2118 U+01800 - U+01805
2119 U+01d00 - U+01d77
2120 U+01d9b - U+01dbf
2121 U+0200b - U+0200f
2122 U+030fc - U+030fe
2123 U+03260 - U+0327f
2124 U+0fb46 - U+0fbb1
2125 U+10450 - U+1049d
2126
2127 12. The -o option (show only the matching part of a line) for pcregrep was not
2128 compatible with GNU grep in that, if there was more than one match in a
2129 line, it showed only the first of them. It now behaves in the same way as
2130 GNU grep.
2131
2132 13. If the -o and -v options were combined for pcregrep, it printed a blank
2133 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2134 does the same. The return code can be used to tell if there were any
2135 non-matching lines.
2136
2137 14. Added --file-offsets and --line-offsets to pcregrep.
2138
2139 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2140 infinitely looping recursion. The bug was that positive lookaheads were not
2141 being skipped when checking for a possible empty match (negative lookaheads
2142 and both kinds of lookbehind were skipped).
2143
2144 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2145 inclusion of <windows.h> to before rather than after the definition of
2146 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2147
2148 17. Specifying a possessive quantifier with a specific limit for a Unicode
2149 character property caused pcre_compile() to compile bad code, which led at
2150 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2151 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2152 caused the error; without that there was no problem.
2153
2154 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2155
2156 19. Added --enable-pcretest-libreadline.
2157
2158 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2159 RE::GlobalReplace(). As a result, the number of replacements returned was
2160 double what it should be. I removed one of the increments, but Craig sent a
2161 later patch that removed the other one (the right fix) and added unit tests
2162 that check the return values (which was not done before).
2163
2164 21. Several CMake things:
2165
2166 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2167 the names libpcre and libpcreposix, not just pcre and pcreposix.
2168
2169 (2) The above change means that pcretest and pcregrep are now correctly
2170 linked with the newly-built libraries, not previously installed ones.
2171
2172 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2173
2174 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2175 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2176 UTF-8 newline character). The key issue is that the pattern starts .*;
2177 this means that the match must be either at the beginning, or after a
2178 newline. The bug was in the code for advancing after a failed match and
2179 checking that the new position followed a newline. It was not taking
2180 account of UTF-8 characters correctly.
2181
2182 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2183 character classes. PCRE was not treating the sequence [:...:] as a
2184 character class unless the ... were all letters. Perl, however, seems to
2185 allow any characters between [: and :], though of course it rejects as
2186 unknown any "names" that contain non-letters, because all the known class
2187 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2188 for example, whereas PCRE did not - it did not recognize a POSIX character
2189 class. This seemed a bit dangerous, so the code has been changed to be
2190 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2191 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2192 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2193 Perl does, and where it didn't before.
2194
2195 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2196 Windows environments %n is disabled by default.
2197
2198
2199 Version 7.4 21-Sep-07
2200 ---------------------
2201
2202 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2203 means that a class such as [\s] counted as "explicit reference to CR or
2204 LF". That isn't really right - the whole point of the change was to try to
2205 help when there was an actual mention of one of the two characters. So now
2206 the change happens only if \r or \n (or a literal CR or LF) character is
2207 encountered.
2208
2209 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2210 of both had grown to the point where there were only 3 bits left.
2211 Fortunately, there was spare space in the data structure, and so I have
2212 moved the internal flags into a new 16-bit field to free up more option
2213 bits.
2214
2215 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2216 but did not set the internal JCHANGED flag - either of these is enough to
2217 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2218 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2219 start sets both bits.
2220
2221 4. Added options (at build time, compile time, exec time) to change \R from
2222 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2223
2224 5. doc/pcresyntax.html was missing from the distribution.
2225
2226 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2227 compatibility, even though it is no longer used.
2228
2229 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2230 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2231 windows.h file is present (where different names are used). [This was
2232 reversed later after testing - see 16 below.]
2233
2234 8. Changed all #include <config.h> to #include "config.h". There were also
2235 some further <pcre.h> cases that I changed to "pcre.h".
2236
2237 9. When pcregrep was used with the --colour option, it missed the line ending
2238 sequence off the lines that it output.
2239
2240 10. It was pointed out to me that arrays of string pointers cause lots of
2241 relocations when a shared library is dynamically loaded. A technique of
2242 using a single long string with a table of offsets can drastically reduce
2243 these. I have refactored PCRE in four places to do this. The result is
2244 dramatic:
2245
2246 Originally: 290
2247 After changing UCP table: 187
2248 After changing error message table: 43
2249 After changing table of "verbs" 36
2250 After changing table of Posix names 22
2251
2252 Thanks to the folks working on Gregex for glib for this insight.
2253
2254 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2255 unicode-properties was also set.
2256
2257 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2258
2259 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2260 checked only for CRLF.
2261
2262 14. Added casts to pcretest.c to avoid compiler warnings.
2263
2264 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2265
2266 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2267 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2268 entirely. This removes changes made in 7 above.
2269
2270 17. The CMake files have been updated, and there is now more information about
2271 building with CMake in the NON-UNIX-USE document.
2272
2273
2274 Version 7.3 28-Aug-07
2275 ---------------------
2276
2277 1. In the rejigging of the build system that eventually resulted in 7.1, the
2278 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2279 brackets there is not right, since it causes compilers to look for an
2280 installed pcre.h, not the version that is in the source that is being
2281 compiled (which of course may be different). I have changed it back to:
2282
2283 #include "pcre.h"
2284
2285 I have a vague recollection that the change was concerned with compiling in
2286 different directories, but in the new build system, that is taken care of
2287 by the VPATH setting the Makefile.
2288
2289 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2290 when the subject happened to end in the byte 0x85 (e.g. if the last
2291 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2292 characters but of course it shouldn't be taken as a newline when it is part
2293 of another character. The bug was that, for an unlimited repeat of . in
2294 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2295 characters when looking for a newline.
2296
2297 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2298
2299 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2300 in debug output.
2301
2302 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2303 long printing in the pcrecpp unittest when running under MinGW.
2304
2305 6. ESC_K was left out of the EBCDIC table.
2306
2307 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2308 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2309 limit also applies to "virtual nesting" when a pattern is recursive, and in
2310 this case 1000 isn't so big. I have been able to remove this limit at the
2311 expense of backing off one optimization in certain circumstances. Normally,
2312 when pcre_exec() would call its internal match() function recursively and
2313 immediately return the result unconditionally, it uses a "tail recursion"
2314 feature to save stack. However, when a subpattern that can match an empty
2315 string has an unlimited repetition quantifier, it no longer makes this
2316 optimization. That gives it a stack frame in which to save the data for
2317 checking that an empty string has been matched. Previously this was taken
2318 from the 1000-entry workspace that had been reserved. So now there is no
2319 explicit limit, but more stack is used.
2320
2321 8. Applied Daniel's patches to solve problems with the import/export magic
2322 syntax that is required for Windows, and which was going wrong for the
2323 pcreposix and pcrecpp parts of the library. These were overlooked when this
2324 problem was solved for the main library.
2325
2326 9. There were some crude static tests to avoid integer overflow when computing
2327 the size of patterns that contain repeated groups with explicit upper
2328 limits. As the maximum quantifier is 65535, the maximum group length was
2329 set at 30,000 so that the product of these two numbers did not overflow a
2330 32-bit integer. However, it turns out that people want to use groups that
2331 are longer than 30,000 bytes (though not repeat them that many times).
2332 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2333 made it possible to implement the integer overflow checks in a much more
2334 dynamic way, which I have now done. The artificial limitation on group
2335 length has been removed - we now have only the limit on the total length of
2336 the compiled pattern, which depends on the LINK_SIZE setting.
2337
2338 10. Fixed a bug in the documentation for get/copy named substring when
2339 duplicate names are permitted. If none of the named substrings are set, the
2340 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2341 empty string.
2342
2343 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2344 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2345 because the ] is interpreted as the first data character and the
2346 terminating ] is not found. PCRE has been made compatible with Perl in this
2347 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2348 cause memory overwriting.
2349
2350 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2351 string has been matched (to stop an infinite loop). It was not recognizing
2352 a conditional subpattern that could match an empty string if that
2353 subpattern was within another subpattern. For example, it looped when
2354 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2355 condition was not nested. This bug has been fixed.
2356
2357 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2358 past the start of the subject in the presence of bytes with the top bit
2359 set, for example "\x8aBCD".
2360
2361 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2362 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2363
2364 14. Optimized (?!) to (*FAIL).
2365
2366 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2367 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2368 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2369 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2370 does: it's just the validity check that is more restrictive.
2371
2372 16. Inserted checks for integer overflows during escape sequence (backslash)
2373 processing, and also fixed erroneous offset values for syntax errors during
2374 backslash processing.
2375
2376 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2377 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2378
2379 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2380 caused an overrun.
2381
2382 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2383 something other than just ASCII characters) inside a group that had an
2384 unlimited repeat caused a loop at compile time (while checking to see
2385 whether the group could match an empty string).
2386
2387 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2388 [\P{Any}] did so. (Error in the code for printing property names.)
2389
2390 21. An orphan \E inside a character class could cause a crash.
2391
2392 22. A repeated capturing bracket such as (A)? could cause a wild memory
2393 reference during compilation.
2394
2395 23. There are several functions in pcre_compile() that scan along a compiled
2396 expression for various reasons (e.g. to see if it's fixed length for look
2397 behind). There were bugs in these functions when a repeated \p or \P was
2398 present in the pattern. These operators have additional parameters compared
2399 with \d, etc, and these were not being taken into account when moving along
2400 the compiled data. Specifically:
2401
2402 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2403 length.
2404
2405 (b) An item such as \pL+ within a repeated group could cause crashes or
2406 loops.
2407
2408 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2409 "reference to non-existent subpattern" error.
2410
2411 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2412
2413 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2414 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2415
2416 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2417
2418 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2419 character were causing crashes (broken optimization).
2420
2421 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2422 \p or \P) caused a compile-time loop.
2423
2424 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2425 break. For example, the unstudied pattern [\r\n]A does not match the string
2426 "\r\nA" because change 7.0/46 below moves the current point on by two
2427 characters after failing to match at the start. However, the pattern \nA
2428 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2429 the same is true. There doesn't seem any very clean way out of this, but
2430 what I have chosen to do makes the common cases work: PCRE now takes note
2431 of whether there can be an explicit match for \r or \n anywhere in the
2432 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2433 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2434 pattern has explicit CR or LF references.
2435
2436 29. Added (*CR) etc for changing newline setting at start of pattern.
2437
2438
2439 Version 7.2 19-Jun-07
2440 ---------------------
2441
2442 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2443 which is apparently normally available under Windows.
2444
2445 2. Re-jig the pcregrep tests with different newline settings in an attempt
2446 to make them independent of the local environment's newline setting.
2447
2448 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2449
2450 4. Some of the "internals" tests were previously cut out when the link size
2451 was not 2, because the output contained actual offsets. The recent new
2452 "Z" feature of pcretest means that these can be cut out, making the tests
2453 usable with all link sizes.
2454
2455 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2456 stack recursion. This gives a massive performance boost under BSD, but just
2457 a small improvement under Linux. However, it saves one field in the frame
2458 in all cases.
2459
2460 6. Added more features from the forthcoming Perl 5.10:
2461
2462 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2463 recursion call. It refers to the nth most recently opened parentheses.
2464
2465 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2466 to be opened parentheses.
2467
2468 (c) Conditions that refer to capturing parentheses can be specified
2469 relatively, for example, (?(-2)... or (?(+3)...
2470
2471 (d) \K resets the start of the current match so that everything before
2472 is not part of it.
2473
2474 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2475
2476 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2477 reference syntax.
2478
2479 (g) (?| introduces a group in which the numbering of parentheses in each
2480 alternative starts with the same number.
2481
2482 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2483
2484 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2485 PCRE_INFO_JCHANGED.
2486
2487 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2488 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2489 for detecting groups that can match an empty string.
2490
2491 9. A pattern with a very large number of alternatives (more than several
2492 hundred) was running out of internal workspace during the pre-compile
2493 phase, where pcre_compile() figures out how much memory will be needed. A
2494 bit of new cunning has reduced the workspace needed for groups with
2495 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2496 workspace instead of running out of the 4096 that are available.
2497
2498 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2499
2500 11. Applied patch from Google to remove an optimization that didn't quite work.
2501 The report of the bug said:
2502
2503 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2504 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2505 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2506
2507 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2508 it matched the wrong number of bytes.
2509
2510
2511 Version 7.1 24-Apr-07
2512 ---------------------
2513
2514 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2515 that is more "standard", making use of automake and other Autotools. There
2516 is some re-arrangement of the files and adjustment of comments consequent
2517 on this.
2518
2519 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2520 for recursive directory scanning broke on some systems because the files
2521 are not scanned in any specific order and on different systems the order
2522 was different. A call to "sort" has been inserted into RunGrepTest for the
2523 approprate test as a short-term fix. In the longer term there may be an
2524 alternative.
2525
2526 3. I had an email from Eric Raymond about problems translating some of PCRE's
2527 man pages to HTML (despite the fact that I distribute HTML pages, some
2528 people do their own conversions for various reasons). The problems
2529 concerned the use of low-level troff macros .br and .in. I have therefore
2530 removed all such uses from the man pages (some were redundant, some could
2531 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2532 HTML has been updated to handle .nf/.fi and to complain if it encounters
2533 .br or .in.
