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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1357 - (show annotations)
Tue Aug 27 15:49:16 2013 UTC (6 years, 2 months ago) by ph10
File size: 250824 byte(s)
Error occurred while calculating annotation data.
Add -T and -TM to pcretest.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.34 xx-xxxx-201x
5 --------------------------
6
7 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8 executable memory. Patch inspired by Carsten Klein.
9
10 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11 this macro is never tested and has no effect, because the work to support
12 coverage involves only compiling and linking options and special targets in
13 the Makefile. The comment in config.h implied that defining the macro would
14 enable coverage support, which is totally false. There was also support for
15 setting this macro in the CMake files (my fault, I just copied it from
16 configure). SUPPORT_GCOV has now been removed.
17
18 3. Make a small performance improvement in strlen16() and strlen32() in
19 pcretest.
20
21 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23 these cases). There was also one in pcretest.c.
24
25 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
26
27 6. In UTF mode, the code for checking whether a group could match an empty
28 string (which is used for indefinitely repeated groups to allow for
29 breaking an infinite loop) was broken when the group contained a repeated
30 negated single-character class with a character that occupied more than one
31 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
32 UTF-8 mode). The effect was undefined: the group might or might not be
33 deemed as matching an empty string, or the program might have crashed.
34
35 7. The code for checking whether a group could match an empty string was not
36 recognizing that \h, \H, \v, \V, and \R must match a character.
37
38 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
39 an empty string. If it can, pcretest shows this in its information output.
40
41 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
42 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
43 matched by pcre_exec() without using JIT:
44
45 (a) If the rest of the pattern did not match after a maximal run of
46 grapheme clusters, the code for backing up to try with fewer of them
47 did not always back up over a full grapheme when characters that do not
48 have the modifier quality were involved, e.g. Hangul syllables.
49
50 (b) If the match point in a subject started with modifier character, and
51 there was no match, the code could incorrectly back up beyond the match
52 point, and potentially beyond the first character in the subject,
53 leading to a segfault or an incorrect match result.
54
55 10. A conditional group with an assertion condition could lead to PCRE
56 recording an incorrect first data item for a match if no other first data
57 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58 first data item, and therefore matched "ca" after "c" instead of at the
59 start.
60
61 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
62 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
63
64 12. The source of pcregrep now includes z/OS-specific code so that it can be
65 compiled for z/OS as part of the special z/OS distribution.
66
67 13. Added the -T and -TM options to pcretest.
68
69
70 Version 8.33 28-May-2013
71 ------------------------
72
73 1. Added 'U' to some constants that are compared to unsigned integers, to
74 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
75 variables that are added to signed variables, to ensure the result is
76 signed and can be negated.
77
78 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
79 CMake config files.
80
81 3. Revise the creation of config.h.generic so that all boolean macros are
82 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
83 overriding via -D on the command line possible.
84
85 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
86 to unsigned int is reported to make a quite noticeable speed difference in
87 a specific Windows environment. Testing on Linux did also appear to show
88 some benefit (and it is clearly not harmful). Also fixed the definition of
89 Xop which should be unsigned.
90
91 5. Related to (4), changing the definition of the intermediate variable cc
92 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
93 improvements.
94
95 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
96 unnecessary spaces.
97
98 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
99 and later.
100
101 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
102
103 9. Optimizing fast_forward_start_bits in JIT.
104
105 10. Adding support for callouts in JIT, and fixing some issues revealed
106 during this work. Namely:
107
108 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
109
110 (b) Minimum length was not checked before the matching is started.
111
112 11. The value of capture_last that is passed to callouts was incorrect in some
113 cases when there was a capture on one path that was subsequently abandoned
114 after a backtrack. Also, the capture_last value is now reset after a
115 recursion, since all captures are also reset in this case.
116
117 12. The interpreter no longer returns the "too many substrings" error in the
118 case when an overflowing capture is in a branch that is subsequently
119 abandoned after a backtrack.
120
121 13. In the pathological case when an offset vector of size 2 is used, pcretest
122 now prints out the matched string after a yield of 0 or 1.
123
124 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
125 Only supported by the JIT compiler at the moment.
126
127 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
128
129 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
130 offset of the starting point of the matching process, provided the offsets
131 vector is large enough.
132
133 17. The \A escape now records a lookbehind value of 1, though its execution
134 does not actually inspect the previous character. This is to ensure that,
135 in partial multi-segment matching, at least one character from the old
136 segment is retained when a new segment is processed. Otherwise, if there
137 are no lookbehinds in the pattern, \A might match incorrectly at the start
138 of a new segment.
139
140 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
141
142 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
143 gives some modest performance improvement in 8-bit mode.
144
145 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
146 be expressed in certain programming languages using Universal Character
147 Names.
148
149 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
150 which points out that "non characters" are not "characters that may not
151 appear in Unicode strings" but rather "characters that are reserved for
152 internal use and have only local meaning".
153
154 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
155 there was a conditional group that depended on an assertion, if the
156 assertion was false, the callout that immediately followed the alternation
157 in the condition was skipped when pcre_exec() was used for matching.
158
159 23. Allow an explicit callout to be inserted before an assertion that is the
160 condition for a conditional group, for compatibility with automatic
161 callouts, which always insert a callout at this point.
162
163 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
164 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
165
166 25. (*PRUNE) is now supported by the JIT compiler.
167
168 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
169
170 27. Fix the case where there are two or more SKIPs with arguments that may be
171 ignored.
172
173 28. (*SKIP) is now supported by the JIT compiler.
174
175 29. (*THEN) is now supported by the JIT compiler.
176
177 30. Update RunTest with additional test selector options.
178
179 31. The way PCRE handles backtracking verbs has been changed in two ways.
180
181 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
182 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
183 backtracking. In some cases this makes it more Perl-compatible, but Perl's
184 rather obscure rules do not always do the same thing.
185
186 (2) Previously, backtracking verbs were confined within assertions. This is
187 no longer the case for positive assertions, except for (*ACCEPT). Again,
188 this sometimes improves Perl compatibility, and sometimes does not.
189
190 32. A number of tests that were in test 2 because Perl did things differently
191 have been moved to test 1, because either Perl or PCRE has changed, and
192 these tests are now compatible.
193
194 32. Backtracking control verbs are now handled in the same way in JIT and
195 interpreter.
196
197 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
198 contained a forward subroutine reference caused a compile error.
199
200 34. Auto-detect and optimize limited repetitions in JIT.
201
202 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
203 blocking (*UTF) etc.
204
205 36. In the interpreter, maximizing pattern repetitions for characters and
206 character types now use tail recursion, which reduces stack usage.
207
208 37. The value of the max lookbehind was not correctly preserved if a compiled
209 and saved regex was reloaded on a host of different endianness.
210
211 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
212 of the compiled pattern block, expand the flags field from 16 to 32 bits
213 because it was almost full.
214
215 39. Try madvise first before posix_madvise.
216
217 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
218 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
219 original change locked out all matches of empty strings. This has been
220 changed so that one match of an empty string per line is recognized.
221 Subsequent searches on the same line (for colouring or for --only-matching,
222 for example) do not recognize empty strings.
223
224 41. Applied a user patch to fix a number of spelling mistakes in comments.
225
226 42. Data lines longer than 65536 caused pcretest to crash.
227
228 43. Clarified the data type for length and startoffset arguments for pcre_exec
229 and pcre_dfa_exec in the function-specific man pages, where they were
230 explicitly stated to be in bytes, never having been updated. I also added
231 some clarification to the pcreapi man page.
232
233 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
234 a segmentation fault.
235
236
237 Version 8.32 30-November-2012
238 -----------------------------
239
240 1. Improved JIT compiler optimizations for first character search and single
241 character iterators.
242
243 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
244 Patch by Daniel Richard G.
245
246 3. Single character iterator optimizations in the JIT compiler.
247
248 4. Improved JIT compiler optimizations for character ranges.
249
250 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
251 Reported by Giuseppe D'Angelo.
252
253 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
254 of a line, was being set incorrectly in cases where .* appeared inside
255 atomic brackets at the start of a pattern, or where there was a subsequent
256 *PRUNE or *SKIP.
257
258 7. Improved instruction cache flush for POWER/PowerPC.
259 Patch by Daniel Richard G.
260
261 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
262 grep:
263
264 (a) There is now no limit to the number of patterns to be matched.
265
266 (b) An error is given if a pattern is too long.
267
268 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
269 are now supported.
270
271 (d) --exclude-from and --include-from (multiple use) have been added.
272
273 (e) Exclusions and inclusions now apply to all files and directories, not
274 just to those obtained from scanning a directory recursively.
275
276 (f) Multiple uses of -f and --file-list are now supported.
277
278 (g) In a Windows environment, the default for -d has been changed from
279 "read" (the GNU grep default) to "skip", because otherwise the presence
280 of a directory in the file list provokes an error.
281
282 (h) The documentation has been revised and clarified in places.
283
284 9. Improve the matching speed of capturing brackets.
285
286 10. Changed the meaning of \X so that it now matches a Unicode extended
287 grapheme cluster.
288
289 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
290 out POSIX threads when JIT support is configured.
291
292 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
293
294 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
295 structure could go wrong in environments where size_t is not the same size
296 as int.
297
298 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
299 set.
300
301 15. The EBCDIC support had decayed; later updates to the code had included
302 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
303 general tidy up of EBCDIC-related issues, and the documentation was also
304 not quite right. There is now a test that can be run on ASCII systems to
305 check some of the EBCDIC-related things (but is it not a full test).
306
307 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
308 in a small tidy to the code.
309
310 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
311
312 18. If the --only-matching (-o) option in pcregrep is specified multiple
313 times, each one causes appropriate output. For example, -o1 -o2 outputs the
314 substrings matched by the 1st and 2nd capturing parentheses. A separating
315 string can be specified by --om-separator (default empty).
316
317 19. Improving the first n character searches.
318
319 20. Turn case lists for horizontal and vertical white space into macros so that
320 they are defined only once.
321
322 21. This set of changes together give more compatible Unicode case-folding
323 behaviour for characters that have more than one other case when UCP
324 support is available.
325
326 (a) The Unicode property table now has offsets into a new table of sets of
327 three or more characters that are case-equivalent. The MultiStage2.py
328 script that generates these tables (the pcre_ucd.c file) now scans
329 CaseFolding.txt instead of UnicodeData.txt for character case
330 information.
331
332 (b) The code for adding characters or ranges of characters to a character
333 class has been abstracted into a generalized function that also handles
334 case-independence. In UTF-mode with UCP support, this uses the new data
335 to handle characters with more than one other case.
336
337 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
338 whose other case is greater than 256 are now correctly matched
339 caselessly. Previously, the high codepoint matched the low one, but not
340 vice versa.
341
342 (d) The processing of \h, \H, \v, and \ in character classes now makes use
343 of the new class addition function, using character lists defined as
344 macros alongside the case definitions of 20 above.
345
346 (e) Caseless back references now work with characters that have more than
347 one other case.
348
349 (f) General caseless matching of characters with more than one other case
350 is supported.
351
352 22. Unicode character properties were updated from Unicode 6.2.0
353
354 23. Improved CMake support under Windows. Patch by Daniel Richard G.
355
356 24. Add support for 32-bit character strings, and UTF-32
357
358 25. Major JIT compiler update (code refactoring and bugfixing).
359 Experimental Sparc 32 support is added.
360
361 26. Applied a modified version of Daniel Richard G's patch to create
362 pcre.h.generic and config.h.generic by "make" instead of in the
363 PrepareRelease script.
364
365 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
366 pcre_compile.c when checking for a zero character.
367
368 28. Introducing a native interface for JIT. Through this interface, the compiled
369 machine code can be directly executed. The purpose of this interface is to
370 provide fast pattern matching, so several sanity checks are not performed.
371 However, feature tests are still performed. The new interface provides
372 1.4x speedup compared to the old one.
373
374 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
375 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
376 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
377
378 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
379 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
380 the "old" RFC 2279). Instead, it ended up passing a negative length to
381 pcre_exec().
382
383 31. Add support for GCC's visibility feature to hide internal functions.
384
385 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
386 "unknown -C option" after outputting 0 or 1.
387
388 33. There is now support for generating a code coverage report for the test
389 suite in environments where gcc is the compiler and lcov is installed. This
390 is mainly for the benefit of the developers.
391
392 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
393 unaddressable using valgrind annotations, allowing valgrind to detect
394 invalid memory accesses. This is mainly for the benefit of the developers.
395
396 25. (*UTF) can now be used to start a pattern in any of the three libraries.
397
398 26. Give configure error if --enable-cpp but no C++ compiler found.
399
400
401 Version 8.31 06-July-2012
402 -------------------------
403
404 1. Fixing a wrong JIT test case and some compiler warnings.
405
406 2. Removed a bashism from the RunTest script.
407
408 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
409 to unsigned type, result still unsigned" that was given by an MS compiler
410 on encountering the code "-sizeof(xxx)".
411
412 4. Partial matching support is added to the JIT compiler.
413
414 5. Fixed several bugs concerned with partial matching of items that consist
415 of more than one character:
416
417 (a) /^(..)\1/ did not partially match "aba" because checking references was
418 done on an "all or nothing" basis. This also applied to repeated
419 references.
420
421 (b) \R did not give a hard partial match if \r was found at the end of the
422 subject.
423
424 (c) \X did not give a hard partial match after matching one or more
425 characters at the end of the subject.
426
427 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
428 a partial match for the string "\r".
429
430 (e) When newline was set to CRLF, the metacharacter "." did not recognize
431 a partial match for a CR character at the end of the subject string.
432
433 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
434 running pcretest, the text "(JIT)" added to the output whenever JIT is
435 actually used to run the match.
436
437 7. Individual JIT compile options can be set in pcretest by following -s+[+]
438 or /S+[+] with a digit between 1 and 7.
439
440 8. OP_NOT now supports any UTF character not just single-byte ones.
441
442 9. (*MARK) control verb is now supported by the JIT compiler.
443
444 10. The command "./RunTest list" lists the available tests without actually
445 running any of them. (Because I keep forgetting what they all are.)
446
447 11. Add PCRE_INFO_MAXLOOKBEHIND.
448
449 12. Applied a (slightly modified) user-supplied patch that improves performance
450 when the heap is used for recursion (compiled with --disable-stack-for-
451 recursion). Instead of malloc and free for each heap frame each time a
452 logical recursion happens, frames are retained on a chain and re-used where
453 possible. This sometimes gives as much as 30% improvement.
454
455 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
456 call.
457
458 14. As documented, (*COMMIT) is now confined to within a positive assertion.
459
460 15. It is now possible to link pcretest with libedit as an alternative to
461 libreadline.
462
463 16. (*COMMIT) control verb is now supported by the JIT compiler.
464
465 17. The Unicode data tables have been updated to Unicode 6.1.0.
466
467 18. Added --file-list option to pcregrep.
468
469 19. Added binary file support to pcregrep, including the -a, --binary-files,
470 -I, and --text options.
471
472 20. The madvise function is renamed for posix_madvise for QNX compatibility
473 reasons. Fixed by Giuseppe D'Angelo.
474
475 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
476
477 22. Changed the option for creating symbolic links for 16-bit man pages from
478 -s to -sf so that re-installing does not cause issues.
479
480 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
481
482 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
483 matches in certain environments (the workspace was not being correctly
484 retained). Also added to pcre_dfa_exec() a simple plausibility check on
485 some of the workspace data at the beginning of a restart.
486
487 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
488 was not doing so when it should - probably a typo introduced by SVN 528
489 (change 8.10/14).
490
491 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
492 \w+ when the character tables indicated that \x{c4} was a word character.
493 There were several related cases, all because the tests for doing a table
494 lookup were testing for characters less than 127 instead of 255.
495
496 27. If a pattern contains capturing parentheses that are not used in a match,
497 their slots in the ovector are set to -1. For those that are higher than
498 any matched groups, this happens at the end of processing. In the case when
499 there were back references that the ovector was too small to contain
500 (causing temporary malloc'd memory to be used during matching), and the
501 highest capturing number was not used, memory off the end of the ovector
502 was incorrectly being set to -1. (It was using the size of the temporary
503 memory instead of the true size.)
504
505 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
506 ovector size, it uses memory at the end of the block that it has got.
507
508 29. Check for an overlong MARK name and give an error at compile time. The
509 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
510
511 30. JIT compiler update.
512
513 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
514 Rill for the patch.
515
516 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
517
518 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
519
520 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
521 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
522 was enabled.
523
524 35. Improve JIT code generation for greedy plus quantifier.
525
526 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
527 1 to "aa" instead of to an empty string. The bug affected repeated groups
528 that could potentially match an empty string.
529
530 37. Optimizing single character iterators in JIT.
531
532 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
533 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
534 codepoints that are too big for the mode are faulted, and in a UTF mode,
535 disallowed codepoints are also faulted.
536
537 39. If PCRE was compiled with UTF support, in three places in the DFA
538 matcher there was code that should only have been obeyed in UTF mode, but
539 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
540 processing when bytes with values greater than 127 were present. In 16-bit
541 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
542 both cases the values are those that cannot be the first data item in a UTF
543 character. The three items that might have provoked this were recursions,
544 possessively repeated groups, and atomic groups.
545
546 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
547 and pcregrep, because some OS require shared objects to be explicitly
548 passed to ld, causing the link step to fail if they are not.
549
550 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
551 mode, patterns that started with \h* or \R* might be incorrectly matched.
552
553
554 Version 8.30 04-February-2012
555 -----------------------------
556
557 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
558 name is defined in ctype.h.
559
560 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
561 only in quite long subpatterns.
562
563 3. Removed the function pcre_info(), which has been obsolete and deprecated
564 since it was replaced by pcre_fullinfo() in February 2000.
565
566 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
567 match a (*MARK), and the match failed at the start of the subject, a
568 reference to memory before the start of the subject could occur. This bug
569 was introduced by fix 17 of release 8.21.
570
571 5. A reference to an unset group with zero minimum repetition was giving
572 totally wrong answers (in non-JavaScript-compatibility mode). For example,
573 /(another)?(\1?)test/ matched against "hello world test". This bug was
574 introduced in release 8.13.
575
576 6. Add support for 16-bit character strings (a large amount of work involving
577 many changes and refactorings).
578
579 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
580 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
581 from a file.
582
583 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
584 rounding is not applied in this particular case).
585
586 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
587 if they appear, or are escaped, in patterns.
588
589 10. Get rid of a number of -Wunused-but-set-variable warnings.
590
591 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
592 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
593 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
594 also returns the mark "x". This bug applied to capturing parentheses,
595 non-capturing parentheses, and atomic parentheses. It also applied to some
596 assertions.
597
598 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
599 information out of configure.ac instead of relying on pcre.h.generic, which
600 is not stored in the repository.
601
602 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
603 -lreadline.
604
605 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
606
607 15. Applied Graycode's patch to put the top-level frame on the stack rather
608 than the heap when not using the stack for recursion. This gives a
609 performance improvement in many cases when recursion is not deep.
610
611 16. Experimental code added to "pcretest -C" to output the stack frame size.
612
613
614 Version 8.21 12-Dec-2011
615 ------------------------
616
617 1. Updating the JIT compiler.
618
619 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
620 are added as well.
621
622 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
623 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
624 calling _pcre_jit_exec. Some extra comments are added.
625
626 4. (*MARK) settings inside atomic groups that do not contain any capturing
627 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
628 was introduced by change 18 for 8.20.
629
630 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
631 ECMA-262 standard.
632
633 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
634 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
635 This bug was probably introduced by change 9 of 8.13.
636
637 7. While fixing 6 above, I noticed that a number of other items were being
638 incorrectly rejected as "not fixed length". This arose partly because newer
639 opcodes had not been added to the fixed-length checking code. I have (a)
640 corrected the bug and added tests for these items, and (b) arranged for an
641 error to occur if an unknown opcode is encountered while checking for fixed
642 length instead of just assuming "not fixed length". The items that were
643 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
644 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
645 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
646
647 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
648 being incorrectly compiled and would have given unpredicatble results.
649
650 9. A possessively repeated subpattern with minimum repeat count greater than
651 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
652 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
653 the first (A) could occur when it should not.
654
655 10. Add a cast and remove a redundant test from the code.
656
657 11. JIT should use pcre_malloc/pcre_free for allocation.
658
659 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
660 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
661 is anything other than /usr, -L is still shown).
662
663 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
664
665 14. Perl does not support \N without a following name in a [] class; PCRE now
666 also gives an error.
667
668 15. If a forward reference was repeated with an upper limit of around 2000,
669 it caused the error "internal error: overran compiling workspace". The
670 maximum number of forward references (including repeats) was limited by the
671 internal workspace, and dependent on the LINK_SIZE. The code has been
672 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
673 the default depends on LINK_SIZE. There is a new upper limit (for safety)
674 of around 200,000 forward references. While doing this, I also speeded up
675 the filling in of repeated forward references.
676
677 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
678 incorrectly expecting the subject to contain another "a" after the start.
679
680 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
681 in the match, the SKIP should be ignored. This was not happening; instead
682 the SKIP was being treated as NOMATCH. For patterns such as
683 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
684 tested.
685
686 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
687 now much more compatible with Perl, in particular in cases where the result
688 is a non-match for a non-anchored pattern. For example, if
689 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
690 "m", where previously it did not return a name. A side effect of this
691 change is that for partial matches, the last encountered mark name is
692 returned, as for non matches. A number of tests that were previously not
693 Perl-compatible have been moved into the Perl-compatible test files. The
694 refactoring has had the pleasing side effect of removing one argument from
695 the match() function, thus reducing its stack requirements.
696
697 19. If the /S+ option was used in pcretest to study a pattern using JIT,
698 subsequent uses of /S (without +) incorrectly behaved like /S+.
699
700 21. Retrieve executable code size support for the JIT compiler and fixing
701 some warnings.
702
703 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
704 not work when the shorter character appeared right at the end of the
705 subject string.
706
707 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
708 systems.
