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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1362 - (show annotations)
Tue Oct 1 11:27:04 2013 UTC (6 years, 8 months ago) by chpe
File size: 251791 byte(s)
Error occurred while calculating annotation data.
unicode: Update to Unicode 6.3.0 release
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.34 xx-xxxx-201x
5 --------------------------
6
7 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8 executable memory. Patch inspired by Carsten Klein.
9
10 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11 this macro is never tested and has no effect, because the work to support
12 coverage involves only compiling and linking options and special targets in
13 the Makefile. The comment in config.h implied that defining the macro would
14 enable coverage support, which is totally false. There was also support for
15 setting this macro in the CMake files (my fault, I just copied it from
16 configure). SUPPORT_GCOV has now been removed.
17
18 3. Make a small performance improvement in strlen16() and strlen32() in
19 pcretest.
20
21 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23 these cases). There was also one in pcretest.c.
24
25 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
26
27 6. In UTF mode, the code for checking whether a group could match an empty
28 string (which is used for indefinitely repeated groups to allow for
29 breaking an infinite loop) was broken when the group contained a repeated
30 negated single-character class with a character that occupied more than one
31 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
32 UTF-8 mode). The effect was undefined: the group might or might not be
33 deemed as matching an empty string, or the program might have crashed.
34
35 7. The code for checking whether a group could match an empty string was not
36 recognizing that \h, \H, \v, \V, and \R must match a character.
37
38 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
39 an empty string. If it can, pcretest shows this in its information output.
40
41 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
42 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
43 matched by pcre_exec() without using JIT:
44
45 (a) If the rest of the pattern did not match after a maximal run of
46 grapheme clusters, the code for backing up to try with fewer of them
47 did not always back up over a full grapheme when characters that do not
48 have the modifier quality were involved, e.g. Hangul syllables.
49
50 (b) If the match point in a subject started with modifier character, and
51 there was no match, the code could incorrectly back up beyond the match
52 point, and potentially beyond the first character in the subject,
53 leading to a segfault or an incorrect match result.
54
55 10. A conditional group with an assertion condition could lead to PCRE
56 recording an incorrect first data item for a match if no other first data
57 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58 first data item, and therefore matched "ca" after "c" instead of at the
59 start.
60
61 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
62 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
63
64 12. The source of pcregrep now includes z/OS-specific code so that it can be
65 compiled for z/OS as part of the special z/OS distribution.
66
67 13. Added the -T and -TM options to pcretest.
68
69 14. The code in pcre_compile.c for creating the table of named capturing groups
70 has been refactored. Instead of creating the table dynamically during the
71 actual compiling pass, the information is remembered during the pre-compile
72 pass (on the stack unless there are more than 20 named groups, in which
73 case malloc() is used) and the whole table is created before the actual
74 compile happens. This has simplified the code (it is now nearly 150 lines
75 shorter) and prepared the way for better handling of references to groups
76 with duplicate names.
77
78 15. A back reference to a named subpattern when there is more than one of the
79 same name now checks them in the order in which they appear in the pattern.
80 The first one that is set is used for the reference. Previously only the
81 first one was inspected. This change makes PCRE more compatible with Perl.
82
83 16. Unicode character properties were updated from Unicode 6.3.0.
84
85 Version 8.33 28-May-2013
86 ------------------------
87
88 1. Added 'U' to some constants that are compared to unsigned integers, to
89 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
90 variables that are added to signed variables, to ensure the result is
91 signed and can be negated.
92
93 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
94 CMake config files.
95
96 3. Revise the creation of config.h.generic so that all boolean macros are
97 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
98 overriding via -D on the command line possible.
99
100 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
101 to unsigned int is reported to make a quite noticeable speed difference in
102 a specific Windows environment. Testing on Linux did also appear to show
103 some benefit (and it is clearly not harmful). Also fixed the definition of
104 Xop which should be unsigned.
105
106 5. Related to (4), changing the definition of the intermediate variable cc
107 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
108 improvements.
109
110 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
111 unnecessary spaces.
112
113 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
114 and later.
115
116 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
117
118 9. Optimizing fast_forward_start_bits in JIT.
119
120 10. Adding support for callouts in JIT, and fixing some issues revealed
121 during this work. Namely:
122
123 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
124
125 (b) Minimum length was not checked before the matching is started.