2534
2535 4. Updated comments in configure.ac that get placed in config.h.in and also
2536 arranged for config.h to be included in the distribution, with the name
2537 config.h.generic, for the benefit of those who have to compile without
2538 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2539
2540 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2541 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2542 makevp.bat for the current PCRE, using the additional files
2543 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2544
2545 6. A Windows user reported a minor discrepancy with test 2, which turned out
2546 to be caused by a trailing space on an input line that had got lost in his
2547 copy. The trailing space was an accident, so I've just removed it.
2548
2549 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2550 that is needed.
2551
2552 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2553 as "const" (a) because they are and (b) because it helps the PHP
2554 maintainers who have recently made a script to detect big data structures
2555 in the php code that should be moved to the .rodata section. I remembered
2556 to update Builducptable as well, so it won't revert if ucptable.h is ever
2557 re-created.
2558
2559 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2560 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2561 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2562 support is not required. This saves 1.5-2K of code, which is important in
2563 some applications.
2564
2565 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2566 so as not to refer to the tables, even though these functions will never be
2567 called when UTF-8 support is disabled. Otherwise there are problems with a
2568 shared library.
2569
2570 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2571
2572 (a) It was defining its arguments as char * instead of void *.
2573
2574 (b) It was assuming that all moves were upwards in memory; this was true
2575 a long time ago when I wrote it, but is no longer the case.
2576
2577 The emulated memove() is provided for those environments that have neither
2578 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2579 is clearly not the case, as these two bugs were recently reported.
2580
2581 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2582 and Detrail to create the HTML documentation, the .txt form of the man
2583 pages, and it removes trailing spaces from listed files. It also creates
2584 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2585 case, it wraps all the #defines with #ifndefs. This script should be run
2586 before "make dist".
2587
2588 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2589 with Unicode property support.
2590
2591 (a) For a maximizing quantifier, if the two different cases of the
2592 character were of different lengths in their UTF-8 codings (there are
2593 some cases like this - I found 11), and the matching function had to
2594 back up over a mixture of the two cases, it incorrectly assumed they
2595 were both the same length.
2596
2597 (b) When PCRE was configured to use the heap rather than the stack for
2598 recursion during matching, it was not correctly preserving the data for
2599 the other case of a UTF-8 character when checking ahead for a match
2600 while processing a minimizing repeat. If the check also involved
2601 matching a wide character, but failed, corruption could cause an
2602 erroneous result when trying to check for a repeat of the original
2603 character.
2604
2605 13. Some tidying changes to the testing mechanism:
2606
2607 (a) The RunTest script now detects the internal link size and whether there
2608 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2609 values substituted by "configure". (The RunGrepTest script already did
2610 this for UTF-8.) The configure.ac script no longer substitutes the
2611 relevant variables.
2612
2613 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2614 with length and offset values. This means that the output is different
2615 for different internal link sizes. Test 2 is skipped for link sizes
2616 other than 2 because of this, bypassing the problem. Unfortunately,
2617 there was also a test in test 3 (the locale tests) that used /B and
2618 failed for link sizes other than 2. Rather than cut the whole test out,
2619 I have added a new /Z option to pcretest that replaces the length and
2620 offset values with spaces. This is now used to make test 3 independent
2621 of link size. (Test 2 will be tidied up later.)
2622
2623 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2624 segmentation fault instead of returning the appropriate error message.
2625
2626 15. In multiline mode when the newline sequence was set to "any", the pattern
2627 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2628 This doesn't seem right; it now treats the CRLF combination as the line
2629 ending, and so does not match in that case. It's only a pattern such as ^$
2630 that would hit this one: something like ^ABC$ would have failed after \r
2631 and then tried again after \r\n.
2632
2633 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2634 in an attempt to make files that differ only in their line terminators
2635 compare equal. This works on Linux.
2636
2637 17. Under certain error circumstances pcregrep might try to free random memory
2638 as it exited. This is now fixed, thanks to valgrind.
2639
2640 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2641 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2642 was because its rules for how to advance for /g after matching an empty
2643 string at the end of a line did not allow for this case. They now check for
2644 it specially.
2645
2646 20. pcretest is supposed to handle patterns and data of any length, by
2647 extending its buffers when necessary. It was getting this wrong when the
2648 buffer for a data line had to be extended.
2649
2650 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2651 CRLF as a newline sequence.
2652
2653 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2654 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2655 I have nevertheless tidied it up.
2656
2657 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2658
2659 24. Added a man page for pcre-config.
2660
2661
2662 Version 7.0 19-Dec-06
2663 ---------------------
2664
2665 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2666 moving to gcc 4.1.1.
2667
2668 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2669 sys/time.h, which is documented as needed for this function. It doesn't
2670 seem to matter on Linux, but it showed up on some releases of OS X.
2671
2672 3. It seems that there are systems where bytes whose values are greater than
2673 127 match isprint() in the "C" locale. The "C" locale should be the
2674 default when a C program starts up. In most systems, only ASCII printing
2675 characters match isprint(). This difference caused the output from pcretest
2676 to vary, making some of the tests fail. I have changed pcretest so that:
2677
2678 (a) When it is outputting text in the compiled version of a pattern, bytes
2679 other than 32-126 are always shown as hex escapes.
2680
2681 (b) When it is outputting text that is a matched part of a subject string,
2682 it does the same, unless a different locale has been set for the match
2683 (using the /L modifier). In this case, it uses isprint() to decide.
2684
2685 4. Fixed a major bug that caused incorrect computation of the amount of memory
2686 required for a compiled pattern when options that changed within the
2687 pattern affected the logic of the preliminary scan that determines the
2688 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2689 that the computed length was too small. The symptoms of this bug were
2690 either the PCRE error "internal error: code overflow" from pcre_compile(),
2691 or a glibc crash with a message such as "pcretest: free(): invalid next
2692 size (fast)". Examples of patterns that provoked this bug (shown in
2693 pcretest format) are:
2694
2695 /(?-x: )/x
2696 /(?x)(?-x: \s*#\s*)/
2697 /((?i)[\x{c0}])/8
2698 /(?i:[\x{c0}])/8
2699
2700 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2701 is now done differently.
2702
2703 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2704 wrapper classes; (b) implement a new function in the C++ scanner that is
2705 more efficient than the old way of doing things because it avoids levels of
2706 recursion in the regex matching; (c) add a paragraph to the documentation
2707 for the FullMatch() function.
2708
2709 6. The escape sequence \n was being treated as whatever was defined as
2710 "newline". Not only was this contrary to the documentation, which states
2711 that \n is character 10 (hex 0A), but it also went horribly wrong when
2712 "newline" was defined as CRLF. This has been fixed.
2713
2714 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2715 was being set to -1 for the "end of line" case (supposedly a value that no
2716 character can have). Though this value is never used (the check for end of
2717 line is "zero bytes in current character"), it caused compiler complaints.
2718 I've changed it to 0xffffffff.
2719
2720 8. In pcre_version.c, the version string was being built by a sequence of
2721 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2722 string (as it is for production releases) called a macro with an empty
2723 argument. The C standard says the result of this is undefined. The gcc
2724 compiler treats it as an empty string (which was what was wanted) but it is
2725 reported that Visual C gives an error. The source has been hacked around to
2726 avoid this problem.
2727
2728 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2729 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2730 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2731 of them did).
2732
2733 10. Originally, pcretest opened its input and output without "b"; then I was
2734 told that "b" was needed in some environments, so it was added for release
2735 5.0 to both the input and output. (It makes no difference on Unix-like
2736 systems.) Later I was told that it is wrong for the input on Windows. I've
2737 now abstracted the modes into two macros, to make it easier to fiddle with
2738 them, and removed "b" from the input mode under Windows.
2739
2740 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2741
2742 12. Added -help and --help to pcretest as an official way of being reminded
2743 of the options.
2744
2745 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2746 and pcrecpp.cc because they annoy compilers at high warning levels.
2747
2748 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2749
2750 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2751 scripts are not C programs :-) and which was not noticed because it works
2752 on Linux.
2753
2754 16. pcretest is supposed to handle any length of pattern and data line (as one
2755 line or as a continued sequence of lines) by extending its input buffer if
2756 necessary. This feature was broken for very long pattern lines, leading to
2757 a string of junk being passed to pcre_compile() if the pattern was longer
2758 than about 50K.
2759
2760 17. I have done a major re-factoring of the way pcre_compile() computes the
2761 amount of memory needed for a compiled pattern. Previously, there was code
2762 that made a preliminary scan of the pattern in order to do this. That was
2763 OK when PCRE was new, but as the facilities have expanded, it has become
2764 harder and harder to keep it in step with the real compile phase, and there
2765 have been a number of bugs (see for example, 4 above). I have now found a
2766 cunning way of running the real compile function in a "fake" mode that
2767 enables it to compute how much memory it would need, while actually only
2768 ever using a few hundred bytes of working memory and without too many
2769 tests of the mode. This should make future maintenance and development
2770 easier. A side effect of this work is that the limit of 200 on the nesting
2771 depth of parentheses has been removed (though this was never a serious
2772 limitation, I suspect). However, there is a downside: pcre_compile() now
2773 runs more slowly than before (30% or more, depending on the pattern). I
2774 hope this isn't a big issue. There is no effect on runtime performance.
2775
2776 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2777 newline (only possible for the last line of a file) and it was a
2778 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2779
2780 19. Added additional timing features to pcretest. (1) The -tm option now times
2781 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2782 separate command line item, by a number that specifies the number of
2783 repeats to use when timing. The default is 50000; this gives better
2784 precision, but takes uncomfortably long for very large patterns.
2785
2786 20. Extended pcre_study() to be more clever in cases where a branch of a
2787 subpattern has no definite first character. For example, (a*|b*)[cd] would
2788 previously give no result from pcre_study(). Now it recognizes that the
2789 first character must be a, b, c, or d.
2790
2791 21. There was an incorrect error "recursive call could loop indefinitely" if
2792 a subpattern (or the entire pattern) that was being tested for matching an
2793 empty string contained only one non-empty item after a nested subpattern.
2794 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2795 incorrectly, because the \d was being skipped in the check.
2796
2797 22. The pcretest program now has a new pattern option /B and a command line
2798 option -b, which is equivalent to adding /B to every pattern. This causes
2799 it to show the compiled bytecode, without the additional information that
2800 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2801 is the same as /B/I).
2802
2803 23. A new optimization is now able automatically to treat some sequences such
2804 as a*b as a*+b. More specifically, if something simple (such as a character
2805 or a simple class like \d) has an unlimited quantifier, and is followed by
2806 something that cannot possibly match the quantified thing, the quantifier
2807 is automatically "possessified".
2808
2809 24. A recursive reference to a subpattern whose number was greater than 39
2810 went wrong under certain circumstances in UTF-8 mode. This bug could also
2811 have affected the operation of pcre_study().
2812
2813 25. Realized that a little bit of performance could be had by replacing
2814 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2815
2816 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2817
2818 27. Possessive quantifiers such as a++ were previously implemented by turning
2819 them into atomic groups such as ($>a+). Now they have their own opcodes,
2820 which improves performance. This includes the automatically created ones
2821 from 23 above.
2822
2823 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2824 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2825 the first matched character to be a colon. This applied both to named and
2826 numbered groups.
2827
2828 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2829
2830 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2831 building PCRE on Windows easier, so I have included it in the distribution.
2832
2833 31. There is now a check in pcretest against a ridiculously large number being
2834 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2835 loop, the loop is abandoned.
2836
2837 32. Forward references to subpatterns in conditions such as (?(2)...) where
2838 subpattern 2 is defined later cause pcre_compile() to search forwards in
2839 the pattern for the relevant set of parentheses. This search went wrong
2840 when there were unescaped parentheses in a character class, parentheses
2841 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2842
2843 33. "Subroutine" calls and backreferences were previously restricted to
2844 referencing subpatterns earlier in the regex. This restriction has now
2845 been removed.
2846
2847 34. Added a number of extra features that are going to be in Perl 5.10. On the
2848 whole, these are just syntactic alternatives for features that PCRE had
2849 previously implemented using the Python syntax or my own invention. The
2850 other formats are all retained for compatibility.
2851
2852 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2853 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2854 also .NET compatible.
2855
2856 (b) A recursion or subroutine call to a named group can now be defined as
2857 (?&name) as well as (?P>name).
2858
2859 (c) A backreference to a named group can now be defined as \k<name> or
2860 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2861 5.10, are also .NET compatible.
2862
2863 (d) A conditional reference to a named group can now use the syntax
2864 (?(<name>) or (?('name') as well as (?(name).
2865
2866 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2867 groups (named and numbered) that are never evaluated inline, but can be
2868 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2869 is always false. There may be only one alternative in such a group.
2870
2871 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2872 as the simple (?(R). The condition is true only if the most recent
2873 recursion is that of the given number or name. It does not search out
2874 through the entire recursion stack.
2875
2876 (g) The escape \gN or \g{N} has been added, where N is a positive or
2877 negative number, specifying an absolute or relative reference.