709
710 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
711 output it when the /M option is used in pcretest.
712
713 25. The CheckMan script was not being included in the distribution. Also, added
714 an explicit "perl" to run Perl scripts from the PrepareRelease script
715 because this is reportedly needed in Windows.
716
717 26. If study data was being save in a file and studying had not found a set of
718 "starts with" bytes for the pattern, the data written to the file (though
719 never used) was taken from uninitialized memory and so caused valgrind to
720 complain.
721
722 27. Updated RunTest.bat as provided by Sheri Pierce.
723
724 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
725
726 29. Computation of memory usage for the table of capturing group names was
727 giving an unnecessarily large value.
728
729
730 Version 8.20 21-Oct-2011
731 ------------------------
732
733 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
734 a POSIX class. After further experiments with Perl, which convinced me that
735 Perl has bugs and confusions, a closing square bracket is no longer allowed
736 in a POSIX name. This bug also affected patterns with classes that started
737 with full stops.
738
739 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
740 captured substring, but while checking the failing first alternative,
741 substring 1 is temporarily captured. If the output vector supplied to
742 pcre_exec() was not big enough for this capture, the yield of the function
743 was still zero ("insufficient space for captured substrings"). This cannot
744 be totally fixed without adding another stack variable, which seems a lot
745 of expense for a edge case. However, I have improved the situation in cases
746 such as /(a)(b)x|abc/ matched against "abc", where the return code
747 indicates that fewer than the maximum number of slots in the ovector have
748 been set.
749
750 3. Related to (2) above: when there are more back references in a pattern than
751 slots in the output vector, pcre_exec() uses temporary memory during
752 matching, and copies in the captures as far as possible afterwards. It was
753 using the entire output vector, but this conflicts with the specification
754 that only 2/3 is used for passing back captured substrings. Now it uses
755 only the first 2/3, for compatibility. This is, of course, another edge
756 case.
757
758 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
759 main code base, and can be used by building with --enable-jit. When this is
760 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
761 runtime --no-jit option is given.
762
763 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
764 ovector, the return from the function was zero, implying that there were
765 other matches that did not fit. The correct "exactly full" value is now
766 returned.
767
768 6. If a subpattern that was called recursively or as a subroutine contained
769 (*PRUNE) or any other control that caused it to give a non-standard return,
770 invalid errors such as "Error -26 (nested recursion at the same subject
771 position)" or even infinite loops could occur.
772
773 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
774 computing the minimum length on reaching *ACCEPT, and so ended up with the
775 wrong value of 1 rather than 0. Further investigation indicates that
776 computing a minimum subject length in the presence of *ACCEPT is difficult
777 (think back references, subroutine calls), and so I have changed the code
778 so that no minimum is registered for a pattern that contains *ACCEPT.
779
780 8. If (*THEN) was present in the first (true) branch of a conditional group,
781 it was not handled as intended. [But see 16 below.]
782
783 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
784 Sheri Pierce.
785
786 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
787 the first byte in a match must be "a".
788
789 11. Change 17 for 8.13 increased the recursion depth for patterns like
790 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
791 pattern contains any instances of (*THEN). If it does not, the old
792 optimizations are restored. It would be nice to do this on a per-group
793 basis, but at the moment that is not feasible.
794
795 12. In some environments, the output of pcretest -C is CRLF terminated. This
796 broke RunTest's code that checks for the link size. A single white space
797 character after the value is now allowed for.
798
799 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
800 For "fr", it uses the Windows-specific input and output files.
801
802 14. If (*THEN) appeared in a group that was called recursively or as a
803 subroutine, it did not work as intended. [But see next item.]
804
805 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
806 pattern fragments (but not containing any | characters). If A and B are
807 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
808 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
809 D. In other words, Perl considers parentheses that do not contain any |
810 characters to be part of a surrounding alternative, whereas PCRE was
811 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
812 differently. PCRE now behaves in the same way as Perl, except in the case
813 of subroutine/recursion calls such as (?1) which have in any case always
814 been different (but PCRE had them first :-).
815
816 16. Related to 15 above: Perl does not treat the | in a conditional group as
817 creating alternatives. Such a group is treated in the same way as an
818 ordinary group without any | characters when processing (*THEN). PCRE has
819 been changed to match Perl's behaviour.
820
821 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
822 RunGrepTest script failed.
823
824 18. Change 22 for version 13 caused atomic groups to use more stack. This is
825 inevitable for groups that contain captures, but it can lead to a lot of
826 stack use in large patterns. The old behaviour has been restored for atomic
827 groups that do not contain any capturing parentheses.
828
829 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
830 suppress the check for a minimum subject length at run time. (If it was
831 given to pcre_exec() or pcre_dfa_exec() it did work.)
832
833 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
834 fail to work when decoding hex characters in data strings in EBCDIC
835 environments.
836
837 21. It appears that in at least one Mac OS environment, the isxdigit() function
838 is implemented as a macro that evaluates to its argument more than once,
839 contravening the C 90 Standard (I haven't checked a later standard). There
840 was an instance in pcretest which caused it to go wrong when processing
841 \x{...} escapes in subject strings. The has been rewritten to avoid using
842 things like p++ in the argument of isxdigit().
843
844
845 Version 8.13 16-Aug-2011
846 ------------------------
847
848 1. The Unicode data tables have been updated to Unicode 6.0.0.
849
850 2. Two minor typos in pcre_internal.h have been fixed.
851
852 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
853 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
854 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
855
856 4. There were a number of related bugs in the code for matching backrefences
857 caselessly in UTF-8 mode when codes for the characters concerned were
858 different numbers of bytes. For example, U+023A and U+2C65 are an upper
859 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
860 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
861 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
862 2-byte code at the end of the subject (it thought there wasn't enough data
863 left).
864
865 5. Comprehensive information about what went wrong is now returned by
866 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
867 as the output vector has at least 2 elements. The offset of the start of
868 the failing character and a reason code are placed in the vector.
869
870 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
871 now returned is for the first byte of the failing character, instead of the
872 last byte inspected. This is an incompatible change, but I hope it is small
873 enough not to be a problem. It makes the returned offset consistent with
874 pcre_exec() and pcre_dfa_exec().
875
876 7. pcretest now gives a text phrase as well as the error number when
877 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
878 failure, the offset and reason code are output.
879
880 8. When \R was used with a maximizing quantifier it failed to skip backwards
881 over a \r\n pair if the subsequent match failed. Instead, it just skipped
882 back over a single character (\n). This seems wrong (because it treated the
883 two characters as a single entity when going forwards), conflicts with the
884 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
885 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
886 has been changed.
887
888 9. Some internal refactoring has changed the processing so that the handling
889 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
890 time (the PCRE_DOTALL option was changed this way some time ago: version
891 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
892 which was always a bit of a fudge. It also means that there is one less
893 argument for the match() function, which reduces its stack requirements
894 slightly. This change also fixes an incompatibility with Perl: the pattern
895 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
896
897 10. More internal refactoring has drastically reduced the number of recursive
898 calls to match() for possessively repeated groups such as (abc)++ when
899 using pcre_exec().
900
901 11. While implementing 10, a number of bugs in the handling of groups were
902 discovered and fixed:
903
904 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
905 (a|)*(?1) gave a compile-time internal error.
906 ((a|)+)+ did not notice that the outer group could match an empty string.
907 (^a|^)+ was not marked as anchored.
908 (.*a|.*)+ was not marked as matching at start or after a newline.
909
910 12. Yet more internal refactoring has removed another argument from the match()
911 function. Special calls to this function are now indicated by setting a
912 value in a variable in the "match data" data block.
913
914 13. Be more explicit in pcre_study() instead of relying on "default" for
915 opcodes that mean there is no starting character; this means that when new
916 ones are added and accidentally left out of pcre_study(), testing should
917 pick them up.
918
919 14. The -s option of pcretest has been documented for ages as being an old
920 synonym of -m (show memory usage). I have changed it to mean "force study
921 for every regex", that is, assume /S for every regex. This is similar to -i
922 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
923 using it. It makes it easier to run collections of tests with and without
924 study enabled, and thereby test pcre_study() more easily. All the standard
925 tests are now run with and without -s (but some patterns can be marked as
926 "never study" - see 20 below).
927
928 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
929 restoration of the capturing data to the outer values was not happening
930 correctly.
931
932 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
933 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
934 pattern had matched an empty string, and so incorrectly returned a no
935 match.
936
937 17. There was optimizing code for the last branch of non-capturing parentheses,
938 and also for the obeyed branch of a conditional subexpression, which used
939 tail recursion to cut down on stack usage. Unfortunately, now that there is
940 the possibility of (*THEN) occurring in these branches, tail recursion is
941 no longer possible because the return has to be checked for (*THEN). These
942 two optimizations have therefore been removed. [But see 8.20/11 above.]
943
944 18. If a pattern containing \R was studied, it was assumed that \R always
945 matched two bytes, thus causing the minimum subject length to be
946 incorrectly computed because \R can also match just one byte.
947
948 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
949 was incorrectly computed.
950
951 20. If /S is present twice on a test pattern in pcretest input, it now
952 *disables* studying, thereby overriding the use of -s on the command line
953 (see 14 above). This is necessary for one or two tests to keep the output
954 identical in both cases.
955
956 21. When (*ACCEPT) was used in an assertion that matched an empty string and
957 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
958
959 22. When an atomic group that contained a capturing parenthesis was
960 successfully matched, but the branch in which it appeared failed, the
961 capturing was not being forgotten if a higher numbered group was later
962 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
963 group 1 to "a", when in fact it should be unset. This applied to multi-
964 branched capturing and non-capturing groups, repeated or not, and also to
965 positive assertions (capturing in negative assertions does not happen
966 in PCRE) and also to nested atomic groups.
967
968 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
969 subject after a captured substring, to make it easier to tell which of a
970 number of identical substrings has been captured.
971
972 24. The way atomic groups are processed by pcre_exec() has been changed so that
973 if they are repeated, backtracking one repetition now resets captured
974 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
975 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
976 "aaa". Previously, it would have been "a". As part of this code
977 refactoring, the way recursive calls are handled has also been changed.
978
979 25. If an assertion condition captured any substrings, they were not passed
980 back unless some other capturing happened later. For example, if
981 (?(?=(a))a) was matched against "a", no capturing was returned.
982
983 26. When studying a pattern that contained subroutine calls or assertions,
984 the code for finding the minimum length of a possible match was handling
985 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
986 group 1 called group 2 while simultaneously a separate group 2 called group
987 1). A stack overflow occurred in this case. I have fixed this by limiting
988 the recursion depth to 10.
989
990 27. Updated RunTest.bat in the distribution to the version supplied by Tom
991 Fortmann. This supports explicit test numbers on the command line, and has
992 argument validation and error reporting.
993
994 28. An instance of \X with an unlimited repeat could fail if at any point the
995 first character it looked at was a mark character.
996
997 29. Some minor code refactoring concerning Unicode properties and scripts
998 should reduce the stack requirement of match() slightly.
999
1000 30. Added the '=' option to pcretest to check the setting of unused capturing
1001 slots at the end of the pattern, which are documented as being -1, but are
1002 not included in the return count.
1003
1004 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1005 compiled something random. Now it gives a compile-time error (as does
1006 Perl).
1007
1008 32. A *MARK encountered during the processing of a positive assertion is now
1009 recorded and passed back (compatible with Perl).
1010
1011 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1012 had alternative anchored branches, the search for a second match in a line
1013 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1014 matched the line "0102" twice. The same bug affected patterns that started
1015 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1016 twice.
1017
1018 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1019 does, and because of capturing effects, quantifying parenthesized
1020 assertions may at times be useful. Quantifiers are now allowed for
1021 parenthesized assertions.
1022
1023 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1024
1025 36. \g was being checked for fancy things in a character class, when it should
1026 just be a literal "g".
1027
1028 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1029 appearance of a nested POSIX class supersedes an apparent external class.
1030 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1031 unescaped square brackets may also appear as part of class names. For
1032 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1033 more like Perl. (But see 8.20/1 above.)
1034
1035 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1036 was because it thought it was \N{name}, which is not supported).
1037
1038 39. Add minix to OS list not supporting the -S option in pcretest.
1039
1040 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1041 cannot analyze patterns in sufficient detail to catch mutual recursions
1042 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1043 subgroup is called recursively as a subpattern for a second time at the
1044 same position in the subject string. In previous releases this might have
1045 been caught by the recursion limit, or it might have run out of stack.
1046
1047 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1048 happen only once. PCRE was, however incorrectly giving a compile time error
1049 "recursive call could loop indefinitely" because it cannot analyze the
1050 pattern in sufficient detail. The compile time test no longer happens when
1051 PCRE is compiling a conditional subpattern, but actual runaway loops are
1052 now caught at runtime (see 40 above).
1053
1054 42. It seems that Perl allows any characters other than a closing parenthesis
1055 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1056 has been changed to be the same.
1057
1058 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1059 as not to get warnings when autogen.sh is called. Also changed
1060 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1061
1062 44. To help people who use pcregrep to scan files containing exceedingly long
1063 lines, the following changes have been made:
1064
1065 (a) The default value of the buffer size parameter has been increased from
1066 8K to 20K. (The actual buffer used is three times this size.)
1067
1068 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1069 PCRE is built.
1070
1071 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1072 to be set at run time.
1073
1074 (d) Numerical values in pcregrep options can be followed by K or M, for
1075 example --buffer-size=50K.
1076
1077 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1078 given and the return code is set to 2.
1079
1080 45. Add a pointer to the latest mark to the callout data block.
1081
1082 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1083 partial match of an empty string instead of no match. This was specific to
1084 the use of ".".
1085
1086 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1087 complete match instead of a partial match. This bug was dependent on both
1088 the PCRE_UTF8 and PCRE_DOTALL options being set.
1089
1090 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1091 starting byte set, because \b was not being ignored.
1092
1093
1094 Version 8.12 15-Jan-2011
1095 ------------------------
1096
1097 1. Fixed some typos in the markup of the man pages, and wrote a script that
1098 checks for such things as part of the documentation building process.
1099
1100 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1101 --match-limit and --recursion-limit options (added for 8.11). In
1102 particular, this made one of the standard tests fail. (The integer value
1103 went into the wrong half of a long int.)
1104
1105 3. If the --colour option was given to pcregrep with -v (invert match), it
1106 did strange things, either producing crazy output, or crashing. It should,
1107 of course, ignore a request for colour when reporting lines that do not
1108 match.
1109
1110 4. Another pcregrep bug caused similar problems if --colour was specified with
1111 -M (multiline) and the pattern match finished with a line ending.
1112
1113 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1114 matched in multiline mode, the following line was shown as part of the
1115 match. This seems wrong, so I have changed it.
1116
1117 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1118 the check for further matches in the same line (so they could be coloured)
1119 to overrun the end of the current line. If another match was found, it was
1120 incorrectly shown (and then shown again when found in the next line).
1121
1122 7. If pcregrep was compiled under Windows, there was a reference to the
1123 function pcregrep_exit() before it was defined. I am assuming this was
1124 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1125 reported by a user. I've moved the definition above the reference.
1126
1127
1128 Version 8.11 10-Dec-2010
1129 ------------------------
1130
1131 1. (*THEN) was not working properly if there were untried alternatives prior
1132 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1133 backtracked to try for "b" instead of moving to the next alternative branch
1134 at the same level (in this case, to look for "c"). The Perl documentation
1135 is clear that when (*THEN) is backtracked onto, it goes to the "next
1136 alternative in the innermost enclosing group".
1137
1138 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1139 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1140 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1141 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1142 (*THEN).
1143
1144 3. If \s appeared in a character class, it removed the VT character from
1145 the class, even if it had been included by some previous item, for example
1146 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1147 of \s, but is part of the POSIX "space" class.)
1148
1149 4. A partial match never returns an empty string (because you can always
1150 match an empty string at the end of the subject); however the checking for
1151 an empty string was starting at the "start of match" point. This has been
1152 changed to the "earliest inspected character" point, because the returned
1153 data for a partial match starts at this character. This means that, for
1154 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1155 (previously it gave "no match").
1156
1157 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1158 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1159 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1160 has an implication that the given string is incomplete (because a partial
1161 match is preferred over a full match). For this reason, these items now
1162 give a partial match in this situation. [Aside: previously, the one case
1163 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1164 match rather than a full match, which was wrong by the old rules, but is
1165 now correct.]
1166
1167 6. There was a bug in the handling of #-introduced comments, recognized when
1168 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1169 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1170 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1171 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1172 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1173 places in pcre_compile().
1174
1175 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1176 comments when looking ahead for named forward references to subpatterns,
1177 the only newline sequence it recognized was NL. It now handles newlines
1178 according to the set newline convention.
1179
1180 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1181 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1182 cater for a lack of strerror(). These oversights have been fixed.
1183
1184 9. Added --match-limit and --recursion-limit to pcregrep.
1185
1186 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1187
1188 11. When the -o option was used, pcregrep was setting a return code of 1, even
1189 when matches were found, and --line-buffered was not being honoured.
1190
1191 12. Added an optional parentheses number to the -o and --only-matching options
1192 of pcregrep.
1193
1194 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1195 can match an empty string. The code to do it in pcretest and pcredemo
1196 needed fixing:
1197
1198 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1199 only one byte after an empty string match just before CRLF (this case
1200 just got forgotten; "any" and "anycrlf" were OK).
1201
1202 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1203 mode when an empty string match preceded an ASCII character followed by
1204 a non-ASCII character. (The code for advancing by one character rather
1205 than one byte was nonsense.)
1206
1207 (c) The pcredemo.c sample program did not have any code at all to handle
1208 the cases when CRLF is a valid newline sequence.
1209
1210 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1211 as a starting offset was within the subject string. There is now a new
1212 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1213 negative or greater than the length of the string. In order to test this,
1214 pcretest is extended to allow the setting of negative starting offsets.
1215
1216 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1217 starting offset points to the beginning of a UTF-8 character was
1218 unnecessarily clumsy. I tidied it up.
1219
1220 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1221 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1222
1223 17. Nobody had reported that the --include_dir option, which was added in
1224 release 7.7 should have been called --include-dir (hyphen, not underscore)
1225 for compatibility with GNU grep. I have changed it to --include-dir, but
1226 left --include_dir as an undocumented synonym, and the same for
1227 --exclude-dir, though that is not available in GNU grep, at least as of
1228 release 2.5.4.
1229
1230 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1231 characters from a string of bytes) have been redefined so as not to use
1232 loops, in order to improve performance in some environments. At the same
1233 time, I abstracted some of the common code into auxiliary macros to save
1234 repetition (this should not affect the compiled code).
1235
1236 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1237 compile-time error is now given if \c is not followed by an ASCII
1238 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1239 different, and any byte value is allowed.)
1240
1241 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1242 START_OPTIMIZE option, which is now allowed at compile time - but just
1243 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1244 to pcregrep and other applications that have no direct access to PCRE
1245 options. The new /Y option in pcretest sets this option when calling
1246 pcre_compile().
1247
1248 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1249 back references. Groups containing recursive back references were forced to
1250 be atomic by that change, but in the case of named groups, the amount of
1251 memory required was incorrectly computed, leading to "Failed: internal
1252 error: code overflow". This has been fixed.
1253
1254 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1255 pcretest.c, to avoid build problems in some Borland environments.
1256
1257
1258 Version 8.10 25-Jun-2010
1259 ------------------------
1260
1261 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1262 THEN.
1263
1264 2. (*ACCEPT) was not working when inside an atomic group.
1265
1266 3. Inside a character class, \B is treated as a literal by default, but
1267 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1268 causes the error). The code is unchanged, but I tidied the documentation.
1269
1270 4. Inside a character class, PCRE always treated \R and \X as literals,
1271 whereas Perl faults them if its -w option is set. I have changed PCRE so
1272 that it faults them when PCRE_EXTRA is set.
1273
1274 5. Added support for \N, which always matches any character other than
1275 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1276
1277 6. When compiling pcregrep with newer versions of gcc which may have
1278 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1279 declared with attribute warn_unused_result" were given. Just casting the
1280 result to (void) does not stop the warnings; a more elaborate fudge is
1281 needed. I've used a macro to implement this.
1282
1283 7. Minor change to pcretest.c to avoid a compiler warning.
1284
1285 8. Added four artifical Unicode properties to help with an option to make
1286 \s etc use properties (see next item). The new properties are: Xan
1287 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1288
1289 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1290 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1291 this option. Modified pcretest to add /W to test this facility. Added
1292 REG_UCP to make it available via the POSIX interface.
1293
1294 10. Added --line-buffered to pcregrep.
1295
1296 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1297 studied, and the match started with a letter with a code point greater than
1298 127 whose first byte was different to the first byte of the other case of
1299 the letter, the other case of this starting letter was not recognized
1300 (#976).
1301
1302 12. If a pattern that was studied started with a repeated Unicode property
1303 test, for example, \p{Nd}+, there was the theoretical possibility of
1304 setting up an incorrect bitmap of starting bytes, but fortunately it could
1305 not have actually happened in practice until change 8 above was made (it
1306 added property types that matched character-matching opcodes).
1307
1308 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1309 possible starting bytes for non-anchored patterns.
1310
1311 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1312 \R, and also a number of cases that involve Unicode properties, both
1313 explicit and implicit when PCRE_UCP is set.
1314
1315 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1316 input, it could crash or give wrong results if characters with values
1317 greater than 0xc0 were present in the subject string. (Detail: it assumed
1318 UTF-8 input when processing these items.)
1319
1320 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1321 size_t is 64-bit (#991).
1322
1323 17. Added a check for running out of memory when PCRE is compiled with
1324 --disable-stack-for-recursion (#990).
1325
1326 18. If the last data line in a file for pcretest does not have a newline on
1327 the end, a newline was missing in the output.