126
127 11. The value of capture_last that is passed to callouts was incorrect in some
128 cases when there was a capture on one path that was subsequently abandoned
129 after a backtrack. Also, the capture_last value is now reset after a
130 recursion, since all captures are also reset in this case.
131
132 12. The interpreter no longer returns the "too many substrings" error in the
133 case when an overflowing capture is in a branch that is subsequently
134 abandoned after a backtrack.
135
136 13. In the pathological case when an offset vector of size 2 is used, pcretest
137 now prints out the matched string after a yield of 0 or 1.
138
139 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
140 Only supported by the JIT compiler at the moment.
141
142 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
143
144 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
145 offset of the starting point of the matching process, provided the offsets
146 vector is large enough.
147
148 17. The \A escape now records a lookbehind value of 1, though its execution
149 does not actually inspect the previous character. This is to ensure that,
150 in partial multi-segment matching, at least one character from the old
151 segment is retained when a new segment is processed. Otherwise, if there
152 are no lookbehinds in the pattern, \A might match incorrectly at the start
153 of a new segment.
154
155 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
156
157 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
158 gives some modest performance improvement in 8-bit mode.
159
160 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
161 be expressed in certain programming languages using Universal Character
162 Names.
163
164 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
165 which points out that "non characters" are not "characters that may not
166 appear in Unicode strings" but rather "characters that are reserved for
167 internal use and have only local meaning".
168
169 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
170 there was a conditional group that depended on an assertion, if the
171 assertion was false, the callout that immediately followed the alternation
172 in the condition was skipped when pcre_exec() was used for matching.
173
174 23. Allow an explicit callout to be inserted before an assertion that is the
175 condition for a conditional group, for compatibility with automatic
176 callouts, which always insert a callout at this point.
177
178 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
179 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
180
181 25. (*PRUNE) is now supported by the JIT compiler.
182
183 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
184
185 27. Fix the case where there are two or more SKIPs with arguments that may be
186 ignored.
187
188 28. (*SKIP) is now supported by the JIT compiler.
189
190 29. (*THEN) is now supported by the JIT compiler.
191
192 30. Update RunTest with additional test selector options.
193
194 31. The way PCRE handles backtracking verbs has been changed in two ways.
195
196 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
197 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
198 backtracking. In some cases this makes it more Perl-compatible, but Perl's
199 rather obscure rules do not always do the same thing.
200
201 (2) Previously, backtracking verbs were confined within assertions. This is
202 no longer the case for positive assertions, except for (*ACCEPT). Again,
203 this sometimes improves Perl compatibility, and sometimes does not.
204
205 32. A number of tests that were in test 2 because Perl did things differently
206 have been moved to test 1, because either Perl or PCRE has changed, and
207 these tests are now compatible.
208
209 32. Backtracking control verbs are now handled in the same way in JIT and
210 interpreter.
211
212 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
213 contained a forward subroutine reference caused a compile error.
214
215 34. Auto-detect and optimize limited repetitions in JIT.
216
217 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
218 blocking (*UTF) etc.
219
220 36. In the interpreter, maximizing pattern repetitions for characters and
221 character types now use tail recursion, which reduces stack usage.
222
223 37. The value of the max lookbehind was not correctly preserved if a compiled
224 and saved regex was reloaded on a host of different endianness.
225
226 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
227 of the compiled pattern block, expand the flags field from 16 to 32 bits
228 because it was almost full.
229
230 39. Try madvise first before posix_madvise.
231
232 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
233 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
234 original change locked out all matches of empty strings. This has been
235 changed so that one match of an empty string per line is recognized.
236 Subsequent searches on the same line (for colouring or for --only-matching,
237 for example) do not recognize empty strings.
238
239 41. Applied a user patch to fix a number of spelling mistakes in comments.
240
241 42. Data lines longer than 65536 caused pcretest to crash.
242
243 43. Clarified the data type for length and startoffset arguments for pcre_exec
244 and pcre_dfa_exec in the function-specific man pages, where they were
245 explicitly stated to be in bytes, never having been updated. I also added
246 some clarification to the pcreapi man page.
247
248 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
249 a segmentation fault.
250
251
252 Version 8.32 30-November-2012
253 -----------------------------
254
255 1. Improved JIT compiler optimizations for first character search and single
256 character iterators.
257
258 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
259 Patch by Daniel Richard G.
260
261 3. Single character iterator optimizations in the JIT compiler.
262
263 4. Improved JIT compiler optimizations for character ranges.