2878
2879 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2880 some "unreachable code" warnings.
2881
2882 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2883 things, this adds five new scripts.
2884
2885 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2886 There were also incompatibilities regarding the handling of \Q..\E inside
2887 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2888 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2889
2890 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2891 matches an empty string, and forcibly breaks the loop. There were bugs in
2892 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2893 against aaaa the result was just "a" rather than "aaaa", for example. Two
2894 separate and independent bugs (that affected different cases) have been
2895 fixed.
2896
2897 39. Refactored the code to abolish the use of different opcodes for small
2898 capturing bracket numbers. This is a tidy that I avoided doing when I
2899 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2900 The new approach is not only tidier, it makes it possible to reduce the
2901 memory needed to fix the previous bug (38).
2902
2903 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2904 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2905 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2906 mode.
2907
2908 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2909 report.
2910
2911 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2912 copy construction and assignment in the C++ wrapper.
2913
2914 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2915 couple of bugs that could have given wrong results in the "--newline=crlf"
2916 case.
2917
2918 44. Added a number of casts and did some reorganization of signed/unsigned int
2919 variables following suggestions from Dair Grant. Also renamed the variable
2920 "this" as "item" because it is a C++ keyword.
2921
2922 45. Arranged for dftables to add
2923
2924 #include "pcre_internal.h"
2925
2926 to pcre_chartables.c because without it, gcc 4.x may remove the array
2927 definition from the final binary if PCRE is built into a static library and
2928 dead code stripping is activated.
2929
2930 46. For an unanchored pattern, if a match attempt fails at the start of a
2931 newline sequence, and the newline setting is CRLF or ANY, and the next two
2932 characters are CRLF, advance by two characters instead of one.
2933
2934
2935 Version 6.7 04-Jul-06
2936 ---------------------
2937
2938 1. In order to handle tests when input lines are enormously long, pcretest has
2939 been re-factored so that it automatically extends its buffers when
2940 necessary. The code is crude, but this _is_ just a test program. The
2941 default size has been increased from 32K to 50K.
2942
2943 2. The code in pcre_study() was using the value of the re argument before
2944 testing it for NULL. (Of course, in any sensible call of the function, it
2945 won't be NULL.)
2946
2947 3. The memmove() emulation function in pcre_internal.h, which is used on
2948 systems that lack both memmove() and bcopy() - that is, hardly ever -
2949 was missing a "static" storage class specifier.
2950
2951 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2952 containing an extended class (one that cannot be represented by a bitmap
2953 because it contains high-valued characters or Unicode property items, e.g.
2954 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2955 pattern, but PCRE should not loop if you do not (it no longer does).
2956 [Detail: two cases were found: (a) a repeated subpattern containing an
2957 extended class; (b) a recursive reference to a subpattern that followed a
2958 previous extended class. It wasn't skipping over the extended class
2959 correctly when UTF-8 mode was not set.]
2960
2961 5. A negated single-character class was not being recognized as fixed-length
2962 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2963 compile error "lookbehind assertion is not fixed length".
2964
2965 6. The RunPerlTest auxiliary script was showing an unexpected difference
2966 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2967 write a Perl script that can interpret lines of an input file either as
2968 byte characters or as UTF-8, which is what "perltest" was being required to
2969 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2970 can't do is switch easily at run time between having the "use utf8;" pragma
2971 or not. In the end, I fudged it by using the RunPerlTest script to insert
2972 "use utf8;" explicitly for the UTF-8 tests.
2973
2974 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2975 the end of the subject string, contrary to the documentation and to what
2976 Perl does. This was true of both matching functions. Now it matches only at
2977 the start of the subject and immediately after *internal* newlines.
2978
2979 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2980 a pointer to an int instead of a pointer to an unsigned long int. This
2981 caused problems on 64-bit systems.
2982
2983 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2984 instance of the 'standard' template library not being so standard".
2985
2986 10. There was no check on the number of named subpatterns nor the maximum
2987 length of a subpattern name. The product of these values is used to compute
2988 the size of the memory block for a compiled pattern. By supplying a very
2989 long subpattern name and a large number of named subpatterns, the size
2990 computation could be caused to overflow. This is now prevented by limiting
2991 the length of names to 32 characters, and the number of named subpatterns
2992 to 10,000.
2993
2994 11. Subpatterns that are repeated with specific counts have to be replicated in
2995 the compiled pattern. The size of memory for this was computed from the
2996 length of the subpattern and the repeat count. The latter is limited to
2997 65535, but there was no limit on the former, meaning that integer overflow
2998 could in principle occur. The compiled length of a repeated subpattern is
2999 now limited to 30,000 bytes in order to prevent this.
3000
3001 12. Added the optional facility to have named substrings with the same name.
3002
3003 13. Added the ability to use a named substring as a condition, using the
3004 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3005 are numbers (not recommended). Forward references are permitted.
3006
3007 14. Added forward references in named backreferences (if you see what I mean).
3008
3009 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3010 pattern could run off the end of the subject. For example, the pattern
3011 "(?s)(.{1,5})"8 did this with the subject "ab".
3012
3013 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3014 PCRE_CASELESS was set when matching characters that were quantified with ?
3015 or *.
3016
3017 17. A character class other than a single negated character that had a minimum
3018 but no maximum quantifier - for example [ab]{6,} - was not handled
3019 correctly by pce_dfa_exec(). It would match only one character.
3020
3021 18. A valid (though odd) pattern that looked like a POSIX character
3022 class but used an invalid character after [ (for example [[,abc,]]) caused
3023 pcre_compile() to give the error "Failed: internal error: code overflow" or
3024 in some cases to crash with a glibc free() error. This could even happen if
3025 the pattern terminated after [[ but there just happened to be a sequence of
3026 letters, a binary zero, and a closing ] in the memory that followed.
3027
3028 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3029 over the years. Originally (before any Unicode support), just the bottom 8
3030 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3031 output from "man perlunicode" includes this:
3032
3033 The regular expression compiler produces polymorphic opcodes. That
3034 is, the pattern adapts to the data and automatically switches to
3035 the Unicode character scheme when presented with Unicode data--or
3036 instead uses a traditional byte scheme when presented with byte
3037 data.
3038
3039 Sadly, a wide octal escape does not cause a switch, and in a string with
3040 no other multibyte characters, these octal escapes are treated as before.
3041 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3042 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3043 Unicode string.
3044
3045 I have not perpetrated such confusion in PCRE. Up till now, it took just
3046 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3047 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3048 translate to the appropriate multibyte character.
3049
3050 29. Applied some refactoring to reduce the number of warnings from Microsoft
3051 and Borland compilers. This has included removing the fudge introduced
3052 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3053 a warning about an unused variable.
3054
3055 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3056 characters since release 4.0, because Perl (from release 5.004) does not.
3057 [Or at least, is documented not to: some releases seem to be in conflict
3058 with the documentation.] However, when a pattern was studied with
3059 pcre_study() and all its branches started with \s, PCRE still included VT
3060 as a possible starting character. Of course, this did no harm; it just
3061 caused an unnecessary match attempt.
3062
3063 22. Removed a now-redundant internal flag bit that recorded the fact that case
3064 dependency changed within the pattern. This was once needed for "required
3065 byte" processing, but is no longer used. This recovers a now-scarce options
3066 bit. Also moved the least significant internal flag bit to the most-
3067 significant bit of the word, which was not previously used (hangover from
3068 the days when it was an int rather than a uint) to free up another bit for
3069 the future.
3070
3071 23. Added support for CRLF line endings as well as CR and LF. As well as the
3072 default being selectable at build time, it can now be changed at runtime
3073 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3074 specify that it is scanning data with non-default line endings.
3075
3076 24. Changed the definition of CXXLINK to make it agree with the definition of
3077 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3078
3079 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3080 recursions. This makes a big different to stack usage for some patterns.
3081
3082 26. If a subpattern containing a named recursion or subroutine reference such
3083 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3084 the space required for the compiled pattern went wrong and gave too small a
3085 value. Depending on the environment, this could lead to "Failed: internal
3086 error: code overflow at offset 49" or "glibc detected double free or
3087 corruption" errors.
3088
3089 27. Applied patches from Google (a) to support the new newline modes and (b) to
3090 advance over multibyte UTF-8 characters in GlobalReplace.
3091
3092 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3093 difference for some implementation of PCRE in some Windows version.
3094
3095 29. Added some extra testing facilities to pcretest:
3096
3097 \q<number> in a data line sets the "match limit" value
3098 \Q<number> in a data line sets the "match recursion limt" value
3099 -S <number> sets the stack size, where <number> is in megabytes
3100
3101 The -S option isn't available for Windows.
3102
3103
3104 Version 6.6 06-Feb-06
3105 ---------------------
3106
3107 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3108 in pcreposix.h. I have copied the definition from pcre.h.
3109
3110 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3111 because pcre.h is no longer a built file.
3112
3113 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3114 not normally included in the compiled code.
3115
3116
3117 Version 6.5 01-Feb-06
3118 ---------------------
3119
3120 1. When using the partial match feature with pcre_dfa_exec(), it was not
3121 anchoring the second and subsequent partial matches at the new starting
3122 point. This could lead to incorrect results. For example, with the pattern
3123 /1234/, partially matching against "123" and then "a4" gave a match.
3124
3125 2. Changes to pcregrep:
3126
3127 (a) All non-match returns from pcre_exec() were being treated as failures
3128 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3129 error message is output. Some extra information is given for the
3130 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3131 probably the only errors that are likely to be caused by users (by
3132 specifying a regex that has nested indefinite repeats, for instance).
3133 If there are more than 20 of these errors, pcregrep is abandoned.
3134
3135 (b) A binary zero was treated as data while matching, but terminated the
3136 output line if it was written out. This has been fixed: binary zeroes
3137 are now no different to any other data bytes.
3138
3139 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3140 used to set a locale for matching. The --locale=xxxx long option has
3141 been added (no short equivalent) to specify a locale explicitly on the
3142 pcregrep command, overriding the environment variables.
3143
3144 (d) When -B was used with -n, some line numbers in the output were one less
3145 than they should have been.
3146
3147 (e) Added the -o (--only-matching) option.
3148
3149 (f) If -A or -C was used with -c (count only), some lines of context were
3150 accidentally printed for the final match.
3151
3152 (g) Added the -H (--with-filename) option.
3153
3154 (h) The combination of options -rh failed to suppress file names for files
3155 that were found from directory arguments.
3156
3157 (i) Added the -D (--devices) and -d (--directories) options.
3158
3159 (j) Added the -F (--fixed-strings) option.
3160
3161 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3162
3163 (l) Added the --colo(u)r option.
3164
3165 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3166 is not present by default.
3167
3168 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3169 items such as (?R) or (?1), when the recursion could match a number of
3170 alternatives. If it matched one of the alternatives, but subsequently,
3171 outside the recursion, there was a failure, the code tried to back up into
3172 the recursion. However, because of the way PCRE is implemented, this is not
3173 possible, and the result was an incorrect result from the match.
3174
3175 In order to prevent this happening, the specification of recursion has
3176 been changed so that all such subpatterns are automatically treated as
3177 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3178
3179 4. I had overlooked the fact that, in some locales, there are characters for
3180 which isalpha() is true but neither isupper() nor islower() are true. In
3181 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3182 and ordfeminine) are like this. This affected the treatment of \w and \W
3183 when they appeared in character classes, but not when they appeared outside
3184 a character class. The bit map for "word" characters is now created
3185 separately from the results of isalnum() instead of just taking it from the
3186 upper, lower, and digit maps. (Plus the underscore character, of course.)
3187
3188 5. The above bug also affected the handling of POSIX character classes such as
3189 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3190 permanent tables. Instead, the bit maps for such a class were previously
3191 created as the appropriate unions of the upper, lower, and digit bitmaps.
3192 Now they are created by subtraction from the [[:word:]] class, which has
3193 its own bitmap.
3194
3195 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3196 It is created by subtracting the vertical space characters (\x09, \x0a,
3197 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3198 subtraction was done in the overall bitmap for a character class, meaning
3199 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3200 be recognized. This bug has been fixed.
3201
3202 7. Patches from the folks at Google:
3203
3204 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3205 real life, but is still worth protecting against".
3206
3207 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3208 regular expressions".
3209
3210 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3211 have it.
3212
3213 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3214 "configure" and the latter not, in order to fix a problem somebody had
3215 with compiling the Arg class on HP-UX.
3216
3217 (e) Improve the error-handling of the C++ wrapper a little bit.
3218
3219 (f) New tests for checking recursion limiting.
3220
3221 8. The pcre_memmove() function, which is used only if the environment does not
3222 have a standard memmove() function (and is therefore rarely compiled),
3223 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3224 returning a result (though PCRE never actually uses the result).
3225
3226 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3227 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3228 returned instead of calling malloc() with an overflowing number that would
3229 most likely cause subsequent chaos.
3230
3231 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3232
3233 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3234 with this option is matched, the nmatch and pmatch options of regexec() are
3235 ignored.
3236
3237 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3238 provided in case anyone wants to the the POSIX interface with UTF-8
3239 strings.
3240
3241 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3242 C++ linking (needed for some HP-UX environments).