1328
1329 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1330 less than 128) in its various bitmaps. However, there is a facility for
1331 generating tables according to the current locale when PCRE is compiled. It
1332 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1333 characters, are recognized by isspace() and therefore were getting set in
1334 these tables, and indeed these tables seem to approximate to ISO 8859. This
1335 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1336 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1337 which of course cannot start UTF-8 characters. I have changed the code so
1338 that only real ASCII characters (less than 128) and the correct starting
1339 bytes for UTF-8 encodings are set for characters greater than 127 when in
1340 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1341 altogether.)
1342
1343 20. Added the /T option to pcretest so as to be able to run tests with non-
1344 standard character tables, thus making it possible to include the tests
1345 used for 19 above in the standard set of tests.
1346
1347 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1348 reference to a subpattern the other side of a comment that contains an
1349 opening parenthesis caused either an internal compiling error, or a
1350 reference to the wrong subpattern.
1351
1352
1353 Version 8.02 19-Mar-2010
1354 ------------------------
1355
1356 1. The Unicode data tables have been updated to Unicode 5.2.0.
1357
1358 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1359 configured.
1360
1361 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1362 original author of that file, following a query about its status.
1363
1364 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1365 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1366
1367 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1368 quantifier applied to a forward-referencing subroutine call, could compile
1369 incorrect code or give the error "internal error: previously-checked
1370 referenced subpattern not found".
1371
1372 6. Both MS Visual Studio and Symbian OS have problems with initializing
1373 variables to point to external functions. For these systems, therefore,
1374 pcre_malloc etc. are now initialized to local functions that call the
1375 relevant global functions.
1376
1377 7. There were two entries missing in the vectors called coptable and poptable
1378 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1379 I've fixed the data, and added a kludgy way of testing at compile time that
1380 the lengths are correct (equal to the number of opcodes).
1381
1382 8. Following on from 7, I added a similar kludge to check the length of the
1383 eint vector in pcreposix.c.
1384
1385 9. Error texts for pcre_compile() are held as one long string to avoid too
1386 much relocation at load time. To find a text, the string is searched,
1387 counting zeros. There was no check for running off the end of the string,
1388 which could happen if a new error number was added without updating the
1389 string.
1390
1391 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1392
1393 11. \K was not working if it appeared in an atomic group or in a group that
1394 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1395 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1396 the assertion is positive, but not if it is negative.
1397
1398 12. Change 11 fortuitously reduced the size of the stack frame used in the
1399 "match()" function of pcre_exec.c by one pointer. Forthcoming
1400 implementation of support for (*MARK) will need an extra pointer on the
1401 stack; I have reserved it now, so that the stack frame size does not
1402 decrease.
1403
1404 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1405 item in branch that calls a recursion is a subroutine call - as in the
1406 second branch in the above example - was incorrectly given the compile-
1407 time error "recursive call could loop indefinitely" because pcre_compile()
1408 was not correctly checking the subroutine for matching a non-empty string.
1409
1410 14. The checks for overrunning compiling workspace could trigger after an
1411 overrun had occurred. This is a "should never occur" error, but it can be
1412 triggered by pathological patterns such as hundreds of nested parentheses.
1413 The checks now trigger 100 bytes before the end of the workspace.
1414
1415 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1416
1417
1418 Version 8.01 19-Jan-2010
1419 ------------------------
1420
1421 1. If a pattern contained a conditional subpattern with only one branch (in
1422 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1423 computed the wrong minimum data length (which is of course zero for such
1424 subpatterns). This could cause incorrect "no match" results.
1425
1426 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1427 the pattern is reset in the first branch, pcre_compile() failed with
1428 "internal error: code overflow at offset...". This happened only when
1429 the reset was to the original external option setting. (An optimization
1430 abstracts leading options settings into an external setting, which was the
1431 cause of this.)
1432
1433 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1434 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1435 assertion pattern did not match (meaning that the assertion was true), it
1436 was incorrectly treated as false if the SKIP had been reached during the
1437 matching. This also applied to assertions used as conditions.
1438
1439 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1440 assertion subpattern, including such a pattern used as a condition,
1441 unpredictable results occurred, instead of the error return
1442 PCRE_ERROR_DFA_UITEM.
1443
1444 5. The C++ GlobalReplace function was not working like Perl for the special
1445 situation when an empty string is matched. It now does the fancy magic
1446 stuff that is necessary.
1447
1448 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1449 removed. (These were left over from very, very early versions of PCRE.)
1450
1451 7. Some cosmetic changes to the code to make life easier when compiling it
1452 as part of something else:
1453
1454 (a) Change DEBUG to PCRE_DEBUG.
1455
1456 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1457 called "current" as "current_branch", to prevent a collision with the
1458 Linux macro when compiled as a kernel module.
1459
1460 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1461 prevent a collision with the Linux macro when compiled as a kernel
1462 module.
1463
1464 8. In pcre_compile() there are some checks for integer overflows that used to
1465 cast potentially large values to (double). This has been changed to that
1466 when building, a check for int64_t is made, and if it is found, it is used
1467 instead, thus avoiding the use of floating point arithmetic. (There is no
1468 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1469 double.
1470
1471 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1472 2005 (difference between two addresses compared to an unsigned value).
1473
1474 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1475 custom one, because of the following reported problem in Windows:
1476
1477 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1478 under Win32.
1479 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1480 therefore missing the function definition.
1481 - The compiler thus generates a "C" signature for the test function.
1482 - The linker fails to find the "C" function.
1483 - PCRE fails to configure if asked to do so against libbz2.
1484
1485 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1486 messages were output:
1487
1488 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1489 rerunning libtoolize, to keep the correct libtool macros in-tree.
1490 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1491
1492 I have done both of these things.
1493
1494 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1495 most of the time, it *can* run out if it is given a pattern that contains a
1496 runaway infinite recursion. I updated the discussion in the pcrestack man
1497 page.
1498
1499 13. Now that we have gone to the x.xx style of version numbers, the minor
1500 version may start with zero. Using 08 or 09 is a bad idea because users
1501 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1502 interpreted as invalid octal numbers. I've updated the previous comment in
1503 configure.ac, and also added a check that gives an error if 08 or 09 are
1504 used.
1505
1506 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1507 causing partial matching to fail when the end of the subject matched \W
1508 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1509
1510 15. There were some discrepancies between the declarations in pcre_internal.h
1511 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1512 their definitions. The declarations used "const uschar *" and the
1513 definitions used USPTR. Even though USPTR is normally defined as "const
1514 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1515 reported that: "This difference in casting confuses some C++ compilers, for
1516 example, SunCC recognizes above declarations as different functions and
1517 generates broken code for hbpcre." I have changed the declarations to use
1518 USPTR.
1519
1520 16. GNU libtool is named differently on some systems. The autogen.sh script now
1521 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1522 (FreeBSD).
1523
1524 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1525 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1526 comment: "Figure out how to create a longlong from a string: strtoll and
1527 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1528 instance, but it only takes 2 args instead of 3!"
1529
1530 18. A subtle bug concerned with back references has been fixed by a change of
1531 specification, with a corresponding code fix. A pattern such as
1532 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1533 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1534 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1535 same bug. Such groups have to be quantified to be useful, or contained
1536 inside another quantified group. (If there's no repetition, the reference
1537 can never match.) The problem arises because, having left the group and
1538 moved on to the rest of the pattern, a later failure that backtracks into
1539 the group uses the captured value from the final iteration of the group
1540 rather than the correct earlier one. I have fixed this in PCRE by forcing
1541 any group that contains a reference to itself to be an atomic group; that
1542 is, there cannot be any backtracking into it once it has completed. This is
1543 similar to recursive and subroutine calls.
1544
1545
1546 Version 8.00 19-Oct-09
1547 ----------------------
1548
1549 1. The table for translating pcre_compile() error codes into POSIX error codes
1550 was out-of-date, and there was no check on the pcre_compile() error code
1551 being within the table. This could lead to an OK return being given in
1552 error.
1553
1554 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1555 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1556 in a Windows environment.
1557
1558 3. The pcregrep --count option prints the count for each file even when it is
1559 zero, as does GNU grep. However, pcregrep was also printing all files when
1560 --files-with-matches was added. Now, when both options are given, it prints
1561 counts only for those files that have at least one match. (GNU grep just
1562 prints the file name in this circumstance, but including the count seems
1563 more useful - otherwise, why use --count?) Also ensured that the
1564 combination -clh just lists non-zero counts, with no names.
1565
1566 4. The long form of the pcregrep -F option was incorrectly implemented as
1567 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1568 but it seems right to fix it, and I didn't think it was worth preserving
1569 the old behaviour.
1570
1571 5. The command line items --regex=pattern and --regexp=pattern were not
1572 recognized by pcregrep, which required --regex pattern or --regexp pattern
1573 (with a space rather than an '='). The man page documented the '=' forms,
1574 which are compatible with GNU grep; these now work.
1575
1576 6. No libpcreposix.pc file was created for pkg-config; there was just
1577 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1578
1579 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1580 when UCP support is not needed, by modifying the Python script that
1581 generates it from Unicode data files. This should not matter if the module
1582 is correctly used as a library, but I received one complaint about 50K of
1583 unwanted data. My guess is that the person linked everything into his
1584 program rather than using a library. Anyway, it does no harm.
1585
1586 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1587 was a minimum greater than 1 for a wide character in a possessive
1588 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1589 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1590 character. Chaos in the form of incorrect output or a compiling loop could
1591 result.
1592
1593 9. The restrictions on what a pattern can contain when partial matching is
1594 requested for pcre_exec() have been removed. All patterns can now be
1595 partially matched by this function. In addition, if there are at least two
1596 slots in the offset vector, the offset of the earliest inspected character
1597 for the match and the offset of the end of the subject are set in them when
1598 PCRE_ERROR_PARTIAL is returned.
1599
1600 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1601 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1602 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1603 and may be more useful for multi-segment matching.
1604
1605 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1606 used to be given if ever the end of the subject was reached; now it is
1607 given only if matching could not proceed because another character was
1608 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1609 string "Z", which now yields "no match" instead of "partial match". In the
1610 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1611 final character ended with (*FAIL).
1612
1613 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1614 if the pattern had a "must contain" character that was already found in the
1615 earlier partial match, unless partial matching was again requested. For
1616 example, with the pattern /dog.(body)?/, the "must contain" character is
1617 "g". If the first part-match was for the string "dog", restarting with
1618 "sbody" failed. This bug has been fixed.
1619
1620 13. The string returned by pcre_dfa_exec() after a partial match has been
1621 changed so that it starts at the first inspected character rather than the
1622 first character of the match. This makes a difference only if the pattern
1623 starts with a lookbehind assertion or \b or \B (\K is not supported by
1624 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1625 matching functions compatible, and I think it's the right thing to do.
1626
1627 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1628 so that the demonstration program is easily available in environments where
1629 PCRE has not been installed from source.
1630
1631 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1632 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1633 library.
1634
1635 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1636 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1637 is not the first non-POSIX option to be added. Clearly some people find
1638 these options useful.
1639
1640 17. If a caller to the POSIX matching function regexec() passes a non-zero
1641 value for nmatch with a NULL value for pmatch, the value of
1642 nmatch is forced to zero.
1643
1644 18. RunGrepTest did not have a test for the availability of the -u option of
1645 the diff command, as RunTest does. It now checks in the same way as
1646 RunTest, and also checks for the -b option.
1647
1648 19. If an odd number of negated classes containing just a single character
1649 interposed, within parentheses, between a forward reference to a named
1650 subpattern and the definition of the subpattern, compilation crashed with
1651 an internal error, complaining that it could not find the referenced
1652 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1653 [The bug was that it was starting one character too far in when skipping
1654 over the character class, thus treating the ] as data rather than
1655 terminating the class. This meant it could skip too much.]
1656
1657 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1658 /g option in pcretest when the pattern contains \K, which makes it possible
1659 to have an empty string match not at the start, even when the pattern is
1660 anchored. Updated pcretest and pcredemo to use this option.
1661
1662 21. If the maximum number of capturing subpatterns in a recursion was greater
1663 than the maximum at the outer level, the higher number was returned, but
1664 with unset values at the outer level. The correct (outer level) value is
1665 now given.
1666
1667 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1668 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1669 make it do so. The string so far is captured, making this feature
1670 compatible with Perl.
1671
1672 23. The tests have been re-organized, adding tests 11 and 12, to make it
1673 possible to check the Perl 5.10 features against Perl 5.10.
1674
1675 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1676 pattern matches a fixed length string. PCRE did not allow this; now it
1677 does. Neither allows recursion.
1678
1679 25. I finally figured out how to implement a request to provide the minimum
1680 length of subject string that was needed in order to match a given pattern.
1681 (It was back references and recursion that I had previously got hung up
1682 on.) This code has now been added to pcre_study(); it finds a lower bound
1683 to the length of subject needed. It is not necessarily the greatest lower
1684 bound, but using it to avoid searching strings that are too short does give
1685 some useful speed-ups. The value is available to calling programs via
1686 pcre_fullinfo().
1687
1688 26. While implementing 25, I discovered to my embarrassment that pcretest had
1689 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1690 study optimizations had never been tested with that matching function.
1691 Oops. What is worse, even when it was passed study data, there was a bug in
1692 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1693 were also very few tests of studied patterns with pcre_dfa_exec().
1694
1695 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1696 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1697 on the other side of the coin, they are no longer allowed to have different
1698 names, because these cannot be distinguished in PCRE, and this has caused
1699 confusion. (This is a difference from Perl.)
1700
1701 28. When duplicate subpattern names are present (necessarily with different
1702 numbers, as required by 27 above), and a test is made by name in a
1703 conditional pattern, either for a subpattern having been matched, or for
1704 recursion in such a pattern, all the associated numbered subpatterns are
1705 tested, and the overall condition is true if the condition is true for any
1706 one of them. This is the way Perl works, and is also more like the way
1707 testing by number works.
1708
1709
1710 Version 7.9 11-Apr-09
1711 ---------------------
1712
1713 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1714 (pcretest), all targets were linked against these libraries. This included
1715 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1716 libraries. This caused unwanted dependencies to be created. This problem
1717 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1718 pcretest is linked with readline.
1719
1720 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1721 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1722 moved outside it again, because FALSE and TRUE are already defined in AIX,
1723 but BOOL is not.
1724
1725 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1726 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1727
1728 4. The pcregrep documentation said spaces were inserted as well as colons (or
1729 hyphens) following file names and line numbers when outputting matching
1730 lines. This is not true; no spaces are inserted. I have also clarified the
1731 wording for the --colour (or --color) option.
1732
1733 5. In pcregrep, when --colour was used with -o, the list of matching strings
1734 was not coloured; this is different to GNU grep, so I have changed it to be
1735 the same.
1736
1737 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1738 each matching line was coloured. Now it goes on to look for further matches
1739 of any of the test patterns, which is the same behaviour as GNU grep.
1740
1741 7. A pattern that could match an empty string could cause pcregrep to loop; it
1742 doesn't make sense to accept an empty string match in pcregrep, so I have
1743 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1744 seems to be how GNU grep behaves. [But see later change 40 for release
1745 8.33.]
1746
1747 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1748 start or after a newline", because the conditional assertion was not being
1749 correctly handled. The rule now is that both the assertion and what follows
1750 in the first alternative must satisfy the test.
1751
1752 9. If auto-callout was enabled in a pattern with a conditional group whose
1753 condition was an assertion, PCRE could crash during matching, both with
1754 pcre_exec() and pcre_dfa_exec().
1755
1756 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1757 used for matching.
1758
1759 11. Unicode property support in character classes was not working for
1760 characters (bytes) greater than 127 when not in UTF-8 mode.
1761
1762 12. Added the -M command line option to pcretest.
1763
1764 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1765
1766 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1767
1768 16. Added comments and documentation about mis-use of no_arg in the C++
1769 wrapper.
1770
1771 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1772 from Martin Jerabek that uses macro names for all relevant character and
1773 string constants.
1774
1775 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1776 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1777 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1778 these, but not everybody uses configure.
1779
1780 19. A conditional group that had only one branch was not being correctly
1781 recognized as an item that could match an empty string. This meant that an
1782 enclosing group might also not be so recognized, causing infinite looping
1783 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1784 with the subject "ab", where knowledge that the repeated group can match
1785 nothing is needed in order to break the loop.
1786
1787 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1788 exec(), but without supplying a callout function, matching went wrong.
1789
1790 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1791 leak if the size of the offset vector was greater than 30. When the vector
1792 is smaller, the saved offsets during recursion go onto a local stack
1793 vector, but for larger vectors malloc() is used. It was failing to free
1794 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1795 error, in fact).
1796
1797 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1798 heapframe that is used only when UTF-8 support is enabled. This caused no
1799 problem, but was untidy.
1800
1801 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1802 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1803 included within another project.
1804
1805 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1806 slightly modified by me:
1807
1808 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1809 not building pcregrep.
1810
1811 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1812 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1813
1814 25. Forward references, both numeric and by name, in patterns that made use of
1815 duplicate group numbers, could behave incorrectly or give incorrect errors,
1816 because when scanning forward to find the reference group, PCRE was not
1817 taking into account the duplicate group numbers. A pattern such as
1818 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1819
1820 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1821 the feature of a custom pointer more persuasive (as requested by a user).
1822
1823 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1824 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1825 pre-defined.
1826
1827 28. Added support for (*UTF8) at the start of a pattern.
1828
1829 29. Arrange for flags added by the "release type" setting in CMake to be shown
1830 in the configuration summary.
1831
1832
1833 Version 7.8 05-Sep-08
1834 ---------------------
1835
1836 1. Replaced UCP searching code with optimized version as implemented for Ad
1837 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1838 stage table and inline lookup instead of a function, giving speed ups of 2
1839 to 5 times on some simple patterns that I tested. Permission was given to
1840 distribute the MultiStage2.py script that generates the tables (it's not in
1841 the tarball, but is in the Subversion repository).
1842
1843 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1844 scripts.
1845
1846 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1847 a group with a zero qualifier. The result of the study could be incorrect,
1848 or the function might crash, depending on the pattern.
1849
1850 4. Caseless matching was not working for non-ASCII characters in back
1851 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1852 It now works when Unicode Property Support is available.
1853
1854 5. In pcretest, an escape such as \x{de} in the data was always generating
1855 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1856 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1857 truncation.
1858
1859 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1860
1861 7. Added two (int) casts to pcregrep when printing the difference of two
1862 pointers, in case they are 64-bit values.
1863
1864 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1865 test 2 if it fails.
1866
1867 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1868 and a #define of that name to empty if it is not externally set. This is to
1869 allow users of MSVC to set it if necessary.
1870
1871 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1872 the convenience functions in the pcre_get.c source file.
1873
1874 11. An option change at the start of a pattern that had top-level alternatives
1875 could cause overwriting and/or a crash. This command provoked a crash in
1876 some environments:
1877
1878 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1879
1880 This potential security problem was recorded as CVE-2008-2371.
1881
1882 12. For a pattern where the match had to start at the beginning or immediately
1883 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1884 pcre_dfa_exec() could read past the end of the passed subject if there was
1885 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1886 pcretest so that it places the subject at the end of its malloc-ed buffer.
1887
1888 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1889 exec() might read past the end of the data buffer in UTF-8 mode.
1890
1891 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1892 the data contained the byte 0x85 as part of a UTF-8 character within its
1893 first line. This applied both to normal and DFA matching.
1894
1895 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1896 /^[^d]*?$/8 failed to match "abc".
1897
1898 16. Added a missing copyright notice to pcrecpp_internal.h.
1899
1900 17. Make it more clear in the documentation that values returned from
1901 pcre_exec() in ovector are byte offsets, not character counts.
1902
1903 18. Tidied a few places to stop certain compilers from issuing warnings.
1904
1905 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1906 supplied by Stefan Weber. I made a further small update for 7.8 because
1907 there is a change of source arrangements: the pcre_searchfuncs.c module is
1908 replaced by pcre_ucd.c.
1909
1910
1911 Version 7.7 07-May-08
1912 ---------------------
1913
1914 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1915 a string to a long long, pretend we don't even have a long long." This is
1916 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1917
1918 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1919 pre-7.6 versions, which defined a global no_arg variable instead of putting
1920 it in the RE class. (See also #8 below.)
1921
1922 3. Remove a line of dead code, identified by coverity and reported by Nuno
1923 Lopes.
1924
1925 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1926
1927 (1) The include/exclude patterns were being applied to the whole pathnames
1928 of files, instead of just to the final components.
1929
1930 (2) If there was more than one level of directory, the subdirectories were
1931 skipped unless they satisfied the include/exclude conditions. This is
1932 inconsistent with GNU grep (and could even be seen as contrary to the
1933 pcregrep specification - which I improved to make it absolutely clear).
1934 The action now is always to scan all levels of directory, and just
1935 apply the include/exclude patterns to regular files.
1936
1937 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1938 --exclude_dir in the tests to avoid scanning .svn directories.
1939
1940 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1941 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1942 doesn't support NULs in patterns.
1943
1944 7. Added some missing "const"s to declarations of static tables in
1945 pcre_compile.c and pcre_dfa_exec.c.
1946
1947 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1948 caused by fix #2 above. (Subsequently also a second patch to fix the
1949 first patch. And a third patch - this was a messy problem.)
1950
1951 9. Applied Craig's patch to remove the use of push_back().
1952
1953 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1954 matching function regexec().
1955
1956 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1957 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1958 references. PCRE supports relative numbers with this syntax (I don't think
1959 Oniguruma does).
1960
1961 12. Previously, a group with a zero repeat such as (...){0} was completely
1962 omitted from the compiled regex. However, this means that if the group
1963 was called as a subroutine from elsewhere in the pattern, things went wrong
1964 (an internal error was given). Such groups are now left in the compiled
1965 pattern, with a new opcode that causes them to be skipped at execution
1966 time.
1967
1968 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1969 to the way PCRE behaves:
1970
1971 (a) A lone ] character is dis-allowed (Perl treats it as data).
1972
1973 (b) A back reference to an unmatched subpattern matches an empty string
1974 (Perl fails the current match path).