264
265 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
266 Reported by Giuseppe D'Angelo.
267
268 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
269 of a line, was being set incorrectly in cases where .* appeared inside
270 atomic brackets at the start of a pattern, or where there was a subsequent
271 *PRUNE or *SKIP.
272
273 7. Improved instruction cache flush for POWER/PowerPC.
274 Patch by Daniel Richard G.
275
276 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
277 grep:
278
279 (a) There is now no limit to the number of patterns to be matched.
280
281 (b) An error is given if a pattern is too long.
282
283 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
284 are now supported.
285
286 (d) --exclude-from and --include-from (multiple use) have been added.
287
288 (e) Exclusions and inclusions now apply to all files and directories, not
289 just to those obtained from scanning a directory recursively.
290
291 (f) Multiple uses of -f and --file-list are now supported.
292
293 (g) In a Windows environment, the default for -d has been changed from
294 "read" (the GNU grep default) to "skip", because otherwise the presence
295 of a directory in the file list provokes an error.
296
297 (h) The documentation has been revised and clarified in places.
298
299 9. Improve the matching speed of capturing brackets.
300
301 10. Changed the meaning of \X so that it now matches a Unicode extended
302 grapheme cluster.
303
304 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
305 out POSIX threads when JIT support is configured.
306
307 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
308
309 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
310 structure could go wrong in environments where size_t is not the same size
311 as int.
312
313 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
314 set.
315
316 15. The EBCDIC support had decayed; later updates to the code had included
317 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
318 general tidy up of EBCDIC-related issues, and the documentation was also
319 not quite right. There is now a test that can be run on ASCII systems to
320 check some of the EBCDIC-related things (but is it not a full test).
321
322 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
323 in a small tidy to the code.
324
325 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
326
327 18. If the --only-matching (-o) option in pcregrep is specified multiple
328 times, each one causes appropriate output. For example, -o1 -o2 outputs the
329 substrings matched by the 1st and 2nd capturing parentheses. A separating
330 string can be specified by --om-separator (default empty).
331
332 19. Improving the first n character searches.
333
334 20. Turn case lists for horizontal and vertical white space into macros so that
335 they are defined only once.
336
337 21. This set of changes together give more compatible Unicode case-folding
338 behaviour for characters that have more than one other case when UCP
339 support is available.
340
341 (a) The Unicode property table now has offsets into a new table of sets of
342 three or more characters that are case-equivalent. The MultiStage2.py
343 script that generates these tables (the pcre_ucd.c file) now scans
344 CaseFolding.txt instead of UnicodeData.txt for character case
345 information.
346
347 (b) The code for adding characters or ranges of characters to a character
348 class has been abstracted into a generalized function that also handles
349 case-independence. In UTF-mode with UCP support, this uses the new data
350 to handle characters with more than one other case.
351
352 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
353 whose other case is greater than 256 are now correctly matched
354 caselessly. Previously, the high codepoint matched the low one, but not
355 vice versa.
356
357 (d) The processing of \h, \H, \v, and \ in character classes now makes use
358 of the new class addition function, using character lists defined as
359 macros alongside the case definitions of 20 above.
360
361 (e) Caseless back references now work with characters that have more than
362 one other case.
363
364 (f) General caseless matching of characters with more than one other case
365 is supported.
366
367 22. Unicode character properties were updated from Unicode 6.2.0
368
369 23. Improved CMake support under Windows. Patch by Daniel Richard G.
370
371 24. Add support for 32-bit character strings, and UTF-32
372
373 25. Major JIT compiler update (code refactoring and bugfixing).
374 Experimental Sparc 32 support is added.
375
376 26. Applied a modified version of Daniel Richard G's patch to create
377 pcre.h.generic and config.h.generic by "make" instead of in the
378 PrepareRelease script.
379
380 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
381 pcre_compile.c when checking for a zero character.
382
383 28. Introducing a native interface for JIT. Through this interface, the compiled
384 machine code can be directly executed. The purpose of this interface is to
385 provide fast pattern matching, so several sanity checks are not performed.
386 However, feature tests are still performed. The new interface provides
387 1.4x speedup compared to the old one.
388
389 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
390 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
391 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
392
393 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
394 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
395 the "old" RFC 2279). Instead, it ended up passing a negative length to
396 pcre_exec().
397
398 31. Add support for GCC's visibility feature to hide internal functions.
399
400 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
401 "unknown -C option" after outputting 0 or 1.