3243
3244 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3245 (unused parameter) and in the pcre_printint() function (omitted "default"
3246 switch label when the default is to do nothing).
3247
3248 15. Added some code to make it possible, when PCRE is compiled as a C++
3249 library, to replace subject pointers for pcre_exec() with a smart pointer
3250 class, thus making it possible to process discontinuous strings.
3251
3252 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3253 much the same function. They were added by different people who were trying
3254 to make PCRE easy to compile on non-Unix systems. It has been suggested
3255 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3256 for compiling on Windows systems. I have therefore replaced it with
3257 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3258 defaults to "extern" for C or "extern C" for C++, which works fine on
3259 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3260 SCOPE with something explicit in config.h. In addition:
3261
3262 (a) pcreposix.h still had just "extern" instead of either of these macros;
3263 I have replaced it with PCRE_DATA_SCOPE.
3264
3265 (b) Functions such as _pcre_xclass(), which are internal to the library,
3266 but external in the C sense, all had PCRE_EXPORT in their definitions.
3267 This is apparently wrong for the Windows case, so I have removed it.
3268 (It makes no difference on Unix-like systems.)
3269
3270 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3271 of recursive calls to match(). This is different to MATCH_LIMIT because
3272 that limits the total number of calls to match(), not all of which increase
3273 the depth of recursion. Limiting the recursion depth limits the amount of
3274 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3275 when PCRE is compiled, and changed at run time. A patch from Google adds
3276 this functionality to the C++ interface.
3277
3278 18. Changes to the handling of Unicode character properties:
3279
3280 (a) Updated the table to Unicode 4.1.0.
3281
3282 (b) Recognize characters that are not in the table as "Cn" (undefined).
3283
3284 (c) I revised the way the table is implemented to a much improved format
3285 which includes recognition of ranges. It now supports the ranges that
3286 are defined in UnicodeData.txt, and it also amalgamates other
3287 characters into ranges. This has reduced the number of entries in the
3288 table from around 16,000 to around 3,000, thus reducing its size
3289 considerably. I realized I did not need to use a tree structure after
3290 all - a binary chop search is just as efficient. Having reduced the
3291 number of entries, I extended their size from 6 bytes to 8 bytes to
3292 allow for more data.
3293
3294 (d) Added support for Unicode script names via properties such as \p{Han}.
3295
3296 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3297 matching that character.
3298
3299 20. When matching a repeated Unicode property with a minimum greater than zero,
3300 (for example \pL{2,}), PCRE could look past the end of the subject if it
3301 reached it while seeking the minimum number of characters. This could
3302 happen only if some of the characters were more than one byte long, because
3303 there is a check for at least the minimum number of bytes.
3304
3305 21. Refactored the implementation of \p and \P so as to be more general, to
3306 allow for more different types of property in future. This has changed the
3307 compiled form incompatibly. Anybody with saved compiled patterns that use
3308 \p or \P will have to recompile them.
3309
3310 22. Added "Any" and "L&" to the supported property types.
3311
3312 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3313 but give a compile time error if the value is greater than 0xff.
3314
3315 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3316 accidentally not being installed or uninstalled.
3317
3318 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3319 made were to insert the current release number. This seemed silly, because
3320 it made things harder for people building PCRE on systems that don't run
3321 "configure". I have turned pcre.h into a distributed file, no longer built
3322 by "configure", with the version identification directly included. There is
3323 no longer a pcre.h.in file.
3324
3325 However, this change necessitated a change to the pcre-config script as
3326 well. It is built from pcre-config.in, and one of the substitutions was the
3327 release number. I have updated configure.ac so that ./configure now finds
3328 the release number by grepping pcre.h.
3329
3330 26. Added the ability to run the tests under valgrind.
3331
3332
3333 Version 6.4 05-Sep-05
3334 ---------------------
3335
3336 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3337 "--" to be printed when multiple files were scanned, even when none of the
3338 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3339 consider it to be a bug, and have restored the previous behaviour.
3340
3341 2. A couple of code tidies to get rid of compiler warnings.
3342
3343 3. The pcretest program used to cheat by referring to symbols in the library
3344 whose names begin with _pcre_. These are internal symbols that are not
3345 really supposed to be visible externally, and in some environments it is
3346 possible to suppress them. The cheating is now confined to including
3347 certain files from the library's source, which is a bit cleaner.
3348
3349 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3350 file's purpose clearer.
3351
3352 5. Reorganized pcre_ucp_findchar().
3353
3354
3355 Version 6.3 15-Aug-05
3356 ---------------------
3357
3358 1. The file libpcre.pc.in did not have general read permission in the tarball.
3359
3360 2. There were some problems when building without C++ support:
3361
3362 (a) If C++ support was not built, "make install" and "make test" still
3363 tried to test it.
3364
3365 (b) There were problems when the value of CXX was explicitly set. Some
3366 changes have been made to try to fix these, and ...
3367
3368 (c) --disable-cpp can now be used to explicitly disable C++ support.
3369
3370 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3371 backslash in a target when C++ was disabled. This confuses some
3372 versions of "make", apparently. Using an intermediate variable solves
3373 this. (Same for CPP_LOBJ.)
3374
3375 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3376 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3377 necessary on certain architectures.
3378
3379 4. Added a setting of -export-symbols-regex to the link command to remove
3380 those symbols that are exported in the C sense, but actually are local
3381 within the library, and not documented. Their names all begin with
3382 "_pcre_". This is not a perfect job, because (a) we have to except some
3383 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3384 available (and never for static libraries). I have made a note to try to
3385 find a way round (a) in the future.
3386
3387
3388 Version 6.2 01-Aug-05
3389 ---------------------
3390
3391 1. There was no test for integer overflow of quantifier values. A construction
3392 such as {1111111111111111} would give undefined results. What is worse, if
3393 a minimum quantifier for a parenthesized subpattern overflowed and became
3394 negative, the calculation of the memory size went wrong. This could have
3395 led to memory overwriting.
3396
3397 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3398
3399 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3400 operating environments where this matters.
3401
3402 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3403 PCRE options from within the C++ wrapper.
3404
3405 5. Named capturing subpatterns were not being correctly counted when a pattern
3406 was compiled. This caused two problems: (a) If there were more than 100
3407 such subpatterns, the calculation of the memory needed for the whole
3408 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3409 back references of the form \12, where the number was greater than 9, were
3410 not recognized as back references, even though there were sufficient
3411 previous subpatterns.
3412
3413 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3414 versions of gcc, e.g. 2.95.4.
3415
3416
3417 Version 6.1 21-Jun-05
3418 ---------------------
3419
3420 1. There was one reference to the variable "posix" in pcretest.c that was not
3421 surrounded by "#if !defined NOPOSIX".
3422
3423 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3424 the cross-check on the old pcre_info() function, for the benefit of the
3425 cut-down version of PCRE that is currently imported into Exim.
3426
3427 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3428 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3429 patterns that start (?i) but I don't think that matters. The use of (?i) is
3430 just an example; this all applies to the other options as well.
3431
3432 4. Since libtool seems to echo the compile commands it is issuing, the output
3433 from "make" can be reduced a bit by putting "@" in front of each libtool
3434 compile command.
3435
3436 5. Patch from the folks at Google for configure.in to be a bit more thorough
3437 in checking for a suitable C++ installation before trying to compile the
3438 C++ stuff. This should fix a reported problem when a compiler was present,
3439 but no suitable headers.
3440
3441 6. The man pages all had just "PCRE" as their title. I have changed them to
3442 be the relevant file name. I have also arranged that these names are
3443 retained in the file doc/pcre.txt, which is a concatenation in text format
3444 of all the man pages except the little individual ones for each function.
3445
3446 7. The NON-UNIX-USE file had not been updated for the different set of source
3447 files that come with release 6. I also added a few comments about the C++
3448 wrapper.
3449
3450
3451 Version 6.0 07-Jun-05
3452 ---------------------
3453
3454 1. Some minor internal re-organization to help with my DFA experiments.
3455
3456 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3457 didn't matter for the library itself when fully configured, but did matter
3458 when compiling without UCP support, or within Exim, where the ucp files are
3459 not imported.
3460
3461 3. Refactoring of the library code to split up the various functions into
3462 different source modules. The addition of the new DFA matching code (see
3463 below) to a single monolithic source would have made it really too
3464 unwieldy, quite apart from causing all the code to be include in a
3465 statically linked application, when only some functions are used. This is
3466 relevant even without the DFA addition now that patterns can be compiled in
3467 one application and matched in another.
3468
3469 The downside of splitting up is that there have to be some external
3470 functions and data tables that are used internally in different modules of
3471 the library but which are not part of the API. These have all had their
3472 names changed to start with "_pcre_" so that they are unlikely to clash
3473 with other external names.
3474
3475 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3476 a different (DFA) algorithm. Although it is slower than the original
3477 function, it does have some advantages for certain types of matching
3478 problem.
3479
3480 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3481 including restarting after a partial match.
3482
3483 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3484 defined when compiling for Windows was sent to me. I have put it into the
3485 code, though I have no means of testing or verifying it.
3486
3487 7. Added the pcre_refcount() auxiliary function.
3488
3489 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3490 match before or at the first newline in the subject string. In pcretest,
3491 the /f option on a pattern can be used to set this.
3492
3493 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3494 would behave wrongly. This has been present in PCRE since release 4.0.
3495
3496 10. A number of changes to the pcregrep command:
3497
3498 (a) Refactored how -x works; insert ^(...)$ instead of setting
3499 PCRE_ANCHORED and checking the length, in preparation for adding
3500 something similar for -w.
3501
3502 (b) Added the -w (match as a word) option.
3503
3504 (c) Refactored the way lines are read and buffered so as to have more
3505 than one at a time available.
3506
3507 (d) Implemented a pcregrep test script.
3508
3509 (e) Added the -M (multiline match) option. This allows patterns to match
3510 over several lines of the subject. The buffering ensures that at least
3511 8K, or the rest of the document (whichever is the shorter) is available
3512 for matching (and similarly the previous 8K for lookbehind assertions).
3513
3514 (f) Changed the --help output so that it now says
3515
3516 -w, --word-regex(p)
3517
3518 instead of two lines, one with "regex" and the other with "regexp"
3519 because that confused at least one person since the short forms are the
3520 same. (This required a bit of code, as the output is generated
3521 automatically from a table. It wasn't just a text change.)
3522
3523 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3524 option but starts with a hyphen. Could be a pattern or a path name
3525 starting with a hyphen, for instance.
3526
3527 (h) "-" can be given as a file name to represent stdin.
3528
3529 (i) When file names are being printed, "(standard input)" is used for
3530 the standard input, for compatibility with GNU grep. Previously
3531 "<stdin>" was used.
3532
3533 (j) The option --label=xxx can be used to supply a name to be used for
3534 stdin when file names are being printed. There is no short form.
3535
3536 (k) Re-factored the options decoding logic because we are going to add
3537 two more options that take data. Such options can now be given in four
3538 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3539
3540 (l) Added the -A, -B, and -C options for requesting that lines of context
3541 around matches be printed.
3542
3543 (m) Added the -L option to print the names of files that do not contain
3544 any matching lines, that is, the complement of -l.
3545
3546 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3547 continue to scan other files.
3548
3549 (o) The -s option was incorrectly implemented. For compatibility with other
3550 greps, it now suppresses the error message for a non-existent or non-
3551 accessible file (but not the return code). There is a new option called
3552 -q that suppresses the output of matching lines, which was what -s was
3553 previously doing.
3554
3555 (p) Added --include and --exclude options to specify files for inclusion
3556 and exclusion when recursing.
3557
3558 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3559 Hopefully, it now does.
3560
3561 12. Missing cast in pcre_study().
3562
3563 13. Added an "uninstall" target to the makefile.
3564
3565 14. Replaced "extern" in the function prototypes in Makefile.in with
3566 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3567 world, but is set differently for Windows.
3568
3569 15. Added a second compiling function called pcre_compile2(). The only
3570 difference is that it has an extra argument, which is a pointer to an
3571 integer error code. When there is a compile-time failure, this is set
3572 non-zero, in addition to the error test pointer being set to point to an
3573 error message. The new argument may be NULL if no error number is required
3574 (but then you may as well call pcre_compile(), which is now just a
3575 wrapper). This facility is provided because some applications need a
3576 numeric error indication, but it has also enabled me to tidy up the way
3577 compile-time errors are handled in the POSIX wrapper.
3578
3579 16. Added VPATH=.libs to the makefile; this should help when building with one
3580 prefix path and installing with another. (Or so I'm told by someone who
3581 knows more about this stuff than I do.)
3582
3583 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3584 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3585 match everything, including newlines. This is not POSIX-compatible, but
3586 somebody wanted the feature. From pcretest it can be activated by using
3587 both the P and the s flags.
3588
3589 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3590
3591 19. libpcre.pc was being incorrectly installed as executable.
3592
3593 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3594 it now also looks for '\r' so that it will work unmodified on Windows.
3595
3596 21. Added Google's contributed C++ wrapper to the distribution.
3597
3598 22. Added some untidy missing memory free() calls in pcretest, to keep
3599 Electric Fence happy when testing.
3600
3601
3602
3603 Version 5.0 13-Sep-04
3604 ---------------------
3605
3606 1. Internal change: literal characters are no longer packed up into items
3607 containing multiple characters in a single byte-string. Each character
3608 is now matched using a separate opcode. However, there may be more than one
3609 byte in the character in UTF-8 mode.