1975
1976 (c) A data ] in a character class must be notated as \] because if the
1977 first data character in a class is ], it defines an empty class. (In
1978 Perl it is not possible to have an empty class.) The empty class []
1979 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1980 The negative empty class [^] matches any one character, independently
1981 of the DOTALL setting.
1982
1983 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1984 non-existent subpattern following a character class starting with ']' and
1985 containing () gave an internal compiling error instead of "reference to
1986 non-existent subpattern". Fortunately, when the pattern did exist, the
1987 compiled code was correct. (When scanning forwards to check for the
1988 existence of the subpattern, it was treating the data ']' as terminating
1989 the class, so got the count wrong. When actually compiling, the reference
1990 was subsequently set up correctly.)
1991
1992 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1993 it was being rejected as not supported by pcre_dfa_exec(), even though
1994 other assertions are supported. I have made pcre_dfa_exec() support
1995 (*FAIL).
1996
1997 16. The implementation of 13c above involved the invention of a new opcode,
1998 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1999 cannot be changed at match time, I realized I could make a small
2000 improvement to matching performance by compiling OP_ALLANY instead of
2001 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2002 on the OP_ANY path.
2003
2004 17. Compiling pcretest on Windows with readline support failed without the
2005 following two fixes: (1) Make the unistd.h include conditional on
2006 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2007
2008 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2009 ncurses library to be included for pcretest when ReadLine support is
2010 requested, but also to allow for it to be overridden. This patch came from
2011 Daniel Bergström.
2012
2013 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2014 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2015 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2016 spotting this.
2017
2018
2019 Version 7.6 28-Jan-08
2020 ---------------------
2021
2022 1. A character class containing a very large number of characters with
2023 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2024 overflow.
2025
2026 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2027 HAVE_LONG_LONG is not defined.
2028
2029 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2030 bring them up to date and include new features. This patch includes:
2031
2032 - Fixed PH's badly added libz and libbz2 support.
2033 - Fixed a problem with static linking.
2034 - Added pcredemo. [But later removed - see 7 below.]
2035 - Fixed dftables problem and added an option.
2036 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2037 HAVE_LONG_LONG.
2038 - Added readline support for pcretest.
2039 - Added an listing of the option settings after cmake has run.
2040
2041 4. A user submitted a patch to Makefile that makes it easy to create
2042 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2043 Makefile.am that cause it to include this special target, without
2044 affecting anything else. Note that the same mingw target plus all
2045 the other distribution libraries and programs are now supported
2046 when configuring with CMake (see 6 below) instead of with
2047 Configure/Make.
2048
2049 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2050 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2051 exported in the Windows port". It has not yet been confirmed that the patch
2052 solves the problem, but it does no harm.
2053
2054 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2055 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2056 with CMake, and also correct the comment about stack recursion.
2057
2058 7. Remove the automatic building of pcredemo from the ./configure system and
2059 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2060 of a program that users should build themselves after PCRE is installed, so
2061 building it automatically is not really right. What is more, it gave
2062 trouble in some build environments.
2063
2064 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2065
2066
2067 Version 7.5 10-Jan-08
2068 ---------------------
2069
2070 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2071 values in parens when parsing an RE using the C++ wrapper."
2072
2073 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2074 Characters greater than 255 were excluded from the class instead of being
2075 included.
2076
2077 3. The same bug as (2) above applied to negated POSIX classes such as
2078 [:^space:].
2079
2080 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2081 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2082 I have changed it.
2083
2084 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2085 first named subpattern) and a construct such as (?&a) would reference the
2086 first named subpattern whose name started with "a" (in other words, the
2087 length check was missing). Both these problems are fixed. "Subpattern name
2088 expected" is now given for (?&) (a zero-length name), and this patch also
2089 makes it give the same error for \k'' (previously it complained that that
2090 was a reference to a non-existent subpattern).
2091
2092 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2093 this is right because (?- can be followed by option settings as well as by
2094 digits. I have, however, made the messages clearer.
2095
2096 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2097 than the number used in the conditional) now cause a compile-time error.
2098 This is actually not compatible with Perl, which accepts such patterns, but
2099 treats the conditional as always being FALSE (as PCRE used to), but it
2100 seems to me that giving a diagnostic is better.
2101
2102 8. Change "alphameric" to the more common word "alphanumeric" in comments
2103 and messages.
2104
2105 9. Fix two occurrences of "backslash" in comments that should have been
2106 "backspace".
2107
2108 10. Remove two redundant lines of code that can never be obeyed (their function
2109 was moved elsewhere).
2110
2111 11. The program that makes PCRE's Unicode character property table had a bug
2112 which caused it to generate incorrect table entries for sequences of
2113 characters that have the same character type, but are in different scripts.
2114 It amalgamated them into a single range, with the script of the first of
2115 them. In other words, some characters were in the wrong script. There were
2116 thirteen such cases, affecting characters in the following ranges:
2117
2118 U+002b0 - U+002c1
2119 U+0060c - U+0060d
2120 U+0061e - U+00612
2121 U+0064b - U+0065e
2122 U+0074d - U+0076d
2123 U+01800 - U+01805
2124 U+01d00 - U+01d77
2125 U+01d9b - U+01dbf
2126 U+0200b - U+0200f
2127 U+030fc - U+030fe
2128 U+03260 - U+0327f
2129 U+0fb46 - U+0fbb1
2130 U+10450 - U+1049d
2131
2132 12. The -o option (show only the matching part of a line) for pcregrep was not
2133 compatible with GNU grep in that, if there was more than one match in a
2134 line, it showed only the first of them. It now behaves in the same way as
2135 GNU grep.
2136
2137 13. If the -o and -v options were combined for pcregrep, it printed a blank
2138 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2139 does the same. The return code can be used to tell if there were any
2140 non-matching lines.
2141
2142 14. Added --file-offsets and --line-offsets to pcregrep.
2143
2144 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2145 infinitely looping recursion. The bug was that positive lookaheads were not
2146 being skipped when checking for a possible empty match (negative lookaheads
2147 and both kinds of lookbehind were skipped).
2148
2149 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2150 inclusion of <windows.h> to before rather than after the definition of
2151 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2152
2153 17. Specifying a possessive quantifier with a specific limit for a Unicode
2154 character property caused pcre_compile() to compile bad code, which led at
2155 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2156 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2157 caused the error; without that there was no problem.
2158
2159 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2160
2161 19. Added --enable-pcretest-libreadline.
2162
2163 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2164 RE::GlobalReplace(). As a result, the number of replacements returned was
2165 double what it should be. I removed one of the increments, but Craig sent a
2166 later patch that removed the other one (the right fix) and added unit tests
2167 that check the return values (which was not done before).
2168
2169 21. Several CMake things:
2170
2171 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2172 the names libpcre and libpcreposix, not just pcre and pcreposix.
2173
2174 (2) The above change means that pcretest and pcregrep are now correctly
2175 linked with the newly-built libraries, not previously installed ones.
2176
2177 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2178
2179 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2180 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2181 UTF-8 newline character). The key issue is that the pattern starts .*;
2182 this means that the match must be either at the beginning, or after a
2183 newline. The bug was in the code for advancing after a failed match and
2184 checking that the new position followed a newline. It was not taking
2185 account of UTF-8 characters correctly.
2186
2187 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2188 character classes. PCRE was not treating the sequence [:...:] as a
2189 character class unless the ... were all letters. Perl, however, seems to
2190 allow any characters between [: and :], though of course it rejects as
2191 unknown any "names" that contain non-letters, because all the known class
2192 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2193 for example, whereas PCRE did not - it did not recognize a POSIX character
2194 class. This seemed a bit dangerous, so the code has been changed to be
2195 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2196 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2197 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2198 Perl does, and where it didn't before.
2199
2200 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2201 Windows environments %n is disabled by default.
2202
2203
2204 Version 7.4 21-Sep-07
2205 ---------------------
2206
2207 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2208 means that a class such as [\s] counted as "explicit reference to CR or
2209 LF". That isn't really right - the whole point of the change was to try to
2210 help when there was an actual mention of one of the two characters. So now
2211 the change happens only if \r or \n (or a literal CR or LF) character is
2212 encountered.
2213
2214 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2215 of both had grown to the point where there were only 3 bits left.
2216 Fortunately, there was spare space in the data structure, and so I have
2217 moved the internal flags into a new 16-bit field to free up more option
2218 bits.
2219
2220 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2221 but did not set the internal JCHANGED flag - either of these is enough to
2222 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2223 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2224 start sets both bits.
2225
2226 4. Added options (at build time, compile time, exec time) to change \R from
2227 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2228
2229 5. doc/pcresyntax.html was missing from the distribution.
2230
2231 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2232 compatibility, even though it is no longer used.
2233
2234 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2235 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2236 windows.h file is present (where different names are used). [This was
2237 reversed later after testing - see 16 below.]
2238
2239 8. Changed all #include <config.h> to #include "config.h". There were also
2240 some further <pcre.h> cases that I changed to "pcre.h".
2241
2242 9. When pcregrep was used with the --colour option, it missed the line ending
2243 sequence off the lines that it output.
2244
2245 10. It was pointed out to me that arrays of string pointers cause lots of
2246 relocations when a shared library is dynamically loaded. A technique of
2247 using a single long string with a table of offsets can drastically reduce
2248 these. I have refactored PCRE in four places to do this. The result is
2249 dramatic:
2250
2251 Originally: 290
2252 After changing UCP table: 187
2253 After changing error message table: 43
2254 After changing table of "verbs" 36
2255 After changing table of Posix names 22
2256
2257 Thanks to the folks working on Gregex for glib for this insight.
2258
2259 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2260 unicode-properties was also set.
2261
2262 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2263
2264 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2265 checked only for CRLF.
2266
2267 14. Added casts to pcretest.c to avoid compiler warnings.
2268
2269 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2270
2271 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2272 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2273 entirely. This removes changes made in 7 above.
2274
2275 17. The CMake files have been updated, and there is now more information about
2276 building with CMake in the NON-UNIX-USE document.
2277
2278
2279 Version 7.3 28-Aug-07
2280 ---------------------
2281
2282 1. In the rejigging of the build system that eventually resulted in 7.1, the
2283 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2284 brackets there is not right, since it causes compilers to look for an
2285 installed pcre.h, not the version that is in the source that is being
2286 compiled (which of course may be different). I have changed it back to:
2287
2288 #include "pcre.h"
2289
2290 I have a vague recollection that the change was concerned with compiling in
2291 different directories, but in the new build system, that is taken care of
2292 by the VPATH setting the Makefile.
2293
2294 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2295 when the subject happened to end in the byte 0x85 (e.g. if the last
2296 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2297 characters but of course it shouldn't be taken as a newline when it is part
2298 of another character. The bug was that, for an unlimited repeat of . in
2299 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2300 characters when looking for a newline.
2301
2302 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2303
2304 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2305 in debug output.
2306
2307 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2308 long printing in the pcrecpp unittest when running under MinGW.
2309
2310 6. ESC_K was left out of the EBCDIC table.
2311
2312 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2313 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2314 limit also applies to "virtual nesting" when a pattern is recursive, and in
2315 this case 1000 isn't so big. I have been able to remove this limit at the
2316 expense of backing off one optimization in certain circumstances. Normally,
2317 when pcre_exec() would call its internal match() function recursively and
2318 immediately return the result unconditionally, it uses a "tail recursion"
2319 feature to save stack. However, when a subpattern that can match an empty
2320 string has an unlimited repetition quantifier, it no longer makes this
2321 optimization. That gives it a stack frame in which to save the data for
2322 checking that an empty string has been matched. Previously this was taken
2323 from the 1000-entry workspace that had been reserved. So now there is no
2324 explicit limit, but more stack is used.
2325
2326 8. Applied Daniel's patches to solve problems with the import/export magic
2327 syntax that is required for Windows, and which was going wrong for the
2328 pcreposix and pcrecpp parts of the library. These were overlooked when this
2329 problem was solved for the main library.
2330
2331 9. There were some crude static tests to avoid integer overflow when computing
2332 the size of patterns that contain repeated groups with explicit upper
2333 limits. As the maximum quantifier is 65535, the maximum group length was
2334 set at 30,000 so that the product of these two numbers did not overflow a
2335 32-bit integer. However, it turns out that people want to use groups that
2336 are longer than 30,000 bytes (though not repeat them that many times).
2337 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2338 made it possible to implement the integer overflow checks in a much more
2339 dynamic way, which I have now done. The artificial limitation on group
2340 length has been removed - we now have only the limit on the total length of
2341 the compiled pattern, which depends on the LINK_SIZE setting.
2342
2343 10. Fixed a bug in the documentation for get/copy named substring when
2344 duplicate names are permitted. If none of the named substrings are set, the
2345 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2346 empty string.
2347
2348 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2349 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2350 because the ] is interpreted as the first data character and the
2351 terminating ] is not found. PCRE has been made compatible with Perl in this
2352 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2353 cause memory overwriting.
2354
2355 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2356 string has been matched (to stop an infinite loop). It was not recognizing
2357 a conditional subpattern that could match an empty string if that
2358 subpattern was within another subpattern. For example, it looped when
2359 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2360 condition was not nested. This bug has been fixed.
2361
2362 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2363 past the start of the subject in the presence of bytes with the top bit
2364 set, for example "\x8aBCD".
2365
2366 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2367 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2368
2369 14. Optimized (?!) to (*FAIL).
2370
2371 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2372 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2373 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2374 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2375 does: it's just the validity check that is more restrictive.
2376
2377 16. Inserted checks for integer overflows during escape sequence (backslash)
2378 processing, and also fixed erroneous offset values for syntax errors during
2379 backslash processing.
2380
2381 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2382 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2383
2384 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2385 caused an overrun.
2386
2387 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2388 something other than just ASCII characters) inside a group that had an
2389 unlimited repeat caused a loop at compile time (while checking to see
2390 whether the group could match an empty string).
2391
2392 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2393 [\P{Any}] did so. (Error in the code for printing property names.)
2394
2395 21. An orphan \E inside a character class could cause a crash.
2396
2397 22. A repeated capturing bracket such as (A)? could cause a wild memory
2398 reference during compilation.
2399
2400 23. There are several functions in pcre_compile() that scan along a compiled
2401 expression for various reasons (e.g. to see if it's fixed length for look
2402 behind). There were bugs in these functions when a repeated \p or \P was
2403 present in the pattern. These operators have additional parameters compared
2404 with \d, etc, and these were not being taken into account when moving along
2405 the compiled data. Specifically:
2406
2407 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2408 length.
2409
2410 (b) An item such as \pL+ within a repeated group could cause crashes or
2411 loops.
2412
2413 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2414 "reference to non-existent subpattern" error.
2415
2416 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2417
2418 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2419 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2420
2421 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2422
2423 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2424 character were causing crashes (broken optimization).
2425
2426 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2427 \p or \P) caused a compile-time loop.
2428
2429 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2430 break. For example, the unstudied pattern [\r\n]A does not match the string
2431 "\r\nA" because change 7.0/46 below moves the current point on by two
2432 characters after failing to match at the start. However, the pattern \nA
2433 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2434 the same is true. There doesn't seem any very clean way out of this, but
2435 what I have chosen to do makes the common cases work: PCRE now takes note
2436 of whether there can be an explicit match for \r or \n anywhere in the
2437 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2438 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2439 pattern has explicit CR or LF references.
2440
2441 29. Added (*CR) etc for changing newline setting at start of pattern.
2442
2443
2444 Version 7.2 19-Jun-07
2445 ---------------------
2446
2447 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2448 which is apparently normally available under Windows.
2449
2450 2. Re-jig the pcregrep tests with different newline settings in an attempt
2451 to make them independent of the local environment's newline setting.
2452
2453 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2454
2455 4. Some of the "internals" tests were previously cut out when the link size
2456 was not 2, because the output contained actual offsets. The recent new
2457 "Z" feature of pcretest means that these can be cut out, making the tests
2458 usable with all link sizes.
2459
2460 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2461 stack recursion. This gives a massive performance boost under BSD, but just
2462 a small improvement under Linux. However, it saves one field in the frame
2463 in all cases.
2464
2465 6. Added more features from the forthcoming Perl 5.10:
2466
2467 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2468 recursion call. It refers to the nth most recently opened parentheses.
2469
2470 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2471 to be opened parentheses.
2472
2473 (c) Conditions that refer to capturing parentheses can be specified
2474 relatively, for example, (?(-2)... or (?(+3)...
2475
2476 (d) \K resets the start of the current match so that everything before
2477 is not part of it.
2478
2479 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2480
2481 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2482 reference syntax.
2483
2484 (g) (?| introduces a group in which the numbering of parentheses in each
2485 alternative starts with the same number.
2486
2487 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2488
2489 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2490 PCRE_INFO_JCHANGED.
2491
2492 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2493 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2494 for detecting groups that can match an empty string.
2495
2496 9. A pattern with a very large number of alternatives (more than several
2497 hundred) was running out of internal workspace during the pre-compile
2498 phase, where pcre_compile() figures out how much memory will be needed. A
2499 bit of new cunning has reduced the workspace needed for groups with
2500 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2501 workspace instead of running out of the 4096 that are available.
2502
2503 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2504
2505 11. Applied patch from Google to remove an optimization that didn't quite work.
2506 The report of the bug said:
2507
2508 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2509 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2510 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2511
2512 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2513 it matched the wrong number of bytes.
2514
2515
2516 Version 7.1 24-Apr-07
2517 ---------------------
2518
2519 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2520 that is more "standard", making use of automake and other Autotools. There
2521 is some re-arrangement of the files and adjustment of comments consequent
2522 on this.
2523
2524 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2525 for recursive directory scanning broke on some systems because the files
2526 are not scanned in any specific order and on different systems the order
2527 was different. A call to "sort" has been inserted into RunGrepTest for the
2528 approprate test as a short-term fix. In the longer term there may be an
2529 alternative.
2530
2531 3. I had an email from Eric Raymond about problems translating some of PCRE's
2532 man pages to HTML (despite the fact that I distribute HTML pages, some
2533 people do their own conversions for various reasons). The problems
2534 concerned the use of low-level troff macros .br and .in. I have therefore
2535 removed all such uses from the man pages (some were redundant, some could
2536 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2537 HTML has been updated to handle .nf/.fi and to complain if it encounters
2538 .br or .in.
2539
2540 4. Updated comments in configure.ac that get placed in config.h.in and also
2541 arranged for config.h to be included in the distribution, with the name
2542 config.h.generic, for the benefit of those who have to compile without
2543 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2544
2545 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2546 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2547 makevp.bat for the current PCRE, using the additional files
2548 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2549
2550 6. A Windows user reported a minor discrepancy with test 2, which turned out
2551 to be caused by a trailing space on an input line that had got lost in his
2552 copy. The trailing space was an accident, so I've just removed it.
2553
2554 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2555 that is needed.
2556
2557 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2558 as "const" (a) because they are and (b) because it helps the PHP
2559 maintainers who have recently made a script to detect big data structures
2560 in the php code that should be moved to the .rodata section. I remembered
2561 to update Builducptable as well, so it won't revert if ucptable.h is ever
2562 re-created.
2563
2564 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2565 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2566 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2567 support is not required. This saves 1.5-2K of code, which is important in
2568 some applications.
2569
2570 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2571 so as not to refer to the tables, even though these functions will never be
2572 called when UTF-8 support is disabled. Otherwise there are problems with a
2573 shared library.
2574
2575 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2576
2577 (a) It was defining its arguments as char * instead of void *.
2578
2579 (b) It was assuming that all moves were upwards in memory; this was true
2580 a long time ago when I wrote it, but is no longer the case.
2581
2582 The emulated memove() is provided for those environments that have neither
2583 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2584 is clearly not the case, as these two bugs were recently reported.
2585
2586 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2587 and Detrail to create the HTML documentation, the .txt form of the man
2588 pages, and it removes trailing spaces from listed files. It also creates
2589 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2590 case, it wraps all the #defines with #ifndefs. This script should be run
2591 before "make dist".
2592
2593 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2594 with Unicode property support.
2595
2596 (a) For a maximizing quantifier, if the two different cases of the
2597 character were of different lengths in their UTF-8 codings (there are
2598 some cases like this - I found 11), and the matching function had to
2599 back up over a mixture of the two cases, it incorrectly assumed they
2600 were both the same length.
2601
2602 (b) When PCRE was configured to use the heap rather than the stack for
2603 recursion during matching, it was not correctly preserving the data for
2604 the other case of a UTF-8 character when checking ahead for a match
2605 while processing a minimizing repeat. If the check also involved
2606 matching a wide character, but failed, corruption could cause an
2607 erroneous result when trying to check for a repeat of the original
2608 character.
2609
2610 13. Some tidying changes to the testing mechanism:
2611
2612 (a) The RunTest script now detects the internal link size and whether there
2613 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2614 values substituted by "configure". (The RunGrepTest script already did
2615 this for UTF-8.) The configure.ac script no longer substitutes the
2616 relevant variables.
2617
2618 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2619 with length and offset values. This means that the output is different
2620 for different internal link sizes. Test 2 is skipped for link sizes
2621 other than 2 because of this, bypassing the problem. Unfortunately,
2622 there was also a test in test 3 (the locale tests) that used /B and
2623 failed for link sizes other than 2. Rather than cut the whole test out,
2624 I have added a new /Z option to pcretest that replaces the length and
2625 offset values with spaces. This is now used to make test 3 independent
2626 of link size. (Test 2 will be tidied up later.)
2627
2628 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2629 segmentation fault instead of returning the appropriate error message.
2630
2631 15. In multiline mode when the newline sequence was set to "any", the pattern
2632 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2633 This doesn't seem right; it now treats the CRLF combination as the line
2634 ending, and so does not match in that case. It's only a pattern such as ^$
2635 that would hit this one: something like ^ABC$ would have failed after \r
2636 and then tried again after \r\n.
2637
2638 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2639 in an attempt to make files that differ only in their line terminators
2640 compare equal. This works on Linux.
2641
2642 17. Under certain error circumstances pcregrep might try to free random memory
2643 as it exited. This is now fixed, thanks to valgrind.
2644
2645 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2646 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2647 was because its rules for how to advance for /g after matching an empty
2648 string at the end of a line did not allow for this case. They now check for
2649 it specially.
2650
2651 20. pcretest is supposed to handle patterns and data of any length, by
2652 extending its buffers when necessary. It was getting this wrong when the
2653 buffer for a data line had to be extended.
2654
2655 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2656 CRLF as a newline sequence.
2657
2658 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2659 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2660 I have nevertheless tidied it up.
2661
2662 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2663
2664 24. Added a man page for pcre-config.
2665
2666
2667 Version 7.0 19-Dec-06
2668 ---------------------
2669
2670 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2671 moving to gcc 4.1.1.
2672
2673 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2674 sys/time.h, which is documented as needed for this function. It doesn't
2675 seem to matter on Linux, but it showed up on some releases of OS X.
2676
2677 3. It seems that there are systems where bytes whose values are greater than
2678 127 match isprint() in the "C" locale. The "C" locale should be the
2679 default when a C program starts up. In most systems, only ASCII printing
2680 characters match isprint(). This difference caused the output from pcretest
2681 to vary, making some of the tests fail. I have changed pcretest so that:
2682
2683 (a) When it is outputting text in the compiled version of a pattern, bytes
2684 other than 32-126 are always shown as hex escapes.
2685
2686 (b) When it is outputting text that is a matched part of a subject string,
2687 it does the same, unless a different locale has been set for the match
2688 (using the /L modifier). In this case, it uses isprint() to decide.
2689
2690 4. Fixed a major bug that caused incorrect computation of the amount of memory
2691 required for a compiled pattern when options that changed within the
2692 pattern affected the logic of the preliminary scan that determines the
2693 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2694 that the computed length was too small. The symptoms of this bug were
2695 either the PCRE error "internal error: code overflow" from pcre_compile(),
2696 or a glibc crash with a message such as "pcretest: free(): invalid next
2697 size (fast)". Examples of patterns that provoked this bug (shown in
2698 pcretest format) are:
2699
2700 /(?-x: )/x
2701 /(?x)(?-x: \s*#\s*)/
2702 /((?i)[\x{c0}])/8
2703 /(?i:[\x{c0}])/8
2704
2705 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2706 is now done differently.
2707
2708 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2709 wrapper classes; (b) implement a new function in the C++ scanner that is
2710 more efficient than the old way of doing things because it avoids levels of
2711 recursion in the regex matching; (c) add a paragraph to the documentation
2712 for the FullMatch() function.
2713
2714 6. The escape sequence \n was being treated as whatever was defined as
2715 "newline". Not only was this contrary to the documentation, which states
2716 that \n is character 10 (hex 0A), but it also went horribly wrong when
2717 "newline" was defined as CRLF. This has been fixed.
2718
2719 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2720 was being set to -1 for the "end of line" case (supposedly a value that no
2721 character can have). Though this value is never used (the check for end of
2722 line is "zero bytes in current character"), it caused compiler complaints.
2723 I've changed it to 0xffffffff.
2724
2725 8. In pcre_version.c, the version string was being built by a sequence of
2726 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2727 string (as it is for production releases) called a macro with an empty
2728 argument. The C standard says the result of this is undefined. The gcc
2729 compiler treats it as an empty string (which was what was wanted) but it is
2730 reported that Visual C gives an error. The source has been hacked around to
2731 avoid this problem.
2732
2733 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2734 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2735 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2736 of them did).
2737
2738 10. Originally, pcretest opened its input and output without "b"; then I was
2739 told that "b" was needed in some environments, so it was added for release
2740 5.0 to both the input and output. (It makes no difference on Unix-like
2741 systems.) Later I was told that it is wrong for the input on Windows. I've
2742 now abstracted the modes into two macros, to make it easier to fiddle with
2743 them, and removed "b" from the input mode under Windows.
2744
2745 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2746
2747 12. Added -help and --help to pcretest as an official way of being reminded
2748 of the options.
2749
2750 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2751 and pcrecpp.cc because they annoy compilers at high warning levels.
2752
2753 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2754
2755 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2756 scripts are not C programs :-) and which was not noticed because it works
2757 on Linux.
2758
2759 16. pcretest is supposed to handle any length of pattern and data line (as one
2760 line or as a continued sequence of lines) by extending its input buffer if
2761 necessary. This feature was broken for very long pattern lines, leading to
2762 a string of junk being passed to pcre_compile() if the pattern was longer
2763 than about 50K.
2764
2765 17. I have done a major re-factoring of the way pcre_compile() computes the
2766 amount of memory needed for a compiled pattern. Previously, there was code
2767 that made a preliminary scan of the pattern in order to do this. That was
2768 OK when PCRE was new, but as the facilities have expanded, it has become
2769 harder and harder to keep it in step with the real compile phase, and there
2770 have been a number of bugs (see for example, 4 above). I have now found a
2771 cunning way of running the real compile function in a "fake" mode that
2772 enables it to compute how much memory it would need, while actually only
2773 ever using a few hundred bytes of working memory and without too many
2774 tests of the mode. This should make future maintenance and development
2775 easier. A side effect of this work is that the limit of 200 on the nesting
2776 depth of parentheses has been removed (though this was never a serious
2777 limitation, I suspect). However, there is a downside: pcre_compile() now
2778 runs more slowly than before (30% or more, depending on the pattern). I
2779 hope this isn't a big issue. There is no effect on runtime performance.
2780
2781 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2782 newline (only possible for the last line of a file) and it was a
2783 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2784
2785 19. Added additional timing features to pcretest. (1) The -tm option now times
2786 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2787 separate command line item, by a number that specifies the number of
2788 repeats to use when timing. The default is 50000; this gives better
2789 precision, but takes uncomfortably long for very large patterns.
2790
2791 20. Extended pcre_study() to be more clever in cases where a branch of a
2792 subpattern has no definite first character. For example, (a*|b*)[cd] would
2793 previously give no result from pcre_study(). Now it recognizes that the
2794 first character must be a, b, c, or d.
2795
2796 21. There was an incorrect error "recursive call could loop indefinitely" if
2797 a subpattern (or the entire pattern) that was being tested for matching an
2798 empty string contained only one non-empty item after a nested subpattern.
2799 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2800 incorrectly, because the \d was being skipped in the check.
2801
2802 22. The pcretest program now has a new pattern option /B and a command line
2803 option -b, which is equivalent to adding /B to every pattern. This causes
2804 it to show the compiled bytecode, without the additional information that
2805 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2806 is the same as /B/I).
2807
2808 23. A new optimization is now able automatically to treat some sequences such
2809 as a*b as a*+b. More specifically, if something simple (such as a character
2810 or a simple class like \d) has an unlimited quantifier, and is followed by
2811 something that cannot possibly match the quantified thing, the quantifier
2812 is automatically "possessified".
2813
2814 24. A recursive reference to a subpattern whose number was greater than 39
2815 went wrong under certain circumstances in UTF-8 mode. This bug could also
2816 have affected the operation of pcre_study().
2817
2818 25. Realized that a little bit of performance could be had by replacing
2819 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2820
2821 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2822
2823 27. Possessive quantifiers such as a++ were previously implemented by turning
2824 them into atomic groups such as ($>a+). Now they have their own opcodes,
2825 which improves performance. This includes the automatically created ones
2826 from 23 above.
2827
2828 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2829 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2830 the first matched character to be a colon. This applied both to named and
2831 numbered groups.
2832
2833 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2834
2835 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2836 building PCRE on Windows easier, so I have included it in the distribution.
2837
2838 31. There is now a check in pcretest against a ridiculously large number being
2839 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2840 loop, the loop is abandoned.
2841
2842 32. Forward references to subpatterns in conditions such as (?(2)...) where
2843 subpattern 2 is defined later cause pcre_compile() to search forwards in
2844 the pattern for the relevant set of parentheses. This search went wrong
2845 when there were unescaped parentheses in a character class, parentheses
2846 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2847
2848 33. "Subroutine" calls and backreferences were previously restricted to
2849 referencing subpatterns earlier in the regex. This restriction has now
2850 been removed.
2851
2852 34. Added a number of extra features that are going to be in Perl 5.10. On the
2853 whole, these are just syntactic alternatives for features that PCRE had
2854 previously implemented using the Python syntax or my own invention. The
2855 other formats are all retained for compatibility.
2856
2857 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2858 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2859 also .NET compatible.
2860
2861 (b) A recursion or subroutine call to a named group can now be defined as
2862 (?&name) as well as (?P>name).
2863
2864 (c) A backreference to a named group can now be defined as \k<name> or
2865 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2866 5.10, are also .NET compatible.
2867
2868 (d) A conditional reference to a named group can now use the syntax
2869 (?(<name>) or (?('name') as well as (?(name).
2870
2871 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2872 groups (named and numbered) that are never evaluated inline, but can be
2873 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2874 is always false. There may be only one alternative in such a group.
2875
2876 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2877 as the simple (?(R). The condition is true only if the most recent
2878 recursion is that of the given number or name. It does not search out
2879 through the entire recursion stack.
2880
2881 (g) The escape \gN or \g{N} has been added, where N is a positive or
2882 negative number, specifying an absolute or relative reference.
2883
2884 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2885 some "unreachable code" warnings.
2886
2887 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2888 things, this adds five new scripts.
2889
2890 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2891 There were also incompatibilities regarding the handling of \Q..\E inside
2892 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2893 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2894
2895 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2896 matches an empty string, and forcibly breaks the loop. There were bugs in
2897 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2898 against aaaa the result was just "a" rather than "aaaa", for example. Two
2899 separate and independent bugs (that affected different cases) have been
2900 fixed.
2901
2902 39. Refactored the code to abolish the use of different opcodes for small
2903 capturing bracket numbers. This is a tidy that I avoided doing when I
2904 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2905 The new approach is not only tidier, it makes it possible to reduce the
2906 memory needed to fix the previous bug (38).
2907
2908 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2909 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2910 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2911 mode.
2912
2913 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2914 report.
2915
2916 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2917 copy construction and assignment in the C++ wrapper.
2918
2919 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2920 couple of bugs that could have given wrong results in the "--newline=crlf"
2921 case.
2922
2923 44. Added a number of casts and did some reorganization of signed/unsigned int
2924 variables following suggestions from Dair Grant. Also renamed the variable
2925 "this" as "item" because it is a C++ keyword.
2926
2927 45. Arranged for dftables to add
2928
2929 #include "pcre_internal.h"
2930
2931 to pcre_chartables.c because without it, gcc 4.x may remove the array
2932 definition from the final binary if PCRE is built into a static library and
2933 dead code stripping is activated.
2934
2935 46. For an unanchored pattern, if a match attempt fails at the start of a
2936 newline sequence, and the newline setting is CRLF or ANY, and the next two
2937 characters are CRLF, advance by two characters instead of one.
2938
2939
2940 Version 6.7 04-Jul-06
2941 ---------------------
2942
2943 1. In order to handle tests when input lines are enormously long, pcretest has
2944 been re-factored so that it automatically extends its buffers when
2945 necessary. The code is crude, but this _is_ just a test program. The
2946 default size has been increased from 32K to 50K.
2947
2948 2. The code in pcre_study() was using the value of the re argument before
2949 testing it for NULL. (Of course, in any sensible call of the function, it
2950 won't be NULL.)
2951
2952 3. The memmove() emulation function in pcre_internal.h, which is used on
2953 systems that lack both memmove() and bcopy() - that is, hardly ever -
2954 was missing a "static" storage class specifier.
2955
2956 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2957 containing an extended class (one that cannot be represented by a bitmap
2958 because it contains high-valued characters or Unicode property items, e.g.
2959 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2960 pattern, but PCRE should not loop if you do not (it no longer does).
2961 [Detail: two cases were found: (a) a repeated subpattern containing an
2962 extended class; (b) a recursive reference to a subpattern that followed a
2963 previous extended class. It wasn't skipping over the extended class
2964 correctly when UTF-8 mode was not set.]
2965
2966 5. A negated single-character class was not being recognized as fixed-length
2967 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2968 compile error "lookbehind assertion is not fixed length".
2969
2970 6. The RunPerlTest auxiliary script was showing an unexpected difference
2971 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2972 write a Perl script that can interpret lines of an input file either as
2973 byte characters or as UTF-8, which is what "perltest" was being required to
2974 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2975 can't do is switch easily at run time between having the "use utf8;" pragma
2976 or not. In the end, I fudged it by using the RunPerlTest script to insert
2977 "use utf8;" explicitly for the UTF-8 tests.
2978
2979 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2980 the end of the subject string, contrary to the documentation and to what
2981 Perl does. This was true of both matching functions. Now it matches only at
2982 the start of the subject and immediately after *internal* newlines.
2983
2984 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2985 a pointer to an int instead of a pointer to an unsigned long int. This
2986 caused problems on 64-bit systems.
2987
2988 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2989 instance of the 'standard' template library not being so standard".
2990
2991 10. There was no check on the number of named subpatterns nor the maximum
2992 length of a subpattern name. The product of these values is used to compute
2993 the size of the memory block for a compiled pattern. By supplying a very
2994 long subpattern name and a large number of named subpatterns, the size
2995 computation could be caused to overflow. This is now prevented by limiting
2996 the length of names to 32 characters, and the number of named subpatterns
2997 to 10,000.
2998
2999 11. Subpatterns that are repeated with specific counts have to be replicated in
3000 the compiled pattern. The size of memory for this was computed from the
3001 length of the subpattern and the repeat count. The latter is limited to
3002 65535, but there was no limit on the former, meaning that integer overflow
3003 could in principle occur. The compiled length of a repeated subpattern is
3004 now limited to 30,000 bytes in order to prevent this.
3005
3006 12. Added the optional facility to have named substrings with the same name.
3007
3008 13. Added the ability to use a named substring as a condition, using the
3009 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3010 are numbers (not recommended). Forward references are permitted.
3011
3012 14. Added forward references in named backreferences (if you see what I mean).
3013
3014 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3015 pattern could run off the end of the subject. For example, the pattern
3016 "(?s)(.{1,5})"8 did this with the subject "ab".
3017
3018 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3019 PCRE_CASELESS was set when matching characters that were quantified with ?
3020 or *.
3021
3022 17. A character class other than a single negated character that had a minimum
3023 but no maximum quantifier - for example [ab]{6,} - was not handled
3024 correctly by pce_dfa_exec(). It would match only one character.
3025
3026 18. A valid (though odd) pattern that looked like a POSIX character
3027 class but used an invalid character after [ (for example [[,abc,]]) caused
3028 pcre_compile() to give the error "Failed: internal error: code overflow" or
3029 in some cases to crash with a glibc free() error. This could even happen if
3030 the pattern terminated after [[ but there just happened to be a sequence of
3031 letters, a binary zero, and a closing ] in the memory that followed.
3032
3033 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3034 over the years. Originally (before any Unicode support), just the bottom 8
3035 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3036 output from "man perlunicode" includes this:
3037
3038 The regular expression compiler produces polymorphic opcodes. That
3039 is, the pattern adapts to the data and automatically switches to
3040 the Unicode character scheme when presented with Unicode data--or
3041 instead uses a traditional byte scheme when presented with byte
3042 data.
3043
3044 Sadly, a wide octal escape does not cause a switch, and in a string with
3045 no other multibyte characters, these octal escapes are treated as before.
3046 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3047 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3048 Unicode string.
3049
3050 I have not perpetrated such confusion in PCRE. Up till now, it took just
3051 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3052 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3053 translate to the appropriate multibyte character.
3054
3055 29. Applied some refactoring to reduce the number of warnings from Microsoft
3056 and Borland compilers. This has included removing the fudge introduced
3057 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3058 a warning about an unused variable.
3059
3060 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3061 characters since release 4.0, because Perl (from release 5.004) does not.
3062 [Or at least, is documented not to: some releases seem to be in conflict
3063 with the documentation.] However, when a pattern was studied with
3064 pcre_study() and all its branches started with \s, PCRE still included VT
3065 as a possible starting character. Of course, this did no harm; it just
3066 caused an unnecessary match attempt.
3067
3068 22. Removed a now-redundant internal flag bit that recorded the fact that case
3069 dependency changed within the pattern. This was once needed for "required
3070 byte" processing, but is no longer used. This recovers a now-scarce options
3071 bit. Also moved the least significant internal flag bit to the most-
3072 significant bit of the word, which was not previously used (hangover from
3073 the days when it was an int rather than a uint) to free up another bit for
3074 the future.
3075
3076 23. Added support for CRLF line endings as well as CR and LF. As well as the
3077 default being selectable at build time, it can now be changed at runtime
3078 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3079 specify that it is scanning data with non-default line endings.
3080
3081 24. Changed the definition of CXXLINK to make it agree with the definition of
3082 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3083
3084 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3085 recursions. This makes a big different to stack usage for some patterns.
3086
3087 26. If a subpattern containing a named recursion or subroutine reference such
3088 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3089 the space required for the compiled pattern went wrong and gave too small a
3090 value. Depending on the environment, this could lead to "Failed: internal
3091 error: code overflow at offset 49" or "glibc detected double free or
3092 corruption" errors.
3093
3094 27. Applied patches from Google (a) to support the new newline modes and (b) to
3095 advance over multibyte UTF-8 characters in GlobalReplace.
3096
3097 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3098 difference for some implementation of PCRE in some Windows version.
3099
3100 29. Added some extra testing facilities to pcretest:
3101
3102 \q<number> in a data line sets the "match limit" value
3103 \Q<number> in a data line sets the "match recursion limt" value
3104 -S <number> sets the stack size, where <number> is in megabytes
3105
3106 The -S option isn't available for Windows.
3107
3108
3109 Version 6.6 06-Feb-06
3110 ---------------------
3111
3112 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3113 in pcreposix.h. I have copied the definition from pcre.h.
3114
3115 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3116 because pcre.h is no longer a built file.
3117
3118 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3119 not normally included in the compiled code.
3120
3121
3122 Version 6.5 01-Feb-06
3123 ---------------------
3124
3125 1. When using the partial match feature with pcre_dfa_exec(), it was not
3126 anchoring the second and subsequent partial matches at the new starting
3127 point. This could lead to incorrect results. For example, with the pattern
3128 /1234/, partially matching against "123" and then "a4" gave a match.
3129
3130 2. Changes to pcregrep:
3131
3132 (a) All non-match returns from pcre_exec() were being treated as failures
3133 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3134 error message is output. Some extra information is given for the
3135 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3136 probably the only errors that are likely to be caused by users (by
3137 specifying a regex that has nested indefinite repeats, for instance).
3138 If there are more than 20 of these errors, pcregrep is abandoned.
3139
3140 (b) A binary zero was treated as data while matching, but terminated the
3141 output line if it was written out. This has been fixed: binary zeroes
3142 are now no different to any other data bytes.
3143
3144 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3145 used to set a locale for matching. The --locale=xxxx long option has
3146 been added (no short equivalent) to specify a locale explicitly on the
3147 pcregrep command, overriding the environment variables.
3148
3149 (d) When -B was used with -n, some line numbers in the output were one less
3150 than they should have been.
3151
3152 (e) Added the -o (--only-matching) option.
3153
3154 (f) If -A or -C was used with -c (count only), some lines of context were
3155 accidentally printed for the final match.
3156
3157 (g) Added the -H (--with-filename) option.
3158
3159 (h) The combination of options -rh failed to suppress file names for files
3160 that were found from directory arguments.
3161
3162 (i) Added the -D (--devices) and -d (--directories) options.
3163
3164 (j) Added the -F (--fixed-strings) option.
3165
3166 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3167
3168 (l) Added the --colo(u)r option.
3169
3170 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3171 is not present by default.
3172
3173 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3174 items such as (?R) or (?1), when the recursion could match a number of
3175 alternatives. If it matched one of the alternatives, but subsequently,
3176 outside the recursion, there was a failure, the code tried to back up into
3177 the recursion. However, because of the way PCRE is implemented, this is not
3178 possible, and the result was an incorrect result from the match.
3179
3180 In order to prevent this happening, the specification of recursion has
3181 been changed so that all such subpatterns are automatically treated as
3182 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3183
3184 4. I had overlooked the fact that, in some locales, there are characters for
3185 which isalpha() is true but neither isupper() nor islower() are true. In
3186 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3187 and ordfeminine) are like this. This affected the treatment of \w and \W
3188 when they appeared in character classes, but not when they appeared outside
3189 a character class. The bit map for "word" characters is now created
3190 separately from the results of isalnum() instead of just taking it from the
3191 upper, lower, and digit maps. (Plus the underscore character, of course.)
3192
3193 5. The above bug also affected the handling of POSIX character classes such as
3194 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3195 permanent tables. Instead, the bit maps for such a class were previously
3196 created as the appropriate unions of the upper, lower, and digit bitmaps.
3197 Now they are created by subtraction from the [[:word:]] class, which has
3198 its own bitmap.
3199
3200 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3201 It is created by subtracting the vertical space characters (\x09, \x0a,
3202 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3203 subtraction was done in the overall bitmap for a character class, meaning
3204 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3205 be recognized. This bug has been fixed.
3206
3207 7. Patches from the folks at Google:
3208
3209 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3210 real life, but is still worth protecting against".
3211
3212 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3213 regular expressions".
3214
3215 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3216 have it.
3217
3218 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3219 "configure" and the latter not, in order to fix a problem somebody had
3220 with compiling the Arg class on HP-UX.
3221
3222 (e) Improve the error-handling of the C++ wrapper a little bit.
3223
3224 (f) New tests for checking recursion limiting.
3225
3226 8. The pcre_memmove() function, which is used only if the environment does not
3227 have a standard memmove() function (and is therefore rarely compiled),
3228 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3229 returning a result (though PCRE never actually uses the result).
3230
3231 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3232 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3233 returned instead of calling malloc() with an overflowing number that would
3234 most likely cause subsequent chaos.