3610
3611 2. The pcre_callout_block structure has two new fields: pattern_position and
3612 next_item_length. These contain the offset in the pattern to the next match
3613 item, and its length, respectively.
3614
3615 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3616 insertion of callouts before each pattern item. Added the /C option to
3617 pcretest to make use of this.
3618
3619 4. On the advice of a Windows user, the lines
3620
3621 #if defined(_WIN32) || defined(WIN32)
3622 _setmode( _fileno( stdout ), 0x8000 );
3623 #endif /* defined(_WIN32) || defined(WIN32) */
3624
3625 have been added to the source of pcretest. This apparently does useful
3626 magic in relation to line terminators.
3627
3628 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3629 for the benefit of those environments where the "b" makes a difference.
3630
3631 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3632 to know about it. I have put a hack into configure.in that adds in code
3633 to set GCC=yes if CC=icc. This seems to end up at a point in the
3634 generated configure script that is early enough to affect the setting of
3635 compiler options, which is what is needed, but I have no means of testing
3636 whether it really works. (The user who reported this had patched the
3637 generated configure script, which of course I cannot do.)
3638
3639 LATER: After change 22 below (new libtool files), the configure script
3640 seems to know about icc (and also ecc). Therefore, I have commented out
3641 this hack in configure.in.
3642
3643 7. Added support for pkg-config (2 patches were sent in).
3644
3645 8. Negated POSIX character classes that used a combination of internal tables
3646 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3647 [[:^ascii]]. Typically, they would match almost any characters. The other
3648 POSIX classes were not broken in this way.
3649
3650 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3651 to find the match, as PCRE was deluded into thinking that the match had to
3652 start at the start point or following a newline. The same bug applied to
3653 patterns with negative forward assertions or any backward assertions
3654 preceding ".*" at the start, unless the pattern required a fixed first
3655 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3656
3657 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3658 starting at the last subject character, bytes beyond the end of the subject
3659 string were read.
3660
3661 11. Renamed the variable "class" as "classbits" to make life easier for C++
3662 users. (Previously there was a macro definition, but it apparently wasn't
3663 enough.)
3664
3665 12. Added the new field "tables" to the extra data so that tables can be passed
3666 in at exec time, or the internal tables can be re-selected. This allows
3667 a compiled regex to be saved and re-used at a later time by a different
3668 program that might have everything at different addresses.
3669
3670 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3671 -R library as well as a -L library.
3672
3673 14. The debugging options of pcretest (-d on the command line or D on a
3674 pattern) showed incorrect output for anything following an extended class
3675 that contained multibyte characters and which was followed by a quantifier.
3676
3677 15. Added optional support for general category Unicode character properties
3678 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3679 support. It adds about 90K to the size of the library. The meanings of the
3680 inbuilt class escapes such as \d and \s have NOT been changed.
3681
3682 16. Updated pcredemo.c to include calls to free() to release the memory for the
3683 compiled pattern.
3684
3685 17. The generated file chartables.c was being created in the source directory
3686 instead of in the building directory. This caused the build to fail if the
3687 source directory was different from the building directory, and was
3688 read-only.
3689
3690 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3691 file. No doubt somebody will tell me if they don't make sense... Also added
3692 Dan Mooney's comments about building on OpenVMS.
3693
3694 19. Added support for partial matching via the PCRE_PARTIAL option for
3695 pcre_exec() and the \P data escape in pcretest.
3696
3697 20. Extended pcretest with 3 new pattern features:
3698
3699 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3700 write the compiled pattern to the file whose name is "rest-of-line".
3701 This is a straight binary dump of the data, with the saved pointer to
3702 the character tables forced to be NULL. The study data, if any, is
3703 written too. After writing, pcretest reads a new pattern.
3704
3705 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3706 compiled pattern from the given file. There must not be any
3707 occurrences of "<" in the file name (pretty unlikely); if there are,
3708 pcretest will instead treat the initial "<" as a pattern delimiter.
3709 After reading in the pattern, pcretest goes on to read data lines as
3710 usual.
3711
3712 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3713 and 16-bit fields in a compiled pattern, to simulate a pattern that
3714 was compiled on a host of opposite endianness.
3715
3716 21. The pcre-exec() function can now cope with patterns that were compiled on
3717 hosts of opposite endianness, with this restriction:
3718
3719 As for any compiled expression that is saved and used later, the tables
3720 pointer field cannot be preserved; the extra_data field in the arguments
3721 to pcre_exec() should be used to pass in a tables address if a value
3722 other than the default internal tables were used at compile time.
3723
3724 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3725 now diagnosed as an error. Previously, most of the time, a negative number
3726 would have been treated as zero, but if in addition "ovector" was passed as
3727 NULL, a crash could occur.
3728
3729 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3730 new versions from the libtool 1.5 distribution (the last one is a copy of
3731 a file called libtool.m4). This seems to have fixed the need to patch
3732 "configure" to support Darwin 1.3 (which I used to do). However, I still
3733 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3734 workstation).
3735
3736 24. Changed the PCRE licence to be the more standard "BSD" licence.
3737
3738
3739 Version 4.5 01-Dec-03
3740 ---------------------
3741
3742 1. There has been some re-arrangement of the code for the match() function so
3743 that it can be compiled in a version that does not call itself recursively.
3744 Instead, it keeps those local variables that need separate instances for
3745 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3746 needs to "recurse". Keeping track of where control must go is done by means
3747 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3748 hide most of the details from the main code, and operates only if
3749 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3750 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3751 operating.
3752
3753 To make it easier for callers to provide specially tailored get/free
3754 functions for this usage, two new functions, pcre_stack_malloc, and
3755 pcre_stack_free, are used. They are always called in strict stacking order,
3756 and the size of block requested is always the same.
3757
3758 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3759 PCRE has been compiled to use the stack or the heap for recursion. The
3760 -C option of pcretest uses this to show which version is compiled.
3761
3762 A new data escape \S, is added to pcretest; it causes the amounts of store
3763 obtained and freed by both kinds of malloc/free at match time to be added
3764 to the output.
3765
3766 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3767 what's available on my current Linux desktop machine.
3768
3769 3. When matching a UTF-8 string, the test for a valid string at the start has
3770 been extended. If start_offset is not zero, PCRE now checks that it points
3771 to a byte that is the start of a UTF-8 character. If not, it returns
3772 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3773 this is necessary because there may be backward assertions in the pattern.
3774 When matching the same subject several times, it may save resources to use
3775 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3776
3777 4. The code for checking the validity of UTF-8 strings has been tightened so
3778 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3779 containing "overlong sequences".
3780
3781 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3782 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3783 should have been "&", but it just so happened that all the cases this let
3784 through by mistake were picked up later in the function.
3785
3786 6. I had used a variable called "isblank" - this is a C99 function, causing
3787 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3788
3789 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3790 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3791 so the version is shown; (c) stop "make test" echoing "./RunTest".
3792
3793 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3794
3795 9. The prototype for memmove() for systems that don't have it was using
3796 size_t, but the inclusion of the header that defines size_t was later. I've
3797 moved the #includes for the C headers earlier to avoid this.
3798
3799 10. Added some adjustments to the code to make it easier to compiler on certain
3800 special systems:
3801
3802 (a) Some "const" qualifiers were missing.
3803 (b) Added the macro EXPORT before all exported functions; by default this
3804 is defined to be empty.
3805 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3806 that it reads its output file name as an argument instead of writing
3807 to the standard output and assuming this can be redirected.
3808
3809 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3810 class containing characters with values greater than 255, PCRE compilation
3811 went into a loop.
3812
3813 12. A recursive reference to a subpattern that was within another subpattern
3814 that had a minimum quantifier of zero caused PCRE to crash. For example,
3815 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3816 recursion. If the recursively-called subpattern itself had a zero repeat,
3817 that was OK.
3818
3819 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3820 buffer into which it was copied (for escape processing) was still set at
3821 1024, so long lines caused crashes.
3822
3823 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3824 "internal error: code overflow...". This applied to any character class
3825 that was followed by a possessive quantifier.
3826
3827 15. Modified the Makefile to add libpcre.la as a prerequisite for
3828 libpcreposix.la because I was told this is needed for a parallel build to
3829 work.
3830
3831 16. If a pattern that contained .* following optional items at the start was
3832 studied, the wrong optimizing data was generated, leading to matching
3833 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3834 matching string must start with a or b or c. The correct conclusion for
3835 this pattern is that a match can start with any character.
3836
3837
3838 Version 4.4 13-Aug-03
3839 ---------------------
3840
3841 1. In UTF-8 mode, a character class containing characters with values between
3842 127 and 255 was not handled correctly if the compiled pattern was studied.
3843 In fixing this, I have also improved the studying algorithm for such
3844 classes (slightly).
3845
3846 2. Three internal functions had redundant arguments passed to them. Removal
3847 might give a very teeny performance improvement.
3848
3849 3. Documentation bug: the value of the capture_top field in a callout is *one
3850 more than* the number of the hightest numbered captured substring.
3851
3852 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3853 in incorrectly linking with a previously installed version. They now link
3854 explicitly with libpcre.la.
3855
3856 5. configure.in no longer needs to recognize Cygwin specially.
3857
3858 6. A problem in pcre.in for Windows platforms is fixed.
3859
3860 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3861 pcretest, it used to include the size of the study block as part of its
3862 output. Unfortunately, the structure contains a field that has a different
3863 size on different hardware architectures. This meant that the tests that
3864 showed this size failed. As the block is currently always of a fixed size,
3865 this information isn't actually particularly useful in pcretest output, so
3866 I have just removed it.
3867
3868 8. Three pre-processor statements accidentally did not start in column 1.
3869 Sadly, there are *still* compilers around that complain, even though
3870 standard C has not required this for well over a decade. Sigh.
3871
3872 9. In pcretest, the code for checking callouts passed small integers in the
3873 callout_data field, which is a void * field. However, some picky compilers
3874 complained about the casts involved for this on 64-bit systems. Now
3875 pcretest passes the address of the small integer instead, which should get
3876 rid of the warnings.
3877
3878 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3879 both compile and run time, and gives an error if an invalid UTF-8 sequence
3880 is found. There is a option for disabling this check in cases where the
3881 string is known to be correct and/or the maximum performance is wanted.
3882
3883 11. In response to a bug report, I changed one line in Makefile.in from
3884
3885 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3886 to
3887 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3888
3889 to look similar to other lines, but I have no way of telling whether this
3890 is the right thing to do, as I do not use Windows. No doubt I'll get told
3891 if it's wrong...
3892
3893
3894 Version 4.3 21-May-03
3895 ---------------------
3896
3897 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3898 Makefile.
3899
3900 2. Some refactoring to improve the quality of the code:
3901
3902 (i) The utf8_table... variables are now declared "const".
3903
3904 (ii) The code for \cx, which used the "case flipping" table to upper case
3905 lower case letters, now just substracts 32. This is ASCII-specific,
3906 but the whole concept of \cx is ASCII-specific, so it seems
3907 reasonable.
3908
3909 (iii) PCRE was using its character types table to recognize decimal and
3910 hexadecimal digits in the pattern. This is silly, because it handles
3911 only 0-9, a-f, and A-F, but the character types table is locale-
3912 specific, which means strange things might happen. A private
3913 table is now used for this - though it costs 256 bytes, a table is
3914 much faster than multiple explicit tests. Of course, the standard
3915 character types table is still used for matching digits in subject
3916 strings against \d.
3917
3918 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3919 ending in _t are). So I've renamed it as ESC_tee.
3920
3921 3. The first argument for regexec() in the POSIX wrapper should have been
3922 defined as "const".
3923
3924 4. Changed pcretest to use malloc() for its buffers so that they can be
3925 Electric Fenced for debugging.
3926
3927 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3928 to read one or more bytes before the start of the subject string. Often this
3929 had no effect on PCRE's behaviour, but in some circumstances it could
3930 provoke a segmentation fault.
3931
3932 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3933 to try to read one or more bytes before the start of the subject string.
3934
3935 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3936 UTF-8 support could misbehave in various ways if the subject string
3937 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3938 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3939 back over UTF-8 characters.)
3940
3941
3942 Version 4.2 14-Apr-03
3943 ---------------------
3944
3945 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3946
3947 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3948 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3949 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3950 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3951 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3952 and BUILD_EXEEXT
3953 Note: automatic setting of the BUILD variables is not yet working
3954 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3955 compile-time but not at link-time
3956 [LINK]: use for linking executables only
3957 make different versions for Windows and non-Windows
3958 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3959 libraries
3960 [LINK_FOR_BUILD]: new variable
3961 [OBJEXT]: use throughout
3962 [EXEEXT]: use throughout
3963 <winshared>: new target
3964 <wininstall>: new target
3965 <dftables.o>: use native compiler
3966 <dftables>: use native linker
3967 <install>: handle Windows platform correctly
3968 <clean>: ditto
3969 <check>: ditto
3970 copy DLL to top builddir before testing
3971
3972 As part of these changes, -no-undefined was removed again. This was reported
3973 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3974 in any case.