3235
3236 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3237
3238 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3239 with this option is matched, the nmatch and pmatch options of regexec() are
3240 ignored.
3241
3242 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3243 provided in case anyone wants to the the POSIX interface with UTF-8
3244 strings.
3245
3246 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3247 C++ linking (needed for some HP-UX environments).
3248
3249 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3250 (unused parameter) and in the pcre_printint() function (omitted "default"
3251 switch label when the default is to do nothing).
3252
3253 15. Added some code to make it possible, when PCRE is compiled as a C++
3254 library, to replace subject pointers for pcre_exec() with a smart pointer
3255 class, thus making it possible to process discontinuous strings.
3256
3257 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3258 much the same function. They were added by different people who were trying
3259 to make PCRE easy to compile on non-Unix systems. It has been suggested
3260 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3261 for compiling on Windows systems. I have therefore replaced it with
3262 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3263 defaults to "extern" for C or "extern C" for C++, which works fine on
3264 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3265 SCOPE with something explicit in config.h. In addition:
3266
3267 (a) pcreposix.h still had just "extern" instead of either of these macros;
3268 I have replaced it with PCRE_DATA_SCOPE.
3269
3270 (b) Functions such as _pcre_xclass(), which are internal to the library,
3271 but external in the C sense, all had PCRE_EXPORT in their definitions.
3272 This is apparently wrong for the Windows case, so I have removed it.
3273 (It makes no difference on Unix-like systems.)
3274
3275 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3276 of recursive calls to match(). This is different to MATCH_LIMIT because
3277 that limits the total number of calls to match(), not all of which increase
3278 the depth of recursion. Limiting the recursion depth limits the amount of
3279 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3280 when PCRE is compiled, and changed at run time. A patch from Google adds
3281 this functionality to the C++ interface.
3282
3283 18. Changes to the handling of Unicode character properties:
3284
3285 (a) Updated the table to Unicode 4.1.0.
3286
3287 (b) Recognize characters that are not in the table as "Cn" (undefined).
3288
3289 (c) I revised the way the table is implemented to a much improved format
3290 which includes recognition of ranges. It now supports the ranges that
3291 are defined in UnicodeData.txt, and it also amalgamates other
3292 characters into ranges. This has reduced the number of entries in the
3293 table from around 16,000 to around 3,000, thus reducing its size
3294 considerably. I realized I did not need to use a tree structure after
3295 all - a binary chop search is just as efficient. Having reduced the
3296 number of entries, I extended their size from 6 bytes to 8 bytes to
3297 allow for more data.
3298
3299 (d) Added support for Unicode script names via properties such as \p{Han}.
3300
3301 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3302 matching that character.
3303
3304 20. When matching a repeated Unicode property with a minimum greater than zero,
3305 (for example \pL{2,}), PCRE could look past the end of the subject if it
3306 reached it while seeking the minimum number of characters. This could
3307 happen only if some of the characters were more than one byte long, because
3308 there is a check for at least the minimum number of bytes.
3309
3310 21. Refactored the implementation of \p and \P so as to be more general, to
3311 allow for more different types of property in future. This has changed the
3312 compiled form incompatibly. Anybody with saved compiled patterns that use
3313 \p or \P will have to recompile them.
3314
3315 22. Added "Any" and "L&" to the supported property types.
3316
3317 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3318 but give a compile time error if the value is greater than 0xff.
3319
3320 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3321 accidentally not being installed or uninstalled.
3322
3323 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3324 made were to insert the current release number. This seemed silly, because
3325 it made things harder for people building PCRE on systems that don't run
3326 "configure". I have turned pcre.h into a distributed file, no longer built
3327 by "configure", with the version identification directly included. There is
3328 no longer a pcre.h.in file.
3329
3330 However, this change necessitated a change to the pcre-config script as
3331 well. It is built from pcre-config.in, and one of the substitutions was the
3332 release number. I have updated configure.ac so that ./configure now finds
3333 the release number by grepping pcre.h.
3334
3335 26. Added the ability to run the tests under valgrind.
3336
3337
3338 Version 6.4 05-Sep-05
3339 ---------------------
3340
3341 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3342 "--" to be printed when multiple files were scanned, even when none of the
3343 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3344 consider it to be a bug, and have restored the previous behaviour.
3345
3346 2. A couple of code tidies to get rid of compiler warnings.
3347
3348 3. The pcretest program used to cheat by referring to symbols in the library
3349 whose names begin with _pcre_. These are internal symbols that are not
3350 really supposed to be visible externally, and in some environments it is
3351 possible to suppress them. The cheating is now confined to including
3352 certain files from the library's source, which is a bit cleaner.
3353
3354 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3355 file's purpose clearer.
3356
3357 5. Reorganized pcre_ucp_findchar().
3358
3359
3360 Version 6.3 15-Aug-05
3361 ---------------------
3362
3363 1. The file libpcre.pc.in did not have general read permission in the tarball.
3364
3365 2. There were some problems when building without C++ support:
3366
3367 (a) If C++ support was not built, "make install" and "make test" still
3368 tried to test it.
3369
3370 (b) There were problems when the value of CXX was explicitly set. Some
3371 changes have been made to try to fix these, and ...
3372
3373 (c) --disable-cpp can now be used to explicitly disable C++ support.
3374
3375 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3376 backslash in a target when C++ was disabled. This confuses some
3377 versions of "make", apparently. Using an intermediate variable solves
3378 this. (Same for CPP_LOBJ.)
3379
3380 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3381 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3382 necessary on certain architectures.
3383
3384 4. Added a setting of -export-symbols-regex to the link command to remove
3385 those symbols that are exported in the C sense, but actually are local
3386 within the library, and not documented. Their names all begin with
3387 "_pcre_". This is not a perfect job, because (a) we have to except some
3388 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3389 available (and never for static libraries). I have made a note to try to
3390 find a way round (a) in the future.
3391
3392
3393 Version 6.2 01-Aug-05
3394 ---------------------
3395
3396 1. There was no test for integer overflow of quantifier values. A construction
3397 such as {1111111111111111} would give undefined results. What is worse, if
3398 a minimum quantifier for a parenthesized subpattern overflowed and became
3399 negative, the calculation of the memory size went wrong. This could have
3400 led to memory overwriting.
3401
3402 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3403
3404 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3405 operating environments where this matters.
3406
3407 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3408 PCRE options from within the C++ wrapper.
3409
3410 5. Named capturing subpatterns were not being correctly counted when a pattern
3411 was compiled. This caused two problems: (a) If there were more than 100
3412 such subpatterns, the calculation of the memory needed for the whole
3413 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3414 back references of the form \12, where the number was greater than 9, were
3415 not recognized as back references, even though there were sufficient
3416 previous subpatterns.
3417
3418 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3419 versions of gcc, e.g. 2.95.4.
3420
3421
3422 Version 6.1 21-Jun-05
3423 ---------------------
3424
3425 1. There was one reference to the variable "posix" in pcretest.c that was not
3426 surrounded by "#if !defined NOPOSIX".
3427
3428 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3429 the cross-check on the old pcre_info() function, for the benefit of the
3430 cut-down version of PCRE that is currently imported into Exim.
3431
3432 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3433 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3434 patterns that start (?i) but I don't think that matters. The use of (?i) is
3435 just an example; this all applies to the other options as well.
3436
3437 4. Since libtool seems to echo the compile commands it is issuing, the output
3438 from "make" can be reduced a bit by putting "@" in front of each libtool
3439 compile command.
3440
3441 5. Patch from the folks at Google for configure.in to be a bit more thorough
3442 in checking for a suitable C++ installation before trying to compile the
3443 C++ stuff. This should fix a reported problem when a compiler was present,
3444 but no suitable headers.
3445
3446 6. The man pages all had just "PCRE" as their title. I have changed them to
3447 be the relevant file name. I have also arranged that these names are
3448 retained in the file doc/pcre.txt, which is a concatenation in text format
3449 of all the man pages except the little individual ones for each function.
3450
3451 7. The NON-UNIX-USE file had not been updated for the different set of source
3452 files that come with release 6. I also added a few comments about the C++
3453 wrapper.
3454
3455
3456 Version 6.0 07-Jun-05
3457 ---------------------
3458
3459 1. Some minor internal re-organization to help with my DFA experiments.
3460
3461 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3462 didn't matter for the library itself when fully configured, but did matter
3463 when compiling without UCP support, or within Exim, where the ucp files are
3464 not imported.
3465
3466 3. Refactoring of the library code to split up the various functions into
3467 different source modules. The addition of the new DFA matching code (see
3468 below) to a single monolithic source would have made it really too
3469 unwieldy, quite apart from causing all the code to be include in a
3470 statically linked application, when only some functions are used. This is
3471 relevant even without the DFA addition now that patterns can be compiled in
3472 one application and matched in another.
3473
3474 The downside of splitting up is that there have to be some external
3475 functions and data tables that are used internally in different modules of
3476 the library but which are not part of the API. These have all had their
3477 names changed to start with "_pcre_" so that they are unlikely to clash
3478 with other external names.
3479
3480 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3481 a different (DFA) algorithm. Although it is slower than the original
3482 function, it does have some advantages for certain types of matching
3483 problem.
3484
3485 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3486 including restarting after a partial match.
3487
3488 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3489 defined when compiling for Windows was sent to me. I have put it into the
3490 code, though I have no means of testing or verifying it.
3491
3492 7. Added the pcre_refcount() auxiliary function.
3493
3494 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3495 match before or at the first newline in the subject string. In pcretest,
3496 the /f option on a pattern can be used to set this.
3497
3498 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3499 would behave wrongly. This has been present in PCRE since release 4.0.
3500
3501 10. A number of changes to the pcregrep command:
3502
3503 (a) Refactored how -x works; insert ^(...)$ instead of setting
3504 PCRE_ANCHORED and checking the length, in preparation for adding
3505 something similar for -w.
3506
3507 (b) Added the -w (match as a word) option.
3508
3509 (c) Refactored the way lines are read and buffered so as to have more
3510 than one at a time available.
3511
3512 (d) Implemented a pcregrep test script.
3513
3514 (e) Added the -M (multiline match) option. This allows patterns to match
3515 over several lines of the subject. The buffering ensures that at least
3516 8K, or the rest of the document (whichever is the shorter) is available
3517 for matching (and similarly the previous 8K for lookbehind assertions).
3518
3519 (f) Changed the --help output so that it now says
3520
3521 -w, --word-regex(p)
3522
3523 instead of two lines, one with "regex" and the other with "regexp"
3524 because that confused at least one person since the short forms are the
3525 same. (This required a bit of code, as the output is generated
3526 automatically from a table. It wasn't just a text change.)
3527
3528 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3529 option but starts with a hyphen. Could be a pattern or a path name
3530 starting with a hyphen, for instance.
3531
3532 (h) "-" can be given as a file name to represent stdin.
3533
3534 (i) When file names are being printed, "(standard input)" is used for
3535 the standard input, for compatibility with GNU grep. Previously
3536 "<stdin>" was used.
3537
3538 (j) The option --label=xxx can be used to supply a name to be used for
3539 stdin when file names are being printed. There is no short form.
3540
3541 (k) Re-factored the options decoding logic because we are going to add
3542 two more options that take data. Such options can now be given in four
3543 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3544
3545 (l) Added the -A, -B, and -C options for requesting that lines of context
3546 around matches be printed.
3547
3548 (m) Added the -L option to print the names of files that do not contain
3549 any matching lines, that is, the complement of -l.
3550
3551 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3552 continue to scan other files.
3553
3554 (o) The -s option was incorrectly implemented. For compatibility with other
3555 greps, it now suppresses the error message for a non-existent or non-
3556 accessible file (but not the return code). There is a new option called
3557 -q that suppresses the output of matching lines, which was what -s was
3558 previously doing.
3559
3560 (p) Added --include and --exclude options to specify files for inclusion
3561 and exclusion when recursing.
3562
3563 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3564 Hopefully, it now does.
3565
3566 12. Missing cast in pcre_study().
3567
3568 13. Added an "uninstall" target to the makefile.
3569
3570 14. Replaced "extern" in the function prototypes in Makefile.in with
3571 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3572 world, but is set differently for Windows.
3573
3574 15. Added a second compiling function called pcre_compile2(). The only
3575 difference is that it has an extra argument, which is a pointer to an
3576 integer error code. When there is a compile-time failure, this is set
3577 non-zero, in addition to the error test pointer being set to point to an
3578 error message. The new argument may be NULL if no error number is required
3579 (but then you may as well call pcre_compile(), which is now just a
3580 wrapper). This facility is provided because some applications need a
3581 numeric error indication, but it has also enabled me to tidy up the way
3582 compile-time errors are handled in the POSIX wrapper.
3583
3584 16. Added VPATH=.libs to the makefile; this should help when building with one
3585 prefix path and installing with another. (Or so I'm told by someone who
3586 knows more about this stuff than I do.)
3587
3588 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3589 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3590 match everything, including newlines. This is not POSIX-compatible, but
3591 somebody wanted the feature. From pcretest it can be activated by using
3592 both the P and the s flags.
3593
3594 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3595
3596 19. libpcre.pc was being incorrectly installed as executable.
3597
3598 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3599 it now also looks for '\r' so that it will work unmodified on Windows.
3600
3601 21. Added Google's contributed C++ wrapper to the distribution.
3602
3603 22. Added some untidy missing memory free() calls in pcretest, to keep
3604 Electric Fence happy when testing.
3605
3606
3607
3608 Version 5.0 13-Sep-04
3609 ---------------------
3610
3611 1. Internal change: literal characters are no longer packed up into items
3612 containing multiple characters in a single byte-string. Each character
3613 is now matched using a separate opcode. However, there may be more than one
3614 byte in the character in UTF-8 mode.
3615
3616 2. The pcre_callout_block structure has two new fields: pattern_position and
3617 next_item_length. These contain the offset in the pattern to the next match
3618 item, and its length, respectively.
3619
3620 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3621 insertion of callouts before each pattern item. Added the /C option to
3622 pcretest to make use of this.
3623
3624 4. On the advice of a Windows user, the lines
3625
3626 #if defined(_WIN32) || defined(WIN32)
3627 _setmode( _fileno( stdout ), 0x8000 );
3628 #endif /* defined(_WIN32) || defined(WIN32) */
3629
3630 have been added to the source of pcretest. This apparently does useful
3631 magic in relation to line terminators.
3632
3633 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3634 for the benefit of those environments where the "b" makes a difference.
3635
3636 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3637 to know about it. I have put a hack into configure.in that adds in code
3638 to set GCC=yes if CC=icc. This seems to end up at a point in the
3639 generated configure script that is early enough to affect the setting of
3640 compiler options, which is what is needed, but I have no means of testing
3641 whether it really works. (The user who reported this had patched the
3642 generated configure script, which of course I cannot do.)
3643
3644 LATER: After change 22 below (new libtool files), the configure script
3645 seems to know about icc (and also ecc). Therefore, I have commented out
3646 this hack in configure.in.
3647
3648 7. Added support for pkg-config (2 patches were sent in).
3649
3650 8. Negated POSIX character classes that used a combination of internal tables
3651 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3652 [[:^ascii]]. Typically, they would match almost any characters. The other
3653 POSIX classes were not broken in this way.
3654
3655 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3656 to find the match, as PCRE was deluded into thinking that the match had to
3657 start at the start point or following a newline. The same bug applied to
3658 patterns with negative forward assertions or any backward assertions
3659 preceding ".*" at the start, unless the pattern required a fixed first
3660 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3661
3662 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3663 starting at the last subject character, bytes beyond the end of the subject
3664 string were read.
3665
3666 11. Renamed the variable "class" as "classbits" to make life easier for C++
3667 users. (Previously there was a macro definition, but it apparently wasn't
3668 enough.)
3669
3670 12. Added the new field "tables" to the extra data so that tables can be passed
3671 in at exec time, or the internal tables can be re-selected. This allows
3672 a compiled regex to be saved and re-used at a later time by a different
3673 program that might have everything at different addresses.
3674
3675 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3676 -R library as well as a -L library.
3677
3678 14. The debugging options of pcretest (-d on the command line or D on a
3679 pattern) showed incorrect output for anything following an extended class
3680 that contained multibyte characters and which was followed by a quantifier.
3681
3682 15. Added optional support for general category Unicode character properties
3683 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3684 support. It adds about 90K to the size of the library. The meanings of the
3685 inbuilt class escapes such as \d and \s have NOT been changed.
3686
3687 16. Updated pcredemo.c to include calls to free() to release the memory for the
3688 compiled pattern.
3689
3690 17. The generated file chartables.c was being created in the source directory
3691 instead of in the building directory. This caused the build to fail if the
3692 source directory was different from the building directory, and was
3693 read-only.
3694
3695 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3696 file. No doubt somebody will tell me if they don't make sense... Also added
3697 Dan Mooney's comments about building on OpenVMS.
3698
3699 19. Added support for partial matching via the PCRE_PARTIAL option for
3700 pcre_exec() and the \P data escape in pcretest.
3701
3702 20. Extended pcretest with 3 new pattern features:
3703
3704 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3705 write the compiled pattern to the file whose name is "rest-of-line".
3706 This is a straight binary dump of the data, with the saved pointer to
3707 the character tables forced to be NULL. The study data, if any, is
3708 written too. After writing, pcretest reads a new pattern.
3709
3710 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3711 compiled pattern from the given file. There must not be any
3712 occurrences of "<" in the file name (pretty unlikely); if there are,
3713 pcretest will instead treat the initial "<" as a pattern delimiter.
3714 After reading in the pattern, pcretest goes on to read data lines as
3715 usual.
3716
3717 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3718 and 16-bit fields in a compiled pattern, to simulate a pattern that
3719 was compiled on a host of opposite endianness.
3720
3721 21. The pcre-exec() function can now cope with patterns that were compiled on
3722 hosts of opposite endianness, with this restriction:
3723
3724 As for any compiled expression that is saved and used later, the tables
3725 pointer field cannot be preserved; the extra_data field in the arguments
3726 to pcre_exec() should be used to pass in a tables address if a value
3727 other than the default internal tables were used at compile time.
3728
3729 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3730 now diagnosed as an error. Previously, most of the time, a negative number
3731 would have been treated as zero, but if in addition "ovector" was passed as
3732 NULL, a crash could occur.
3733
3734 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3735 new versions from the libtool 1.5 distribution (the last one is a copy of
3736 a file called libtool.m4). This seems to have fixed the need to patch
3737 "configure" to support Darwin 1.3 (which I used to do). However, I still
3738 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3739 workstation).
3740
3741 24. Changed the PCRE licence to be the more standard "BSD" licence.
3742
3743
3744 Version 4.5 01-Dec-03
3745 ---------------------
3746
3747 1. There has been some re-arrangement of the code for the match() function so
3748 that it can be compiled in a version that does not call itself recursively.
3749 Instead, it keeps those local variables that need separate instances for
3750 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3751 needs to "recurse". Keeping track of where control must go is done by means
3752 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3753 hide most of the details from the main code, and operates only if
3754 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3755 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3756 operating.
3757
3758 To make it easier for callers to provide specially tailored get/free
3759 functions for this usage, two new functions, pcre_stack_malloc, and
3760 pcre_stack_free, are used. They are always called in strict stacking order,
3761 and the size of block requested is always the same.
3762
3763 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3764 PCRE has been compiled to use the stack or the heap for recursion. The
3765 -C option of pcretest uses this to show which version is compiled.
3766
3767 A new data escape \S, is added to pcretest; it causes the amounts of store
3768 obtained and freed by both kinds of malloc/free at match time to be added
3769 to the output.
3770
3771 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3772 what's available on my current Linux desktop machine.
3773
3774 3. When matching a UTF-8 string, the test for a valid string at the start has
3775 been extended. If start_offset is not zero, PCRE now checks that it points
3776 to a byte that is the start of a UTF-8 character. If not, it returns
3777 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3778 this is necessary because there may be backward assertions in the pattern.
3779 When matching the same subject several times, it may save resources to use
3780 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3781
3782 4. The code for checking the validity of UTF-8 strings has been tightened so
3783 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3784 containing "overlong sequences".
3785
3786 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3787 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3788 should have been "&", but it just so happened that all the cases this let
3789 through by mistake were picked up later in the function.
3790
3791 6. I had used a variable called "isblank" - this is a C99 function, causing
3792 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3793
3794 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3795 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3796 so the version is shown; (c) stop "make test" echoing "./RunTest".
3797
3798 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3799
3800 9. The prototype for memmove() for systems that don't have it was using
3801 size_t, but the inclusion of the header that defines size_t was later. I've
3802 moved the #includes for the C headers earlier to avoid this.
3803
3804 10. Added some adjustments to the code to make it easier to compiler on certain
3805 special systems:
3806
3807 (a) Some "const" qualifiers were missing.
3808 (b) Added the macro EXPORT before all exported functions; by default this
3809 is defined to be empty.
3810 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3811 that it reads its output file name as an argument instead of writing
3812 to the standard output and assuming this can be redirected.
3813
3814 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3815 class containing characters with values greater than 255, PCRE compilation
3816 went into a loop.
3817
3818 12. A recursive reference to a subpattern that was within another subpattern
3819 that had a minimum quantifier of zero caused PCRE to crash. For example,
3820 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3821 recursion. If the recursively-called subpattern itself had a zero repeat,
3822 that was OK.
3823
3824 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3825 buffer into which it was copied (for escape processing) was still set at
3826 1024, so long lines caused crashes.
3827
3828 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3829 "internal error: code overflow...". This applied to any character class
3830 that was followed by a possessive quantifier.
3831
3832 15. Modified the Makefile to add libpcre.la as a prerequisite for
3833 libpcreposix.la because I was told this is needed for a parallel build to
3834 work.
3835
3836 16. If a pattern that contained .* following optional items at the start was
3837 studied, the wrong optimizing data was generated, leading to matching
3838 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3839 matching string must start with a or b or c. The correct conclusion for
3840 this pattern is that a match can start with any character.