3975
3976 3. Some tidies to get rid of compiler warnings:
3977
3978 . In the match_data structure, match_limit was an unsigned long int, whereas
3979 match_call_count was an int. I've made them both unsigned long ints.
3980
3981 . In pcretest the fact that a const uschar * doesn't automatically cast to
3982 a void * provoked a warning.
3983
3984 . Turning on some more compiler warnings threw up some "shadow" variables
3985 and a few more missing casts.
3986
3987 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3988 option, a class that contained a single character with a value between 128
3989 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3990
3991 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3992 option, a class that contained several characters, but with at least one
3993 whose value was between 128 and 255 caused PCRE to crash.
3994
3995
3996 Version 4.1 12-Mar-03
3997 ---------------------
3998
3999 1. Compiling with gcc -pedantic found a couple of places where casts were
4000 needed, and a string in dftables.c that was longer than standard compilers are
4001 required to support.
4002
4003 2. Compiling with Sun's compiler found a few more places where the code could
4004 be tidied up in order to avoid warnings.
4005
4006 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4007 first of these names is deprecated in the latest Autoconf in favour of the name
4008 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4009 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4010 analogy I have changed it to CFLAGS_FOR_BUILD.
4011
4012 4. Added -no-undefined to the linking command in the Makefile, because this is
4013 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4014 linking step for the pcreposix library.
4015
4016 5. PCRE was failing to diagnose the case of two named groups with the same
4017 name.
4018
4019 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4020 literal character that is needed in the subject for a match, and scans along to
4021 ensure that it is present before embarking on the full matching process. This
4022 saves time in cases of nested unlimited repeats that are never going to match.
4023 Problem: the scan can take a lot of time if the subject is very long (e.g.
4024 megabytes), thus penalizing straightforward matches. It is now done only if the
4025 amount of subject to be scanned is less than 1000 bytes.
4026
4027 7. A lesser problem with the same optimization is that it was recording the
4028 first character of an anchored pattern as "needed", thus provoking a search
4029 right along the subject, even when the first match of the pattern was going to
4030 fail. The "needed" character is now not set for anchored patterns, unless it
4031 follows something in the pattern that is of non-fixed length. Thus, it still
4032 fulfils its original purpose of finding quick non-matches in cases of nested
4033 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4034
4035
4036 Version 4.0 17-Feb-03
4037 ---------------------
4038
4039 1. If a comment in an extended regex that started immediately after a meta-item
4040 extended to the end of string, PCRE compiled incorrect data. This could lead to
4041 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4042
4043 2. Moved to autoconf 2.53 and libtool 1.4.2.
4044
4045 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4046 the special perltest8 script is no longer needed - all the tests can be run
4047 from a single perltest script.
4048
4049 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4050 by \s. It has now been removed in PCRE. This means it isn't recognized as
4051 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4052 class [:space:] *does* include VT, thereby creating a mess.
4053
4054 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4055 space and tab.
4056
4057 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4058 its new features into the main test script, reducing the number of scripts.
4059
4060 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4061 were backward compatible, and made the (?i) apply to the whole pattern, as if
4062 /i were given. Now it behaves more logically, and applies the option setting
4063 only to what follows. PCRE has been changed to follow suit. However, if it
4064 finds options settings right at the start of the pattern, it extracts them into
4065 the global options, as before. Thus, they show up in the info data.
4066
4067 8. Added support for the \Q...\E escape sequence. Characters in between are
4068 treated as literals. This is slightly different from Perl in that $ and @ are
4069 also handled as literals inside the quotes. In Perl, they will cause variable
4070 interpolation. Note the following examples:
4071
4072 Pattern PCRE matches Perl matches
4073
4074 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4075 \Qabc\$xyz\E abc\$xyz abc\$xyz
4076 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4077
4078 For compatibility with Perl, \Q...\E sequences are recognized inside character
4079 classes as well as outside them.
4080
4081 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4082 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4083 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4084 signed/unsigned warnings.
4085
4086 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4087 option for pcretest, so I've replaced it by a simple function that does just
4088 that job.
4089
4090 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4091 "pcregrep -".
4092
4093 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4094 Java package. This provides some syntactic sugar for simple cases of what my
4095 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4096 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4097 item, you can use this simplified notation. Note that only makes sense with
4098 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4099 greediness, whatever the setting of the PCRE_UNGREEDY option.
4100
4101 13. A change of greediness default within a pattern was not taking effect at
4102 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4103 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4104 was abstracted outside.
4105
4106 14. PCRE now supports the \G assertion. It is true when the current matching
4107 position is at the start point of the match. This differs from \A when the
4108 starting offset is non-zero. Used with the /g option of pcretest (or similar
4109 code), it works in the same way as it does for Perl's /g option. If all
4110 alternatives of a regex begin with \G, the expression is anchored to the start
4111 match position, and the "anchored" flag is set in the compiled expression.
4112
4113 15. Some bugs concerning the handling of certain option changes within patterns
4114 have been fixed. These applied to options other than (?ims). For example,
4115 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4116 been the other way round. Some of this was related to change 7 above.
4117
4118 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4119 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4120 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4121 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4122
4123 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4124 mode. Unlike ".", it always matches newline, whatever the setting of
4125 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4126 assertions. Perl allows it, but it doesn't (in general) work because it can't
4127 calculate the length of the lookbehind. At least, that's the case for Perl
4128 5.8.0 - I've been told they are going to document that it doesn't work in
4129 future.
4130
4131 18. Added an error diagnosis for escapes that PCRE does not support: these are
4132 \L, \l, \N, \P, \p, \U, \u, and \X.
4133
4134 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4135 reading past the end of the pattern in cases such as /[abcd/.
4136
4137 20. PCRE was getting more memory than necessary for patterns with classes that
4138 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4139
4140 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4141 compiling PCRE for use with Virtual Pascal.
4142
4143 22. Small fix to the Makefile to make it work properly if the build is done
4144 outside the source tree.
4145
4146 23. Added a new extension: a condition to go with recursion. If a conditional
4147 subpattern starts with (?(R) the "true" branch is used if recursion has
4148 happened, whereas the "false" branch is used only at the top level.
4149
4150 24. When there was a very long string of literal characters (over 255 bytes
4151 without UTF support, over 250 bytes with UTF support), the computation of how
4152 much memory was required could be incorrect, leading to segfaults or other
4153 strange effects.
4154
4155 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4156 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4157 there was a subsequent back reference to those brackets. This meant that, for
4158 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4159 possible to check for precisely this case. All we can do is abandon the
4160 optimization if .* occurs inside capturing brackets when there are any back
4161 references whatsoever. (See below for a better fix that came later.)
4162
4163 26. The handling of the optimization for finding the first character of a
4164 non-anchored pattern, and for finding a character that is required later in the
4165 match were failing in some cases. This didn't break the matching; it just
4166 failed to optimize when it could. The way this is done has been re-implemented.
4167
4168 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4169
4170 28. Added a new feature that provides some of the functionality that Perl
4171 provides with (?{...}). The facility is termed a "callout". The way it is done
4172 in PCRE is for the caller to provide an optional function, by setting
4173 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4174 global variable. By default it is unset, which disables all calling out. To get
4175 the function called, the regex must include (?C) at appropriate points. This
4176 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4177 This provides a means of identifying different callout points. When PCRE
4178 reaches such a point in the regex, if pcre_callout has been set, the external
4179 function is called. It is provided with data in a structure called
4180 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4181 matching continues; if it returns a non-zero value, the match at the current
4182 point fails. However, backtracking will occur if possible. [This was changed
4183 later and other features added - see item 49 below.]
4184
4185 29. pcretest is upgraded to test the callout functionality. It provides a
4186 callout function that displays information. By default, it shows the start of
4187 the match and the current position in the text. There are some new data escapes
4188 to vary what happens:
4189
4190 \C+ in addition, show current contents of captured substrings
4191 \C- do not supply a callout function
4192 \C!n return 1 when callout number n is reached
4193 \C!n!m return 1 when callout number n is reached for the mth time
4194
4195 30. If pcregrep was called with the -l option and just a single file name, it
4196 output "<stdin>" if a match was found, instead of the file name.
4197
4198 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4199 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4200 pcre_exec(). This saves a malloc/free per call. The default value of
4201 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4202 when configuring.
4203
4204 32. The default maximum size of a compiled pattern is 64K. There have been a
4205 few cases of people hitting this limit. The code now uses macros to handle the
4206 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4207 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4208 configuring. Tests 2 and 5 work only with 2-byte links because they output
4209 debugging information about compiled patterns.
4210
4211 33. Internal code re-arrangements:
4212
4213 (a) Moved the debugging function for printing out a compiled regex into
4214 its own source file (printint.c) and used #include to pull it into
4215 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4216 separate copies.
4217
4218 (b) Defined the list of op-code names for debugging as a macro in
4219 internal.h so that it is next to the definition of the opcodes.
4220
4221 (c) Defined a table of op-code lengths for simpler skipping along compiled
4222 code. This is again a macro in internal.h so that it is next to the
4223 definition of the opcodes.
4224
4225 34. Added support for recursive calls to individual subpatterns, along the
4226 lines of Robin Houston's patch (but implemented somewhat differently).
4227
4228 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4229 allow it to read and process whole directories in Win32. This code was
4230 contributed by Lionel Fourquaux; it has not been tested by me.
4231
4232 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4233 used to name a group. Names consist of alphanumerics and underscores, and must
4234 be unique. Back references use the syntax (?P=name) and recursive calls use
4235 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4236 numbers. The function pcre_fullinfo() can be used after compilation to extract
4237 a name/number map. There are three relevant calls:
4238
4239 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4240 PCRE_INFO_NAMECOUNT yields the number of entries
4241 PCRE_INFO_NAMETABLE yields a pointer to the map.
4242
4243 The map is a vector of fixed-size entries. The size of each entry depends on
4244 the length of the longest name used. The first two bytes of each entry are the
4245 group number, most significant byte first. There follows the corresponding
4246 name, zero terminated. The names are in alphabetical order.
4247
4248 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4249 case instead of 255. Making it the same both with and without UTF-8 support
4250 means that the same test output works with both.
4251
4252 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4253 calling malloc() with a zero argument.
4254
4255 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4256 optimization. I've improved things by keeping a bitmap of backreferences with
4257 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4258 fact referenced, the optimization can be applied. It is unlikely that a
4259 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4260 the match to follow \n) will appear inside brackets with a number greater than
4261 31, but if it does, any back reference > 31 suppresses the optimization.
4262
4263 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4264 of disabling numbered capturing parentheses. Any opening parenthesis that is
4265 not followed by ? behaves as if it were followed by ?: but named parentheses
4266 can still be used for capturing (and they will acquire numbers in the usual
4267 way).
4268
4269 41. Redesigned the return codes from the match() function into yes/no/error so
4270 that errors can be passed back from deep inside the nested calls. A malloc
4271 failure while inside a recursive subpattern call now causes the
4272 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4273
4274 42. It is now possible to set a limit on the number of times the match()
4275 function is called in a call to pcre_exec(). This facility makes it possible to
4276 limit the amount of recursion and backtracking, though not in a directly
4277 obvious way, because the match() function is used in a number of different
4278 circumstances. The count starts from zero for each position in the subject
4279 string (for non-anchored patterns). The default limit is, for compatibility, a
4280 large number, namely 10 000 000. You can change this in two ways:
4281
4282 (a) When configuring PCRE before making, you can use --with-match-limit=n
4283 to set a default value for the compiled library.
4284
4285 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4286 a different value is set. See 45 below.
4287
4288 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4289
4290 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4291 of things that can be changed at compile time. The first argument specifies
4292 what is wanted and the second points to where the information is to be placed.
4293 The current list of available information is:
4294
4295 PCRE_CONFIG_UTF8
4296
4297 The output is an integer that is set to one if UTF-8 support is available;
4298 otherwise it is set to zero.
4299
4300 PCRE_CONFIG_NEWLINE
4301
4302 The output is an integer that it set to the value of the code that is used for
4303 newline. It is either LF (10) or CR (13).
4304
4305 PCRE_CONFIG_LINK_SIZE
4306
4307 The output is an integer that contains the number of bytes used for internal
4308 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4309
4310 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4311
4312 The output is an integer that contains the threshold above which the POSIX
4313 interface uses malloc() for output vectors. See item 31 above.
4314
4315 PCRE_CONFIG_MATCH_LIMIT
4316
4317 The output is an unsigned integer that contains the default limit of the number
4318 of match() calls in a pcre_exec() execution. See 42 above.
4319
4320 44. pcretest has been upgraded by the addition of the -C option. This causes it
4321 to extract all the available output from the new pcre_config() function, and to
4322 output it. The program then exits immediately.
4323
4324 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4325 order to support additional features. One way would have been to define
4326 pcre_exec2() (for example) with extra arguments, but this would not have been
4327 extensible, and would also have required all calls to the original function to
4328 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4329 is used for passing in "extra" data from pcre_study().
4330
4331 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4332 contains the following fields:
4333
4334 flags a bitmap indicating which of the following fields are set
4335 study_data opaque data from pcre_study()
4336 match_limit a way of specifying a limit on match() calls for a specific
4337 call to pcre_exec()
4338 callout_data data for callouts (see 49 below)
4339
4340 The flag bits are also defined in pcre.h, and are
4341
4342 PCRE_EXTRA_STUDY_DATA
4343 PCRE_EXTRA_MATCH_LIMIT
4344 PCRE_EXTRA_CALLOUT_DATA
4345
4346 The pcre_study() function now returns one of these new pcre_extra blocks, with
4347 the actual study data pointed to by the study_data field, and the
4348 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4349 before. That is, this change is entirely upwards-compatible and requires no
4350 change to existing code.
4351
4352 If you want to pass in additional data to pcre_exec(), you can either place it
4353 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4354 block.
4355
4356 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4357 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4358 times with different match limits, until it finds the minimum value needed for
4359 pcre_exec() to complete. The value is then output. This can be instructive; for
4360 most simple matches the number is quite small, but for pathological cases it
4361 gets very large very quickly.
4362
4363 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4364 returns the size of the data block pointed to by the study_data field in a
4365 pcre_extra block, that is, the value that was passed as the argument to
4366 pcre_malloc() when PCRE was getting memory in which to place the information
4367 created by pcre_study(). The fourth argument should point to a size_t variable.
4368 pcretest has been extended so that this information is shown after a successful
4369 pcre_study() call when information about the compiled regex is being displayed.
4370
4371 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4372 because what follows is always an absolute path. (Later: it turns out that this
4373 is more than cosmetic for MinGW, because it doesn't like empty path
4374 components.)
4375
4376 49. Some changes have been made to the callout feature (see 28 above):
4377
4378 (i) A callout function now has three choices for what it returns:
4379
4380 0 => success, carry on matching
4381 > 0 => failure at this point, but backtrack if possible
4382 < 0 => serious error, return this value from pcre_exec()
4383
4384 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4385 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4386 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4387 use by callout functions. It will never be used by PCRE itself.
4388
4389 (ii) The pcre_extra structure (see 45 above) has a void * field called
4390 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4391 pcre_callout_block structure has a field of the same name. The contents of
4392 the field passed in the pcre_extra structure are passed to the callout
4393 function in the corresponding field in the callout block. This makes it
4394 easier to use the same callout-containing regex from multiple threads. For
4395 testing, the pcretest program has a new data escape
4396
4397 \C*n pass the number n (may be negative) as callout_data
4398
4399 If the callout function in pcretest receives a non-zero value as
4400 callout_data, it returns that value.
4401
4402 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4403 there were some redundant $(CFLAGS) in commands that are now specified as
4404 $(LINK), which already includes $(CFLAGS).
4405
4406 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4407 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4408 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4409 one-byte characters throughout. Note that case-insensitive matching applies
4410 only to characters whose values are less than 256. PCRE doesn't support the
4411 notion of cases for higher-valued characters.
4412
4413 (i) A character class whose characters are all within 0-255 is handled as
4414 a bit map, and the map is inverted for negative classes. Previously, a
4415 character > 255 always failed to match such a class; however it should
4416 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4417
4418 (ii) A negated character class with a single character < 255 is coded as
4419 "not this character" (OP_NOT). This wasn't working properly when the test
4420 character was multibyte, either singly or repeated.
4421
4422 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4423 mode, for example: \x{100}{2,3}.
4424
4425 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4426 singly or repeated) now correctly test multibyte characters. However,
4427 PCRE doesn't recognize any characters with values greater than 255 as
4428 digits, spaces, or word characters. Such characters always match \D, \S,
4429 and \W, and never match \d, \s, or \w.
4430
4431 (v) Classes may now contain characters and character ranges with values
4432 greater than 255. For example: [ab\x{100}-\x{400}].
4433
4434 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4435 PCRE in UTF-8 mode.
4436
4437 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4438 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4439 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4440 value.)
4441
4442 53. The single man page has become too large. I have therefore split it up into
4443 a number of separate man pages. These also give rise to individual HTML pages;
4444 these are now put in a separate directory, and there is an index.html page that
4445 lists them all. Some hyperlinking between the pages has been installed.
4446
4447 54. Added convenience functions for handling named capturing parentheses.
4448
4449 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4450 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4451 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4452 are faulted.
4453
4454 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4455 calling configure. These values are used when compiling the dftables.c program
4456 which is run to generate the source of the default character tables. They
4457 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4458 you will need to set these values.
4459
4460 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4461
4462
4463 Version 3.9 02-Jan-02
4464 ---------------------
4465
4466 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4467
4468 2. If --disable-static was given, the building process failed when trying to
4469 build pcretest and pcregrep. (For some reason it was using libtool to compile
4470 them, which is not right, as they aren't part of the library.)
4471
4472
4473 Version 3.8 18-Dec-01
4474 ---------------------
4475
4476 1. The experimental UTF-8 code was completely screwed up. It was packing the
4477 bytes in the wrong order. How dumb can you get?
4478
4479
4480 Version 3.7 29-Oct-01
4481 ---------------------
4482
4483 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4484 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4485 this didn't happen under Solaris 8, where I normally test things.
4486
4487 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4488 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4489 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4490 this fix an hour or so after the initial 3.7 release.)
4491
4492
4493 Version 3.6 23-Oct-01
4494 ---------------------
4495
4496 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4497 offsets passed as NULL with zero offset count.
4498
4499 2. The config.guess and config.sub files had not been updated when I moved to
4500 the latest autoconf.
4501
4502
4503 Version 3.5 15-Aug-01
4504 ---------------------
4505
4506 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4507 had been forgotten.
4508
4509 2. By using declared but undefined structures, we can avoid using "void"
4510 definitions in pcre.h while keeping the internal definitions of the structures
4511 private.
4512
4513 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4514 user point of view, this means that both static and shared libraries are built
4515 by default, but this can be individually controlled. More of the work of
4516 handling this static/shared cases is now inside libtool instead of PCRE's make
4517 file.
4518
4519 4. The pcretest utility is now installed along with pcregrep because it is
4520 useful for users (to test regexs) and by doing this, it automatically gets
4521 relinked by libtool. The documentation has been turned into a man page, so
4522 there are now .1, .txt, and .html versions in /doc.
4523
4524 5. Upgrades to pcregrep:
4525 (i) Added long-form option names like gnu grep.
4526 (ii) Added --help to list all options with an explanatory phrase.
4527 (iii) Added -r, --recursive to recurse into sub-directories.
4528 (iv) Added -f, --file to read patterns from a file.
4529
4530 6. pcre_exec() was referring to its "code" argument before testing that
4531 argument for NULL (and giving an error if it was NULL).
4532
4533 7. Upgraded Makefile.in to allow for compiling in a different directory from
4534 the source directory.
4535
4536 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4537 options bits, the pointer it was passed was to an int instead of to an unsigned
4538 long int. This mattered only on 64-bit systems.
4539
4540 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4541 generated) instead of pcre.in, which it its source. Also made the same change
4542 in several of the .c files.
4543
4544 10. A new release of gcc defines printf() as a macro, which broke pcretest
4545 because it had an ifdef in the middle of a string argument for printf(). Fixed
4546 by using separate calls to printf().
4547
4548 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4549 script, to force use of CR or LF instead of \n in the source. On non-Unix
4550 systems, the value can be set in config.h.
4551
4552 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4553 absolute limit. Changed the text of the error message to make this clear, and
4554 likewise updated the man page.
4555
4556 13. The limit of 99 on the number of capturing subpatterns has been removed.
4557 The new limit is 65535, which I hope will not be a "real" limit.
4558
4559
4560 Version 3.4 22-Aug-00
4561 ---------------------
4562
4563 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4564
4565 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4566
4567
4568 Version 3.3 01-Aug-00
4569 ---------------------
4570
4571 1. If an octal character was given, but the value was greater than \377, it
4572 was not getting masked to the least significant bits, as documented. This could
4573 lead to crashes in some systems.
4574
4575 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4576 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4577
4578 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4579 These just pass their arguments on to (pcre_free)(), but they are provided
4580 because some uses of PCRE bind it to non-C systems that can call its functions,
4581 but cannot call free() or pcre_free() directly.
4582
4583 4. Add "make test" as a synonym for "make check". Corrected some comments in
4584 the Makefile.
4585
4586 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4587 Makefile.
4588
4589 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4590 command called pgrep for grepping around the active processes.
4591
4592 7. Added the beginnings of support for UTF-8 character strings.
4593
4594 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4595 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4596 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4597 out for the ar command.)
4598
4599
4600 Version 3.2 12-May-00
4601 ---------------------
4602
4603 This is purely a bug fixing release.
4604
4605 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4606 of ZA. This was just one example of several cases that could provoke this bug,
4607 which was introduced by change 9 of version 2.00. The code for breaking
4608 infinite loops after an iteration that matches an empty string was't working
4609 correctly.
4610
4611 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4612 when matched against abbab (for example). After matching an empty string, it
4613 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4614 caused it to match further down the string than it should.
4615
4616 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4617 was there because it doesn't seem to be needed, and it causes trouble on some
4618 systems, as it is not a Standard C header. It has been removed.
4619
4620 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4621 were reported on the Macintosh. The changes were from
4622
4623 while ((c = *(++ptr)) != 0 && c != '\n');
4624 to
4625 while ((c = *(++ptr)) != 0 && c != '\n') ;
4626
4627 Totally extraordinary, but if that's what it takes...
4628
4629 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4630 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4631 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4632 assumes the way PCRE uses memmove() (always moving upwards).
4633
4634 6. PCRE is being used in one environment where strchr() is not available. There
4635 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4636 faster code anyway.
4637
4638
4639 Version 3.1 09-Feb-00
4640 ---------------------
4641
4642 The only change in this release is the fixing of some bugs in Makefile.in for
4643 the "install" target:
4644
4645 (1) It was failing to install pcreposix.h.
4646
4647 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4648
4649
4650 Version 3.0 01-Feb-00
4651 ---------------------
4652
4653 1. Add support for the /+ modifier to perltest (to output $` like it does in
4654 pcretest).
4655
4656 2. Add support for the /g modifier to perltest.
4657
4658 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4659 matches null strings.
4660
4661 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4662 pattern. Perl treats empty patterns specially - it reuses the most recent
4663 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4664 effect.
4665
4666 5. The POSIX interface was broken in that it was just handing over the POSIX
4667 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4668 required a bigger vector, with some working space on the end. This means that
4669 the POSIX wrapper now has to get and free some memory, and copy the results.
4670
4671 6. Added some simple autoconf support, placing the test data and the
4672 documentation in separate directories, re-organizing some of the
4673 information files, and making it build pcre-config (a GNU standard). Also added
4674 libtool support for building PCRE as a shared library, which is now the
4675 default.
4676
4677 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4678 09 are not valid octal constants. Single digits will be used for minor values
4679 less than 10.
4680
4681 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4682 existing programs that set these in the POSIX interface can use PCRE without
4683 modification.
4684
4685 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4686 return all that pcre_info() returns, plus additional data. The pcre_info()
4687 function is retained for compatibility, but is considered to be obsolete.
4688
4689 10. Added experimental recursion feature (?R) to handle one common case that
4690 Perl 5.6 will be able to do with (?p{...}).
4691
4692 11. Added support for POSIX character classes like [:alpha:], which Perl is
4693 adopting.
4694
4695
4696 Version 2.08 31-Aug-99
4697 ----------------------
4698
4699 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4700 trying to match at the startoffset position, but instead was moving forward to
4701 the next newline as if a previous match had failed.
4702
4703 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4704 and could get into a loop if a null string was matched other than at the start
4705 of the subject.
4706
4707 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4708 be distinguished at compile time, and for completeness also added PCRE_DATE.
4709
4710 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4711 in GnuWin32 environments.
4712
4713
4714 Version 2.07 29-Jul-99
4715 ----------------------
4716
4717 1. The documentation is now supplied in plain text form and HTML as well as in
4718 the form of man page sources.
4719
4720 2. C++ compilers don't like assigning (void *) values to other pointer types.
4721 In particular this affects malloc(). Although there is no problem in Standard
4722 C, I've put in casts to keep C++ compilers happy.
4723
4724 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4725 should be (const char *).
4726
4727 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4728 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4729 However, I haven't made this a standard facility. The documentation doesn't
4730 mention it, and the Makefile doesn't support it.
4731
4732 5. The Makefile now contains an "install" target, with editable destinations at
4733 the top of the file. The pcretest program is not installed.
4734
4735 6. pgrep -V now gives the PCRE version number and date.
4736
4737 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4738 causing the entire string to be ignored, instead of just the last character.
4739
4740 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4741 non-matching string, it can take a very, very long time, even for strings of
4742 quite modest length, because of the nested recursion. PCRE now does better in
4743 some of these cases. It does this by remembering the last required literal
4744 character in the pattern, and pre-searching the subject to ensure it is present
4745 before running the real match. In other words, it applies a heuristic to detect
4746 some types of certain failure quickly, and in the above example, if presented
4747 with a string that has no trailing " it gives "no match" very quickly.
4748
4749 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4750 other alternatives are tried instead.
4751
4752
4753 Version 2.06 09-Jun-99
4754 ----------------------
4755
4756 1. Change pcretest's output for amount of store used to show just the code
4757 space, because the remainder (the data block) varies in size between 32-bit and
4758 64-bit systems.