3841
3842
3843 Version 4.4 13-Aug-03
3844 ---------------------
3845
3846 1. In UTF-8 mode, a character class containing characters with values between
3847 127 and 255 was not handled correctly if the compiled pattern was studied.
3848 In fixing this, I have also improved the studying algorithm for such
3849 classes (slightly).
3850
3851 2. Three internal functions had redundant arguments passed to them. Removal
3852 might give a very teeny performance improvement.
3853
3854 3. Documentation bug: the value of the capture_top field in a callout is *one
3855 more than* the number of the hightest numbered captured substring.
3856
3857 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3858 in incorrectly linking with a previously installed version. They now link
3859 explicitly with libpcre.la.
3860
3861 5. configure.in no longer needs to recognize Cygwin specially.
3862
3863 6. A problem in pcre.in for Windows platforms is fixed.
3864
3865 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3866 pcretest, it used to include the size of the study block as part of its
3867 output. Unfortunately, the structure contains a field that has a different
3868 size on different hardware architectures. This meant that the tests that
3869 showed this size failed. As the block is currently always of a fixed size,
3870 this information isn't actually particularly useful in pcretest output, so
3871 I have just removed it.
3872
3873 8. Three pre-processor statements accidentally did not start in column 1.
3874 Sadly, there are *still* compilers around that complain, even though
3875 standard C has not required this for well over a decade. Sigh.
3876
3877 9. In pcretest, the code for checking callouts passed small integers in the
3878 callout_data field, which is a void * field. However, some picky compilers
3879 complained about the casts involved for this on 64-bit systems. Now
3880 pcretest passes the address of the small integer instead, which should get
3881 rid of the warnings.
3882
3883 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3884 both compile and run time, and gives an error if an invalid UTF-8 sequence
3885 is found. There is a option for disabling this check in cases where the
3886 string is known to be correct and/or the maximum performance is wanted.
3887
3888 11. In response to a bug report, I changed one line in Makefile.in from
3889
3890 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3891 to
3892 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3893
3894 to look similar to other lines, but I have no way of telling whether this
3895 is the right thing to do, as I do not use Windows. No doubt I'll get told
3896 if it's wrong...
3897
3898
3899 Version 4.3 21-May-03
3900 ---------------------
3901
3902 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3903 Makefile.
3904
3905 2. Some refactoring to improve the quality of the code:
3906
3907 (i) The utf8_table... variables are now declared "const".
3908
3909 (ii) The code for \cx, which used the "case flipping" table to upper case
3910 lower case letters, now just substracts 32. This is ASCII-specific,
3911 but the whole concept of \cx is ASCII-specific, so it seems
3912 reasonable.
3913
3914 (iii) PCRE was using its character types table to recognize decimal and
3915 hexadecimal digits in the pattern. This is silly, because it handles
3916 only 0-9, a-f, and A-F, but the character types table is locale-
3917 specific, which means strange things might happen. A private
3918 table is now used for this - though it costs 256 bytes, a table is
3919 much faster than multiple explicit tests. Of course, the standard
3920 character types table is still used for matching digits in subject
3921 strings against \d.
3922
3923 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3924 ending in _t are). So I've renamed it as ESC_tee.
3925
3926 3. The first argument for regexec() in the POSIX wrapper should have been
3927 defined as "const".
3928
3929 4. Changed pcretest to use malloc() for its buffers so that they can be
3930 Electric Fenced for debugging.
3931
3932 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3933 to read one or more bytes before the start of the subject string. Often this
3934 had no effect on PCRE's behaviour, but in some circumstances it could
3935 provoke a segmentation fault.
3936
3937 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3938 to try to read one or more bytes before the start of the subject string.
3939
3940 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3941 UTF-8 support could misbehave in various ways if the subject string
3942 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3943 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3944 back over UTF-8 characters.)
3945
3946
3947 Version 4.2 14-Apr-03
3948 ---------------------
3949
3950 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3951
3952 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3953 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3954 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3955 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3956 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3957 and BUILD_EXEEXT
3958 Note: automatic setting of the BUILD variables is not yet working
3959 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3960 compile-time but not at link-time
3961 [LINK]: use for linking executables only
3962 make different versions for Windows and non-Windows
3963 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3964 libraries
3965 [LINK_FOR_BUILD]: new variable
3966 [OBJEXT]: use throughout
3967 [EXEEXT]: use throughout
3968 <winshared>: new target
3969 <wininstall>: new target
3970 <dftables.o>: use native compiler
3971 <dftables>: use native linker
3972 <install>: handle Windows platform correctly
3973 <clean>: ditto
3974 <check>: ditto
3975 copy DLL to top builddir before testing
3976
3977 As part of these changes, -no-undefined was removed again. This was reported
3978 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3979 in any case.
3980
3981 3. Some tidies to get rid of compiler warnings:
3982
3983 . In the match_data structure, match_limit was an unsigned long int, whereas
3984 match_call_count was an int. I've made them both unsigned long ints.
3985
3986 . In pcretest the fact that a const uschar * doesn't automatically cast to
3987 a void * provoked a warning.
3988
3989 . Turning on some more compiler warnings threw up some "shadow" variables
3990 and a few more missing casts.
3991
3992 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3993 option, a class that contained a single character with a value between 128
3994 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3995
3996 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3997 option, a class that contained several characters, but with at least one
3998 whose value was between 128 and 255 caused PCRE to crash.
3999
4000
4001 Version 4.1 12-Mar-03
4002 ---------------------
4003
4004 1. Compiling with gcc -pedantic found a couple of places where casts were
4005 needed, and a string in dftables.c that was longer than standard compilers are
4006 required to support.
4007
4008 2. Compiling with Sun's compiler found a few more places where the code could
4009 be tidied up in order to avoid warnings.
4010
4011 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4012 first of these names is deprecated in the latest Autoconf in favour of the name
4013 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4014 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4015 analogy I have changed it to CFLAGS_FOR_BUILD.
4016
4017 4. Added -no-undefined to the linking command in the Makefile, because this is
4018 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4019 linking step for the pcreposix library.
4020
4021 5. PCRE was failing to diagnose the case of two named groups with the same
4022 name.
4023
4024 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4025 literal character that is needed in the subject for a match, and scans along to
4026 ensure that it is present before embarking on the full matching process. This
4027 saves time in cases of nested unlimited repeats that are never going to match.
4028 Problem: the scan can take a lot of time if the subject is very long (e.g.
4029 megabytes), thus penalizing straightforward matches. It is now done only if the
4030 amount of subject to be scanned is less than 1000 bytes.
4031
4032 7. A lesser problem with the same optimization is that it was recording the
4033 first character of an anchored pattern as "needed", thus provoking a search
4034 right along the subject, even when the first match of the pattern was going to
4035 fail. The "needed" character is now not set for anchored patterns, unless it
4036 follows something in the pattern that is of non-fixed length. Thus, it still
4037 fulfils its original purpose of finding quick non-matches in cases of nested
4038 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4039
4040
4041 Version 4.0 17-Feb-03
4042 ---------------------
4043
4044 1. If a comment in an extended regex that started immediately after a meta-item
4045 extended to the end of string, PCRE compiled incorrect data. This could lead to
4046 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4047
4048 2. Moved to autoconf 2.53 and libtool 1.4.2.
4049
4050 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4051 the special perltest8 script is no longer needed - all the tests can be run
4052 from a single perltest script.
4053
4054 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4055 by \s. It has now been removed in PCRE. This means it isn't recognized as
4056 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4057 class [:space:] *does* include VT, thereby creating a mess.
4058
4059 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4060 space and tab.
4061
4062 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4063 its new features into the main test script, reducing the number of scripts.
4064
4065 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4066 were backward compatible, and made the (?i) apply to the whole pattern, as if
4067 /i were given. Now it behaves more logically, and applies the option setting
4068 only to what follows. PCRE has been changed to follow suit. However, if it
4069 finds options settings right at the start of the pattern, it extracts them into
4070 the global options, as before. Thus, they show up in the info data.
4071
4072 8. Added support for the \Q...\E escape sequence. Characters in between are
4073 treated as literals. This is slightly different from Perl in that $ and @ are
4074 also handled as literals inside the quotes. In Perl, they will cause variable
4075 interpolation. Note the following examples:
4076
4077 Pattern PCRE matches Perl matches
4078
4079 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4080 \Qabc\$xyz\E abc\$xyz abc\$xyz
4081 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4082
4083 For compatibility with Perl, \Q...\E sequences are recognized inside character
4084 classes as well as outside them.
4085
4086 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4087 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4088 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4089 signed/unsigned warnings.
4090
4091 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4092 option for pcretest, so I've replaced it by a simple function that does just
4093 that job.
4094
4095 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4096 "pcregrep -".
4097
4098 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4099 Java package. This provides some syntactic sugar for simple cases of what my
4100 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4101 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4102 item, you can use this simplified notation. Note that only makes sense with
4103 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4104 greediness, whatever the setting of the PCRE_UNGREEDY option.
4105
4106 13. A change of greediness default within a pattern was not taking effect at
4107 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4108 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4109 was abstracted outside.
4110
4111 14. PCRE now supports the \G assertion. It is true when the current matching
4112 position is at the start point of the match. This differs from \A when the
4113 starting offset is non-zero. Used with the /g option of pcretest (or similar
4114 code), it works in the same way as it does for Perl's /g option. If all
4115 alternatives of a regex begin with \G, the expression is anchored to the start
4116 match position, and the "anchored" flag is set in the compiled expression.
4117
4118 15. Some bugs concerning the handling of certain option changes within patterns
4119 have been fixed. These applied to options other than (?ims). For example,
4120 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4121 been the other way round. Some of this was related to change 7 above.
4122
4123 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4124 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4125 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4126 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4127
4128 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4129 mode. Unlike ".", it always matches newline, whatever the setting of
4130 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4131 assertions. Perl allows it, but it doesn't (in general) work because it can't
4132 calculate the length of the lookbehind. At least, that's the case for Perl
4133 5.8.0 - I've been told they are going to document that it doesn't work in
4134 future.
4135
4136 18. Added an error diagnosis for escapes that PCRE does not support: these are
4137 \L, \l, \N, \P, \p, \U, \u, and \X.
4138
4139 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4140 reading past the end of the pattern in cases such as /[abcd/.
4141
4142 20. PCRE was getting more memory than necessary for patterns with classes that
4143 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4144
4145 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4146 compiling PCRE for use with Virtual Pascal.
4147
4148 22. Small fix to the Makefile to make it work properly if the build is done
4149 outside the source tree.
4150
4151 23. Added a new extension: a condition to go with recursion. If a conditional
4152 subpattern starts with (?(R) the "true" branch is used if recursion has
4153 happened, whereas the "false" branch is used only at the top level.
4154
4155 24. When there was a very long string of literal characters (over 255 bytes
4156 without UTF support, over 250 bytes with UTF support), the computation of how
4157 much memory was required could be incorrect, leading to segfaults or other
4158 strange effects.
4159
4160 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4161 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4162 there was a subsequent back reference to those brackets. This meant that, for
4163 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4164 possible to check for precisely this case. All we can do is abandon the
4165 optimization if .* occurs inside capturing brackets when there are any back
4166 references whatsoever. (See below for a better fix that came later.)
4167
4168 26. The handling of the optimization for finding the first character of a
4169 non-anchored pattern, and for finding a character that is required later in the
4170 match were failing in some cases. This didn't break the matching; it just
4171 failed to optimize when it could. The way this is done has been re-implemented.
4172
4173 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4174
4175 28. Added a new feature that provides some of the functionality that Perl
4176 provides with (?{...}). The facility is termed a "callout". The way it is done
4177 in PCRE is for the caller to provide an optional function, by setting
4178 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4179 global variable. By default it is unset, which disables all calling out. To get
4180 the function called, the regex must include (?C) at appropriate points. This
4181 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4182 This provides a means of identifying different callout points. When PCRE
4183 reaches such a point in the regex, if pcre_callout has been set, the external
4184 function is called. It is provided with data in a structure called
4185 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4186 matching continues; if it returns a non-zero value, the match at the current
4187 point fails. However, backtracking will occur if possible. [This was changed
4188 later and other features added - see item 49 below.]
4189
4190 29. pcretest is upgraded to test the callout functionality. It provides a
4191 callout function that displays information. By default, it shows the start of
4192 the match and the current position in the text. There are some new data escapes
4193 to vary what happens:
4194
4195 \C+ in addition, show current contents of captured substrings
4196 \C- do not supply a callout function
4197 \C!n return 1 when callout number n is reached
4198 \C!n!m return 1 when callout number n is reached for the mth time
4199
4200 30. If pcregrep was called with the -l option and just a single file name, it
4201 output "<stdin>" if a match was found, instead of the file name.
4202
4203 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4204 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4205 pcre_exec(). This saves a malloc/free per call. The default value of
4206 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4207 when configuring.
4208
4209 32. The default maximum size of a compiled pattern is 64K. There have been a
4210 few cases of people hitting this limit. The code now uses macros to handle the
4211 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4212 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4213 configuring. Tests 2 and 5 work only with 2-byte links because they output
4214 debugging information about compiled patterns.
4215
4216 33. Internal code re-arrangements:
4217
4218 (a) Moved the debugging function for printing out a compiled regex into
4219 its own source file (printint.c) and used #include to pull it into
4220 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4221 separate copies.
4222
4223 (b) Defined the list of op-code names for debugging as a macro in
4224 internal.h so that it is next to the definition of the opcodes.
4225
4226 (c) Defined a table of op-code lengths for simpler skipping along compiled
4227 code. This is again a macro in internal.h so that it is next to the
4228 definition of the opcodes.
4229
4230 34. Added support for recursive calls to individual subpatterns, along the
4231 lines of Robin Houston's patch (but implemented somewhat differently).
4232
4233 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4234 allow it to read and process whole directories in Win32. This code was
4235 contributed by Lionel Fourquaux; it has not been tested by me.
4236
4237 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4238 used to name a group. Names consist of alphanumerics and underscores, and must
4239 be unique. Back references use the syntax (?P=name) and recursive calls use
4240 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4241 numbers. The function pcre_fullinfo() can be used after compilation to extract
4242 a name/number map. There are three relevant calls:
4243
4244 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4245 PCRE_INFO_NAMECOUNT yields the number of entries
4246 PCRE_INFO_NAMETABLE yields a pointer to the map.
4247
4248 The map is a vector of fixed-size entries. The size of each entry depends on
4249 the length of the longest name used. The first two bytes of each entry are the
4250 group number, most significant byte first. There follows the corresponding
4251 name, zero terminated. The names are in alphabetical order.
4252
4253 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4254 case instead of 255. Making it the same both with and without UTF-8 support
4255 means that the same test output works with both.
4256
4257 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4258 calling malloc() with a zero argument.
4259
4260 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4261 optimization. I've improved things by keeping a bitmap of backreferences with
4262 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4263 fact referenced, the optimization can be applied. It is unlikely that a
4264 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4265 the match to follow \n) will appear inside brackets with a number greater than
4266 31, but if it does, any back reference > 31 suppresses the optimization.
4267
4268 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4269 of disabling numbered capturing parentheses. Any opening parenthesis that is
4270 not followed by ? behaves as if it were followed by ?: but named parentheses
4271 can still be used for capturing (and they will acquire numbers in the usual
4272 way).
4273
4274 41. Redesigned the return codes from the match() function into yes/no/error so
4275 that errors can be passed back from deep inside the nested calls. A malloc
4276 failure while inside a recursive subpattern call now causes the
4277 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4278
4279 42. It is now possible to set a limit on the number of times the match()
4280 function is called in a call to pcre_exec(). This facility makes it possible to
4281 limit the amount of recursion and backtracking, though not in a directly
4282 obvious way, because the match() function is used in a number of different
4283 circumstances. The count starts from zero for each position in the subject
4284 string (for non-anchored patterns). The default limit is, for compatibility, a
4285 large number, namely 10 000 000. You can change this in two ways:
4286
4287 (a) When configuring PCRE before making, you can use --with-match-limit=n
4288 to set a default value for the compiled library.
4289
4290 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4291 a different value is set. See 45 below.
4292
4293 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4294
4295 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4296 of things that can be changed at compile time. The first argument specifies
4297 what is wanted and the second points to where the information is to be placed.
4298 The current list of available information is:
4299
4300 PCRE_CONFIG_UTF8
4301
4302 The output is an integer that is set to one if UTF-8 support is available;
4303 otherwise it is set to zero.
4304
4305 PCRE_CONFIG_NEWLINE
4306
4307 The output is an integer that it set to the value of the code that is used for
4308 newline. It is either LF (10) or CR (13).
4309
4310 PCRE_CONFIG_LINK_SIZE
4311
4312 The output is an integer that contains the number of bytes used for internal
4313 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4314
4315 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4316
4317 The output is an integer that contains the threshold above which the POSIX
4318 interface uses malloc() for output vectors. See item 31 above.
4319
4320 PCRE_CONFIG_MATCH_LIMIT
4321
4322 The output is an unsigned integer that contains the default limit of the number
4323 of match() calls in a pcre_exec() execution. See 42 above.
4324
4325 44. pcretest has been upgraded by the addition of the -C option. This causes it
4326 to extract all the available output from the new pcre_config() function, and to
4327 output it. The program then exits immediately.
4328
4329 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4330 order to support additional features. One way would have been to define
4331 pcre_exec2() (for example) with extra arguments, but this would not have been
4332 extensible, and would also have required all calls to the original function to
4333 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4334 is used for passing in "extra" data from pcre_study().
4335
4336 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4337 contains the following fields:
4338
4339 flags a bitmap indicating which of the following fields are set
4340 study_data opaque data from pcre_study()
4341 match_limit a way of specifying a limit on match() calls for a specific
4342 call to pcre_exec()
4343 callout_data data for callouts (see 49 below)
4344
4345 The flag bits are also defined in pcre.h, and are
4346
4347 PCRE_EXTRA_STUDY_DATA
4348 PCRE_EXTRA_MATCH_LIMIT
4349 PCRE_EXTRA_CALLOUT_DATA
4350
4351 The pcre_study() function now returns one of these new pcre_extra blocks, with
4352 the actual study data pointed to by the study_data field, and the
4353 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4354 before. That is, this change is entirely upwards-compatible and requires no
4355 change to existing code.
4356
4357 If you want to pass in additional data to pcre_exec(), you can either place it
4358 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4359 block.
4360
4361 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4362 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4363 times with different match limits, until it finds the minimum value needed for
4364 pcre_exec() to complete. The value is then output. This can be instructive; for
4365 most simple matches the number is quite small, but for pathological cases it
4366 gets very large very quickly.
4367
4368 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4369 returns the size of the data block pointed to by the study_data field in a
4370 pcre_extra block, that is, the value that was passed as the argument to
4371 pcre_malloc() when PCRE was getting memory in which to place the information
4372 created by pcre_study(). The fourth argument should point to a size_t variable.
4373 pcretest has been extended so that this information is shown after a successful
4374 pcre_study() call when information about the compiled regex is being displayed.
4375
4376 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4377 because what follows is always an absolute path. (Later: it turns out that this
4378 is more than cosmetic for MinGW, because it doesn't like empty path
4379 components.)
4380
4381 49. Some changes have been made to the callout feature (see 28 above):
4382
4383 (i) A callout function now has three choices for what it returns:
4384
4385 0 => success, carry on matching
4386 > 0 => failure at this point, but backtrack if possible
4387 < 0 => serious error, return this value from pcre_exec()
4388
4389 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4390 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4391 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4392 use by callout functions. It will never be used by PCRE itself.
4393
4394 (ii) The pcre_extra structure (see 45 above) has a void * field called
4395 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4396 pcre_callout_block structure has a field of the same name. The contents of
4397 the field passed in the pcre_extra structure are passed to the callout
4398 function in the corresponding field in the callout block. This makes it
4399 easier to use the same callout-containing regex from multiple threads. For
4400 testing, the pcretest program has a new data escape
4401
4402 \C*n pass the number n (may be negative) as callout_data
4403
4404 If the callout function in pcretest receives a non-zero value as
4405 callout_data, it returns that value.
4406
4407 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4408 there were some redundant $(CFLAGS) in commands that are now specified as
4409 $(LINK), which already includes $(CFLAGS).
4410
4411 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4412 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4413 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4414 one-byte characters throughout. Note that case-insensitive matching applies
4415 only to characters whose values are less than 256. PCRE doesn't support the
4416 notion of cases for higher-valued characters.
4417
4418 (i) A character class whose characters are all within 0-255 is handled as
4419 a bit map, and the map is inverted for negative classes. Previously, a
4420 character > 255 always failed to match such a class; however it should
4421 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4422
4423 (ii) A negated character class with a single character < 255 is coded as
4424 "not this character" (OP_NOT). This wasn't working properly when the test
4425 character was multibyte, either singly or repeated.
4426
4427 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4428 mode, for example: \x{100}{2,3}.
4429
4430 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4431 singly or repeated) now correctly test multibyte characters. However,
4432 PCRE doesn't recognize any characters with values greater than 255 as
4433 digits, spaces, or word characters. Such characters always match \D, \S,
4434 and \W, and never match \d, \s, or \w.
4435
4436 (v) Classes may now contain characters and character ranges with values
4437 greater than 255. For example: [ab\x{100}-\x{400}].
4438
4439 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4440 PCRE in UTF-8 mode.
4441
4442 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4443 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4444 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4445 value.)
4446
4447 53. The single man page has become too large. I have therefore split it up into
4448 a number of separate man pages. These also give rise to individual HTML pages;
4449 these are now put in a separate directory, and there is an index.html page that
4450 lists them all. Some hyperlinking between the pages has been installed.
4451
4452 54. Added convenience functions for handling named capturing parentheses.
4453
4454 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4455 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4456 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4457 are faulted.
4458
4459 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4460 calling configure. These values are used when compiling the dftables.c program
4461 which is run to generate the source of the default character tables. They
4462 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4463 you will need to set these values.
4464
4465 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4466
4467
4468 Version 3.9 02-Jan-02
4469 ---------------------
4470
4471 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4472
4473 2. If --disable-static was given, the building process failed when trying to
4474 build pcretest and pcregrep. (For some reason it was using libtool to compile
4475 them, which is not right, as they aren't part of the library.)
4476
4477
4478 Version 3.8 18-Dec-01
4479 ---------------------
4480
4481 1. The experimental UTF-8 code was completely screwed up. It was packing the
4482 bytes in the wrong order. How dumb can you get?
4483
4484
4485 Version 3.7 29-Oct-01
4486 ---------------------
4487
4488 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4489 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4490 this didn't happen under Solaris 8, where I normally test things.
4491
4492 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4493 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4494 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4495 this fix an hour or so after the initial 3.7 release.)
4496
4497
4498 Version 3.6 23-Oct-01
4499 ---------------------
4500
4501 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4502 offsets passed as NULL with zero offset count.
4503
4504 2. The config.guess and config.sub files had not been updated when I moved to
4505 the latest autoconf.
4506
4507
4508 Version 3.5 15-Aug-01
4509 ---------------------
4510
4511 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4512 had been forgotten.
4513
4514 2. By using declared but undefined structures, we can avoid using "void"
4515 definitions in pcre.h while keeping the internal definitions of the structures
4516 private.
4517
4518 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4519 user point of view, this means that both static and shared libraries are built
4520 by default, but this can be individually controlled. More of the work of
4521 handling this static/shared cases is now inside libtool instead of PCRE's make
4522 file.
4523
4524 4. The pcretest utility is now installed along with pcregrep because it is
4525 useful for users (to test regexs) and by doing this, it automatically gets
4526 relinked by libtool. The documentation has been turned into a man page, so
4527 there are now .1, .txt, and .html versions in /doc.
4528
4529 5. Upgrades to pcregrep:
4530 (i) Added long-form option names like gnu grep.
4531 (ii) Added --help to list all options with an explanatory phrase.
4532 (iii) Added -r, --recursive to recurse into sub-directories.
4533 (iv) Added -f, --file to read patterns from a file.
4534
4535 6. pcre_exec() was referring to its "code" argument before testing that
4536 argument for NULL (and giving an error if it was NULL).
4537
4538 7. Upgraded Makefile.in to allow for compiling in a different directory from
4539 the source directory.
4540
4541 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4542 options bits, the pointer it was passed was to an int instead of to an unsigned
4543 long int. This mattered only on 64-bit systems.
4544
4545 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4546 generated) instead of pcre.in, which it its source. Also made the same change
4547 in several of the .c files.
4548
4549 10. A new release of gcc defines printf() as a macro, which broke pcretest
4550 because it had an ifdef in the middle of a string argument for printf(). Fixed
4551 by using separate calls to printf().
4552
4553 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4554 script, to force use of CR or LF instead of \n in the source. On non-Unix
4555 systems, the value can be set in config.h.
4556
4557 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4558 absolute limit. Changed the text of the error message to make this clear, and
4559 likewise updated the man page.
4560
4561 13. The limit of 99 on the number of capturing subpatterns has been removed.
4562 The new limit is 65535, which I hope will not be a "real" limit.
4563
4564
4565 Version 3.4 22-Aug-00
4566 ---------------------
4567
4568 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4569
4570 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4571
4572
4573 Version 3.3 01-Aug-00
4574 ---------------------
4575
4576 1. If an octal character was given, but the value was greater than \377, it
4577 was not getting masked to the least significant bits, as documented. This could
4578 lead to crashes in some systems.
4579
4580 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4581 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4582
4583 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4584 These just pass their arguments on to (pcre_free)(), but they are provided
4585 because some uses of PCRE bind it to non-C systems that can call its functions,
4586 but cannot call free() or pcre_free() directly.
4587
4588 4. Add "make test" as a synonym for "make check". Corrected some comments in
4589 the Makefile.
4590
4591 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4592 Makefile.
4593
4594 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4595 command called pgrep for grepping around the active processes.
4596
4597 7. Added the beginnings of support for UTF-8 character strings.
4598
4599 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4600 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4601 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4602 out for the ar command.)
4603
4604
4605 Version 3.2 12-May-00
4606 ---------------------
4607
4608 This is purely a bug fixing release.
4609
4610 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4611 of ZA. This was just one example of several cases that could provoke this bug,
4612 which was introduced by change 9 of version 2.00. The code for breaking
4613 infinite loops after an iteration that matches an empty string was't working
4614 correctly.
4615
4616 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4617 when matched against abbab (for example). After matching an empty string, it
4618 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4619 caused it to match further down the string than it should.
4620
4621 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4622 was there because it doesn't seem to be needed, and it causes trouble on some
4623 systems, as it is not a Standard C header. It has been removed.
4624
4625 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4626 were reported on the Macintosh. The changes were from
4627
4628 while ((c = *(++ptr)) != 0 && c != '\n');
4629 to
4630 while ((c = *(++ptr)) != 0 && c != '\n') ;
4631
4632 Totally extraordinary, but if that's what it takes...
4633
4634 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4635 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4636 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4637 assumes the way PCRE uses memmove() (always moving upwards).
4638
4639 6. PCRE is being used in one environment where strchr() is not available. There
4640 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4641 faster code anyway.
4642
4643
4644 Version 3.1 09-Feb-00
4645 ---------------------
4646
4647 The only change in this release is the fixing of some bugs in Makefile.in for
4648 the "install" target:
4649
4650 (1) It was failing to install pcreposix.h.
4651
4652 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4653
4654
4655 Version 3.0 01-Feb-00
4656 ---------------------
4657
4658 1. Add support for the /+ modifier to perltest (to output $` like it does in
4659 pcretest).
4660
4661 2. Add support for the /g modifier to perltest.
4662
4663 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4664 matches null strings.
4665
4666 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4667 pattern. Perl treats empty patterns specially - it reuses the most recent
4668 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4669 effect.
4670
4671 5. The POSIX interface was broken in that it was just handing over the POSIX
4672 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4673 required a bigger vector, with some working space on the end. This means that
4674 the POSIX wrapper now has to get and free some memory, and copy the results.
4675
4676 6. Added some simple autoconf support, placing the test data and the
4677 documentation in separate directories, re-organizing some of the
4678 information files, and making it build pcre-config (a GNU standard). Also added
4679 libtool support for building PCRE as a shared library, which is now the
4680 default.
4681
4682 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4683 09 are not valid octal constants. Single digits will be used for minor values
4684 less than 10.
4685
4686 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4687 existing programs that set these in the POSIX interface can use PCRE without
4688 modification.
4689
4690 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4691 return all that pcre_info() returns, plus additional data. The pcre_info()
4692 function is retained for compatibility, but is considered to be obsolete.
4693
4694 10. Added experimental recursion feature (?R) to handle one common case that
4695 Perl 5.6 will be able to do with (?p{...}).
4696
4697 11. Added support for POSIX character classes like [:alpha:], which Perl is
4698 adopting.
4699
4700
4701 Version 2.08 31-Aug-99
4702 ----------------------
4703
4704 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4705 trying to match at the startoffset position, but instead was moving forward to
4706 the next newline as if a previous match had failed.
4707
4708 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4709 and could get into a loop if a null string was matched other than at the start
4710 of the subject.
4711
4712 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4713 be distinguished at compile time, and for completeness also added PCRE_DATE.
4714
4715 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4716 in GnuWin32 environments.
4717
4718
4719 Version 2.07 29-Jul-99
4720 ----------------------
4721
4722 1. The documentation is now supplied in plain text form and HTML as well as in
4723 the form of man page sources.
4724
4725 2. C++ compilers don't like assigning (void *) values to other pointer types.
4726 In particular this affects malloc(). Although there is no problem in Standard
4727 C, I've put in casts to keep C++ compilers happy.
4728
4729 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4730 should be (const char *).
4731
4732 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4733 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4734 However, I haven't made this a standard facility. The documentation doesn't
4735 mention it, and the Makefile doesn't support it.
4736
4737 5. The Makefile now contains an "install" target, with editable destinations at
4738 the top of the file. The pcretest program is not installed.
4739
4740 6. pgrep -V now gives the PCRE version number and date.
4741
4742 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4743 causing the entire string to be ignored, instead of just the last character.
4744
4745 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4746 non-matching string, it can take a very, very long time, even for strings of
4747 quite modest length, because of the nested recursion. PCRE now does better in
4748 some of these cases. It does this by remembering the last required literal
4749 character in the pattern, and pre-searching the subject to ensure it is present
4750 before running the real match. In other words, it applies a heuristic to detect
4751 some types of certain failure quickly, and in the above example, if presented
4752 with a string that has no trailing " it gives "no match" very quickly.
4753
4754 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4755 other alternatives are tried instead.
4756
4757
4758 Version 2.06 09-Jun-99
4759 ----------------------
4760
4761 1. Change pcretest's output for amount of store used to show just the code
4762 space, because the remainder (the data block) varies in size between 32-bit and
4763 64-bit systems.
4764
4765 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4766 start matching at. This allows lookbehinds to work when searching for multiple
4767 occurrences in a string.
4768
4769 3. Added additional options to pcretest for testing multiple occurrences:
4770
4771 /+ outputs the rest of the string that follows a match
4772 /g loops for multiple occurrences, using the new startoffset argument
4773 /G loops for multiple occurrences by passing an incremented pointer
4774
4775 4. PCRE wasn't doing the "first character" optimization for patterns starting
4776 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4777 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4778 the letter 'x'. On long subject strings, this gives a significant speed-up.
4779
4780
4781 Version 2.05 21-Apr-99
4782 ----------------------
4783
4784 1. Changed the type of magic_number from int to long int so that it works
4785 properly on 16-bit systems.
4786
4787 2. Fixed a bug which caused patterns starting with .* not to work correctly
4788 when the subject string contained newline characters. PCRE was assuming
4789 anchoring for such patterns in all cases, which is not correct because .* will
4790 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4791 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4792 must be retried after every newline in the subject.
4793
4794
4795 Version 2.04 18-Feb-99
4796 ----------------------
4797
4798 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4799 computation of the store needed to hold the pattern was incorrect (too large).
4800 If such patterns were nested a few deep, this could multiply and become a real
4801 problem.
4802
4803 2. Added /M option to pcretest to show the memory requirement of a specific
4804 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4805
4806 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4807 compiled in such a way that the backtracking after subsequent failure was
4808 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4809 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4810
4811
4812 Version 2.03 02-Feb-99
4813 ----------------------
4814
4815 1. Fixed typo and small mistake in man page.
4816
4817 2. Added 4th condition (GPL supersedes if conflict) and created separate
4818 LICENCE file containing the conditions.
4819
4820 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4821 Perl, that is the internal \ allows the delimiter to be included in the
4822 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4823 the final delimiter, add \ to the end of the pattern (to test the error).
4824
4825 4. Added the convenience functions for extracting substrings after a successful
4826 match. Updated pcretest to make it able to test these functions.
4827
4828
4829 Version 2.02 14-Jan-99
4830 ----------------------
4831
4832 1. Initialized the working variables associated with each extraction so that
4833 their saving and restoring doesn't refer to uninitialized store.
4834
4835 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4836 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4837 fix the problem.
4838
4839 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4840 calls, and wasn't printing the correct value for compiling calls. Increased the
4841 default value of LOOPREPEAT, and the number of significant figures in the
4842 times.
4843
4844 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4845
4846 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4847 a building problem on Windows NT with a FAT file system.
4848
4849
4850 Version 2.01 21-Oct-98
4851 ----------------------
4852
4853 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4854 to character tables built by pcre_maketables() in the current locale. If NULL
4855 is passed, the default tables are used.
4856
4857
4858 Version 2.00 24-Sep-98
4859 ----------------------
4860
4861 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4862 it any more.
4863
4864 2. Allow quantification of (?>) groups, and make it work correctly.
4865
4866 3. The first character computation wasn't working for (?>) groups.
4867
4868 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4869 end of the subject) and add 5.005's \z, which really does match only at the
4870 very end of the subject.
4871
4872 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4873
4874 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4875 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4876 localized options. All options to pcre_study() were also removed.
4877
4878 7. Add other new features from 5.005:
4879
4880 $(?<= positive lookbehind
4881 $(?<! negative lookbehind
4882 (?imsx-imsx) added the unsetting capability
4883 such a setting is global if at outer level; local otherwise
4884 (?imsx-imsx:) non-capturing groups with option setting
4885 (?(cond)re|re) conditional pattern matching
4886
4887 A backreference to itself in a repeated group matches the previous
4888 captured string.
4889
4890 8. General tidying up of studying (both automatic and via "study")
4891 consequential on the addition of new assertions.
4892
4893 9. As in 5.005, unlimited repeated groups that could match an empty substring
4894 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4895 runtime if any iteration does actually match an empty substring.
4896
4897 10. Include the RunTest script in the distribution.
4898
4899 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4900 discrepancies, some of which were old and were also with respect to 5.004. They
4901 have now been fixed.
4902
4903
4904 Version 1.09 28-Apr-98
4905 ----------------------
4906
4907 1. A negated single character class followed by a quantifier with a minimum
4908 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4909 program crashes, or just wrong answers. This did not apply to negated classes
4910 containing more than one character, or to minima other than one.
4911
4912
4913 Version 1.08 27-Mar-98
4914 ----------------------
4915
4916 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4917
4918 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4919 latter must appear before anything that relies on it in the pattern.
4920
4921
4922 Version 1.07 16-Feb-98
4923 ----------------------
4924
4925 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4926 repeat of a potentially empty string).
4927
4928
4929 Version 1.06 23-Jan-98
4930 ----------------------
4931
4932 1. Added Markus Oberhumer's little patches for C++.
4933
4934 2. Literal strings longer than 255 characters were broken.
4935
4936
4937 Version 1.05 23-Dec-97
4938 ----------------------
4939
4940 1. Negated character classes containing more than one character were failing if
4941 PCRE_CASELESS was set at run time.
4942
4943
4944 Version 1.04 19-Dec-97
4945 ----------------------
4946
4947 1. Corrected the man page, where some "const" qualifiers had been omitted.
4948
4949 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4950 input syntax.
4951
4952 3. Fixed memory leak which occurred when a regex with back references was
4953 matched with an offsets vector that wasn't big enough. The temporary memory
4954 that is used in this case wasn't being freed if the match failed.
4955
4956 4. Tidied pcretest to ensure it frees memory that it gets.
4957
4958 5. Temporary memory was being obtained in the case where the passed offsets
4959 vector was exactly big enough.
4960
4961 6. Corrected definition of offsetof() from change 5 below.
4962
4963 7. I had screwed up change 6 below and broken the rules for the use of
4964 setjmp(). Now fixed.
4965
4966
4967 Version 1.03 18-Dec-97
4968 ----------------------
4969
4970 1. A erroneous regex with a missing opening parenthesis was correctly
4971 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4972 on some systems.
4973
4974 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4975 it was reported that one broken compiler failed on the former because "code" is
4976 also an independent variable.
4977
4978 3. The erroneous regex a[]b caused an array overrun reference.
4979
4980 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4981 fail on data ending with that character. (It was going on too far, and checking
4982 the next character, typically a binary zero.) This was specific to the
4983 optimized code for single-character negative classes.
4984
4985 5. Added a contributed patch from the TIN world which does the following:
4986
4987 + Add an undef for memmove, in case the the system defines a macro for it.
4988
4989 + Add a definition of offsetof(), in case there isn't one. (I don't know
4990 the reason behind this - offsetof() is part of the ANSI standard - but
4991 it does no harm).
4992
4993 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4994 most of the places where whitespace preceded '#'. I have given up and
4995 allowed the remaining 2 cases to be at the margin.
4996
4997 + Rename some variables in pcre to eliminate shadowing. This seems very
4998 pedantic, but does no harm, of course.
4999
5000 6. Moved the call to setjmp() into its own function, to get rid of warnings
5001 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5002
5003 7. Constructs such as \d{8,} were compiling into the equivalent of
5004 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5005 outcome, but in this particular case used more store than had been allocated,
5006 which caused the bug to be discovered because it threw up an internal error.
5007
5008 8. The debugging code in both pcre and pcretest for outputting the compiled
5009 form of a regex was going wrong in the case of back references followed by
5010 curly-bracketed repeats.
5011
5012
5013 Version 1.02 12-Dec-97
5014 ----------------------
5015
5016 1. Typos in pcre.3 and comments in the source fixed.
5017
5018 2. Applied a contributed patch to get rid of places where it used to remove
5019 'const' from variables, and fixed some signed/unsigned and uninitialized
5020 variable warnings.
5021
5022 3. Added the "runtest" target to Makefile.
5023
5024 4. Set default compiler flag to -O2 rather than just -O.
5025
5026
5027 Version 1.01 19-Nov-97
5028 ----------------------
5029
5030 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5031 like /([ab]*)*/, that is, for classes with more than one character in them.
5032
5033 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5034 as /((?>a*))*/ (a PCRE_EXTRA facility).
5035
5036
5037 Version 1.00 18-Nov-97
5038 ----------------------
5039
5040 1. Added compile-time macros to support systems such as SunOS4 which don't have
5041 memmove() or strerror() but have other things that can be used instead.
5042
5043 2. Arranged that "make clean" removes the executables.
5044
5045
5046 Version 0.99 27-Oct-97
5047 ----------------------
5048
5049 1. Fixed bug in code for optimizing classes with only one character. It was
5050 initializing a 32-byte map regardless, which could cause it to run off the end
5051 of the memory it had got.
5052
5053 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5054
5055
5056 Version 0.98 22-Oct-97
5057 ----------------------
5058
5059 1. Fixed bug in code for handling temporary memory usage when there are more
5060 back references than supplied space in the ovector. This could cause segfaults.
5061
5062
5063 Version 0.97 21-Oct-97
5064 ----------------------
5065
5066 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5067
5068 2. Optimized negated single characters not to use a bit map.
5069
5070 3. Brought error texts together as macro definitions; clarified some of them;
5071 fixed one that was wrong - it said "range out of order" when it meant "invalid
5072 escape sequence".
5073
5074 4. Changed some char * arguments to const char *.
5075
5076 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5077
5078 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5079 pcretest.
5080
5081
5082 Version 0.96 16-Oct-97
5083 ----------------------
5084
5085 1. Added a simple "pgrep" utility to the distribution.
5086
5087 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5088 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5089 where "ddd" means "one or more decimal digits".
5090
5091 3. Fixed serious bug. If a pattern had a back reference, but the call to
5092 pcre_exec() didn't supply a large enough ovector to record the related
5093 identifying subpattern, the match always failed. PCRE now remembers the number
5094 of the largest back reference, and gets some temporary memory in which to save
5095 the offsets during matching if necessary, in order to ensure that
5096 backreferences always work.
5097
5098 4. Increased the compatibility with Perl in a number of ways:
5099
5100 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5101 to request this handling. The option can be set at compile or exec time.
5102
5103 (b) $ matches before a terminating newline by default; an option
5104 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5105 mode). The option can be set at compile or exec time.
5106
5107 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5108 the same as Perl's. If the decimal number it represents is less than 10
5109 or there aren't that many previous left capturing parentheses, an octal
5110 escape is read. Inside a character class, it's always an octal escape,
5111 even if it is a single digit.
5112
5113 (d) An escaped but undefined alphabetic character is taken as a literal,
5114 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5115 escapes.
5116
5117 (e) {0} is now permitted. (The previous item is removed from the compiled
5118 pattern).
5119
5120 5. Changed all the names of code files so that the basic parts are no longer
5121 than 10 characters, and abolished the teeny "globals.c" file.
5122
5123 6. Changed the handling of character classes; they are now done with a 32-byte
5124 bit map always.
5125
5126 7. Added the -d and /D options to pcretest to make it possible to look at the
5127 internals of compilation without having to recompile pcre.
5128
5129
5130 Version 0.95 23-Sep-97
5131 ----------------------
5132
5133 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5134 \x20 at the start of a run of normal characters. These were being treated as
5135 real characters, instead of the source characters being re-checked.
5136
5137
5138 Version 0.94 18-Sep-97
5139 ----------------------
5140
5141 1. The functions are now thread-safe, with the caveat that the global variables
5142 containing pointers to malloc() and free() or alternative functions are the
5143 same for all threads.
5144
5145 2. Get pcre_study() to generate a bitmap of initial characters for non-
5146 anchored patterns when this is possible, and use it if passed to pcre_exec().
5147
5148
5149 Version 0.93 15-Sep-97
5150 ----------------------
5151
5152 1. /(b)|(:+)/ was computing an incorrect first character.
5153
5154 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5155 but not actually doing anything yet.
5156
5157 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5158 as Perl does (e.g. [-az] or [az-]).
5159
5160 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5161 all possible positions.
5162
5163 5. Split up into different modules to avoid including unneeded functions in a
5164 compiled binary. However, compile and exec are still in one module. The "study"
5165 function is split off.
5166
5167 6. The character tables are now in a separate module whose source is generated
5168 by an auxiliary program - but can then be edited by hand if required. There are
5169 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5170 toupper() in the code.
5171
5172 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5173 make them global. Abolish the function for setting them, as the caller can now
5174 set them directly.
5175
5176
5177 Version 0.92 11-Sep-97
5178 ----------------------
5179
5180 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5181 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5182
5183 2. Caseless matching was not working in character classes if the characters in
5184 the pattern were in upper case.
5185
5186 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5187
5188 4. Make PCRE_ANCHORED public and accept as a compile option.
5189
5190 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5191 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5192 pass them.
5193
5194 6. Give an error if bad option bits passed at compile or run time.
5195
5196 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5197 pcretest to cause it to pass that flag.
5198
5199 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5200 options, and the first character, if set.
5201
5202 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5203
5204
5205 Version 0.91 10-Sep-97
5206 ----------------------
5207
5208 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5209 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5210
5211 2. PCRE was looping on encountering an indefinitely repeated back reference to
5212 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5213 Perl does - treats the match as successful.
5214
5215 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5