4759
4760 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4761 start matching at. This allows lookbehinds to work when searching for multiple
4762 occurrences in a string.
4763
4764 3. Added additional options to pcretest for testing multiple occurrences:
4765
4766 /+ outputs the rest of the string that follows a match
4767 /g loops for multiple occurrences, using the new startoffset argument
4768 /G loops for multiple occurrences by passing an incremented pointer
4769
4770 4. PCRE wasn't doing the "first character" optimization for patterns starting
4771 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4772 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4773 the letter 'x'. On long subject strings, this gives a significant speed-up.
4774
4775
4776 Version 2.05 21-Apr-99
4777 ----------------------
4778
4779 1. Changed the type of magic_number from int to long int so that it works
4780 properly on 16-bit systems.
4781
4782 2. Fixed a bug which caused patterns starting with .* not to work correctly
4783 when the subject string contained newline characters. PCRE was assuming
4784 anchoring for such patterns in all cases, which is not correct because .* will
4785 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4786 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4787 must be retried after every newline in the subject.
4788
4789
4790 Version 2.04 18-Feb-99
4791 ----------------------
4792
4793 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4794 computation of the store needed to hold the pattern was incorrect (too large).
4795 If such patterns were nested a few deep, this could multiply and become a real
4796 problem.
4797
4798 2. Added /M option to pcretest to show the memory requirement of a specific
4799 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4800
4801 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4802 compiled in such a way that the backtracking after subsequent failure was
4803 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4804 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4805
4806
4807 Version 2.03 02-Feb-99
4808 ----------------------
4809
4810 1. Fixed typo and small mistake in man page.
4811
4812 2. Added 4th condition (GPL supersedes if conflict) and created separate
4813 LICENCE file containing the conditions.
4814
4815 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4816 Perl, that is the internal \ allows the delimiter to be included in the
4817 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4818 the final delimiter, add \ to the end of the pattern (to test the error).
4819
4820 4. Added the convenience functions for extracting substrings after a successful
4821 match. Updated pcretest to make it able to test these functions.
4822
4823
4824 Version 2.02 14-Jan-99
4825 ----------------------
4826
4827 1. Initialized the working variables associated with each extraction so that
4828 their saving and restoring doesn't refer to uninitialized store.
4829
4830 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4831 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4832 fix the problem.
4833
4834 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4835 calls, and wasn't printing the correct value for compiling calls. Increased the
4836 default value of LOOPREPEAT, and the number of significant figures in the
4837 times.
4838
4839 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4840
4841 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4842 a building problem on Windows NT with a FAT file system.
4843
4844
4845 Version 2.01 21-Oct-98
4846 ----------------------
4847
4848 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4849 to character tables built by pcre_maketables() in the current locale. If NULL
4850 is passed, the default tables are used.
4851
4852
4853 Version 2.00 24-Sep-98
4854 ----------------------
4855
4856 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4857 it any more.
4858
4859 2. Allow quantification of (?>) groups, and make it work correctly.
4860
4861 3. The first character computation wasn't working for (?>) groups.
4862
4863 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4864 end of the subject) and add 5.005's \z, which really does match only at the
4865 very end of the subject.
4866
4867 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4868
4869 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4870 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4871 localized options. All options to pcre_study() were also removed.
4872
4873 7. Add other new features from 5.005:
4874
4875 $(?<= positive lookbehind
4876 $(?<! negative lookbehind
4877 (?imsx-imsx) added the unsetting capability
4878 such a setting is global if at outer level; local otherwise
4879 (?imsx-imsx:) non-capturing groups with option setting
4880 (?(cond)re|re) conditional pattern matching
4881
4882 A backreference to itself in a repeated group matches the previous
4883 captured string.
4884
4885 8. General tidying up of studying (both automatic and via "study")
4886 consequential on the addition of new assertions.
4887
4888 9. As in 5.005, unlimited repeated groups that could match an empty substring
4889 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4890 runtime if any iteration does actually match an empty substring.
4891
4892 10. Include the RunTest script in the distribution.
4893
4894 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4895 discrepancies, some of which were old and were also with respect to 5.004. They
4896 have now been fixed.
4897
4898
4899 Version 1.09 28-Apr-98
4900 ----------------------
4901
4902 1. A negated single character class followed by a quantifier with a minimum
4903 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4904 program crashes, or just wrong answers. This did not apply to negated classes
4905 containing more than one character, or to minima other than one.
4906
4907
4908 Version 1.08 27-Mar-98
4909 ----------------------
4910
4911 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4912
4913 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4914 latter must appear before anything that relies on it in the pattern.
4915
4916
4917 Version 1.07 16-Feb-98
4918 ----------------------
4919
4920 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4921 repeat of a potentially empty string).
4922
4923
4924 Version 1.06 23-Jan-98
4925 ----------------------
4926
4927 1. Added Markus Oberhumer's little patches for C++.
4928
4929 2. Literal strings longer than 255 characters were broken.
4930
4931
4932 Version 1.05 23-Dec-97
4933 ----------------------
4934
4935 1. Negated character classes containing more than one character were failing if
4936 PCRE_CASELESS was set at run time.
4937
4938
4939 Version 1.04 19-Dec-97
4940 ----------------------
4941
4942 1. Corrected the man page, where some "const" qualifiers had been omitted.
4943
4944 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4945 input syntax.
4946
4947 3. Fixed memory leak which occurred when a regex with back references was
4948 matched with an offsets vector that wasn't big enough. The temporary memory
4949 that is used in this case wasn't being freed if the match failed.
4950
4951 4. Tidied pcretest to ensure it frees memory that it gets.
4952
4953 5. Temporary memory was being obtained in the case where the passed offsets
4954 vector was exactly big enough.
4955
4956 6. Corrected definition of offsetof() from change 5 below.
4957
4958 7. I had screwed up change 6 below and broken the rules for the use of
4959 setjmp(). Now fixed.
4960
4961
4962 Version 1.03 18-Dec-97
4963 ----------------------
4964
4965 1. A erroneous regex with a missing opening parenthesis was correctly
4966 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4967 on some systems.
4968
4969 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4970 it was reported that one broken compiler failed on the former because "code" is
4971 also an independent variable.
4972
4973 3. The erroneous regex a[]b caused an array overrun reference.
4974
4975 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4976 fail on data ending with that character. (It was going on too far, and checking
4977 the next character, typically a binary zero.) This was specific to the
4978 optimized code for single-character negative classes.
4979
4980 5. Added a contributed patch from the TIN world which does the following:
4981
4982 + Add an undef for memmove, in case the the system defines a macro for it.
4983
4984 + Add a definition of offsetof(), in case there isn't one. (I don't know
4985 the reason behind this - offsetof() is part of the ANSI standard - but
4986 it does no harm).
4987
4988 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4989 most of the places where whitespace preceded '#'. I have given up and
4990 allowed the remaining 2 cases to be at the margin.
4991
4992 + Rename some variables in pcre to eliminate shadowing. This seems very
4993 pedantic, but does no harm, of course.
4994
4995 6. Moved the call to setjmp() into its own function, to get rid of warnings
4996 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4997
4998 7. Constructs such as \d{8,} were compiling into the equivalent of
4999 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5000 outcome, but in this particular case used more store than had been allocated,
5001 which caused the bug to be discovered because it threw up an internal error.
5002
5003 8. The debugging code in both pcre and pcretest for outputting the compiled
5004 form of a regex was going wrong in the case of back references followed by
5005 curly-bracketed repeats.
5006
5007
5008 Version 1.02 12-Dec-97
5009 ----------------------
5010
5011 1. Typos in pcre.3 and comments in the source fixed.
5012
5013 2. Applied a contributed patch to get rid of places where it used to remove
5014 'const' from variables, and fixed some signed/unsigned and uninitialized
5015 variable warnings.
5016
5017 3. Added the "runtest" target to Makefile.
5018
5019 4. Set default compiler flag to -O2 rather than just -O.
5020
5021
5022 Version 1.01 19-Nov-97
5023 ----------------------
5024
5025 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5026 like /([ab]*)*/, that is, for classes with more than one character in them.
5027
5028 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5029 as /((?>a*))*/ (a PCRE_EXTRA facility).
5030
5031
5032 Version 1.00 18-Nov-97
5033 ----------------------
5034
5035 1. Added compile-time macros to support systems such as SunOS4 which don't have
5036 memmove() or strerror() but have other things that can be used instead.
5037
5038 2. Arranged that "make clean" removes the executables.
5039
5040
5041 Version 0.99 27-Oct-97
5042 ----------------------
5043
5044 1. Fixed bug in code for optimizing classes with only one character. It was
5045 initializing a 32-byte map regardless, which could cause it to run off the end
5046 of the memory it had got.
5047
5048 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5049
5050
5051 Version 0.98 22-Oct-97
5052 ----------------------
5053
5054 1. Fixed bug in code for handling temporary memory usage when there are more
5055 back references than supplied space in the ovector. This could cause segfaults.
5056
5057
5058 Version 0.97 21-Oct-97
5059 ----------------------
5060
5061 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5062
5063 2. Optimized negated single characters not to use a bit map.
5064
5065 3. Brought error texts together as macro definitions; clarified some of them;
5066 fixed one that was wrong - it said "range out of order" when it meant "invalid
5067 escape sequence".
5068
5069 4. Changed some char * arguments to const char *.
5070
5071 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5072
5073 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5074 pcretest.
5075
5076
5077 Version 0.96 16-Oct-97
5078 ----------------------
5079
5080 1. Added a simple "pgrep" utility to the distribution.
5081
5082 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5083 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5084 where "ddd" means "one or more decimal digits".
5085
5086 3. Fixed serious bug. If a pattern had a back reference, but the call to
5087 pcre_exec() didn't supply a large enough ovector to record the related
5088 identifying subpattern, the match always failed. PCRE now remembers the number
5089 of the largest back reference, and gets some temporary memory in which to save
5090 the offsets during matching if necessary, in order to ensure that
5091 backreferences always work.
5092
5093 4. Increased the compatibility with Perl in a number of ways:
5094
5095 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5096 to request this handling. The option can be set at compile or exec time.
5097
5098 (b) $ matches before a terminating newline by default; an option
5099 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5100 mode). The option can be set at compile or exec time.
5101
5102 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5103 the same as Perl's. If the decimal number it represents is less than 10
5104 or there aren't that many previous left capturing parentheses, an octal
5105 escape is read. Inside a character class, it's always an octal escape,
5106 even if it is a single digit.
5107
5108 (d) An escaped but undefined alphabetic character is taken as a literal,
5109 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5110 escapes.
5111
5112 (e) {0} is now permitted. (The previous item is removed from the compiled
5113 pattern).
5114
5115 5. Changed all the names of code files so that the basic parts are no longer
5116 than 10 characters, and abolished the teeny "globals.c" file.
5117
5118 6. Changed the handling of character classes; they are now done with a 32-byte
5119 bit map always.
5120
5121 7. Added the -d and /D options to pcretest to make it possible to look at the
5122 internals of compilation without having to recompile pcre.
5123
5124
5125 Version 0.95 23-Sep-97
5126 ----------------------
5127
5128 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5129 \x20 at the start of a run of normal characters. These were being treated as
5130 real characters, instead of the source characters being re-checked.
5131
5132
5133 Version 0.94 18-Sep-97
5134 ----------------------
5135
5136 1. The functions are now thread-safe, with the caveat that the global variables
5137 containing pointers to malloc() and free() or alternative functions are the
5138 same for all threads.
5139
5140 2. Get pcre_study() to generate a bitmap of initial characters for non-
5141 anchored patterns when this is possible, and use it if passed to pcre_exec().
5142
5143
5144 Version 0.93 15-Sep-97
5145 ----------------------
5146
5147 1. /(b)|(:+)/ was computing an incorrect first character.
5148
5149 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5150 but not actually doing anything yet.
5151
5152 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5153 as Perl does (e.g. [-az] or [az-]).
5154
5155 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5156 all possible positions.
5157
5158 5. Split up into different modules to avoid including unneeded functions in a
5159 compiled binary. However, compile and exec are still in one module. The "study"
5160 function is split off.
5161
5162 6. The character tables are now in a separate module whose source is generated
5163 by an auxiliary program - but can then be edited by hand if required. There are
5164 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5165 toupper() in the code.
5166
5167 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5168 make them global. Abolish the function for setting them, as the caller can now
5169 set them directly.
5170
5171
5172 Version 0.92 11-Sep-97
5173 ----------------------
5174
5175 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5176 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5177
5178 2. Caseless matching was not working in character classes if the characters in
5179 the pattern were in upper case.
5180
5181 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5182
5183 4. Make PCRE_ANCHORED public and accept as a compile option.
5184
5185 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5186 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5187 pass them.
5188
5189 6. Give an error if bad option bits passed at compile or run time.
5190
5191 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5192 pcretest to cause it to pass that flag.
5193
5194 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5195 options, and the first character, if set.
5196
5197 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5198
5199
5200 Version 0.91 10-Sep-97
5201 ----------------------
5202
5203 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5204 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5205
5206 2. PCRE was looping on encountering an indefinitely repeated back reference to
5207 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5208 Perl does - treats the match as successful.
5209
5210 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5