402
403 33. There is now support for generating a code coverage report for the test
404 suite in environments where gcc is the compiler and lcov is installed. This
405 is mainly for the benefit of the developers.
406
407 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
408 unaddressable using valgrind annotations, allowing valgrind to detect
409 invalid memory accesses. This is mainly for the benefit of the developers.
410
411 25. (*UTF) can now be used to start a pattern in any of the three libraries.
412
413 26. Give configure error if --enable-cpp but no C++ compiler found.
414
415
416 Version 8.31 06-July-2012
417 -------------------------
418
419 1. Fixing a wrong JIT test case and some compiler warnings.
420
421 2. Removed a bashism from the RunTest script.
422
423 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
424 to unsigned type, result still unsigned" that was given by an MS compiler
425 on encountering the code "-sizeof(xxx)".
426
427 4. Partial matching support is added to the JIT compiler.
428
429 5. Fixed several bugs concerned with partial matching of items that consist
430 of more than one character:
431
432 (a) /^(..)\1/ did not partially match "aba" because checking references was
433 done on an "all or nothing" basis. This also applied to repeated
434 references.
435
436 (b) \R did not give a hard partial match if \r was found at the end of the
437 subject.
438
439 (c) \X did not give a hard partial match after matching one or more
440 characters at the end of the subject.
441
442 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
443 a partial match for the string "\r".
444
445 (e) When newline was set to CRLF, the metacharacter "." did not recognize
446 a partial match for a CR character at the end of the subject string.
447
448 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
449 running pcretest, the text "(JIT)" added to the output whenever JIT is
450 actually used to run the match.
451
452 7. Individual JIT compile options can be set in pcretest by following -s+[+]
453 or /S+[+] with a digit between 1 and 7.
454
455 8. OP_NOT now supports any UTF character not just single-byte ones.
456
457 9. (*MARK) control verb is now supported by the JIT compiler.
458
459 10. The command "./RunTest list" lists the available tests without actually
460 running any of them. (Because I keep forgetting what they all are.)
461
462 11. Add PCRE_INFO_MAXLOOKBEHIND.
463
464 12. Applied a (slightly modified) user-supplied patch that improves performance
465 when the heap is used for recursion (compiled with --disable-stack-for-
466 recursion). Instead of malloc and free for each heap frame each time a
467 logical recursion happens, frames are retained on a chain and re-used where
468 possible. This sometimes gives as much as 30% improvement.
469
470 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
471 call.
472
473 14. As documented, (*COMMIT) is now confined to within a positive assertion.
474
475 15. It is now possible to link pcretest with libedit as an alternative to
476 libreadline.
477
478 16. (*COMMIT) control verb is now supported by the JIT compiler.
479
480 17. The Unicode data tables have been updated to Unicode 6.1.0.
481
482 18. Added --file-list option to pcregrep.
483
484 19. Added binary file support to pcregrep, including the -a, --binary-files,
485 -I, and --text options.
486
487 20. The madvise function is renamed for posix_madvise for QNX compatibility
488 reasons. Fixed by Giuseppe D'Angelo.
489
490 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
491
492 22. Changed the option for creating symbolic links for 16-bit man pages from
493 -s to -sf so that re-installing does not cause issues.
494
495 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
496
497 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
498 matches in certain environments (the workspace was not being correctly
499 retained). Also added to pcre_dfa_exec() a simple plausibility check on
500 some of the workspace data at the beginning of a restart.
501
502 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
503 was not doing so when it should - probably a typo introduced by SVN 528
504 (change 8.10/14).
505
506 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
507 \w+ when the character tables indicated that \x{c4} was a word character.
508 There were several related cases, all because the tests for doing a table
509 lookup were testing for characters less than 127 instead of 255.
510
511 27. If a pattern contains capturing parentheses that are not used in a match,
512 their slots in the ovector are set to -1. For those that are higher than
513 any matched groups, this happens at the end of processing. In the case when
514 there were back references that the ovector was too small to contain
515 (causing temporary malloc'd memory to be used during matching), and the
516 highest capturing number was not used, memory off the end of the ovector
517 was incorrectly being set to -1. (It was using the size of the temporary
518 memory instead of the true size.)
519
520 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
521 ovector size, it uses memory at the end of the block that it has got.
522
523 29. Check for an overlong MARK name and give an error at compile time. The
524 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
525
526 30. JIT compiler update.
527
528 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
529 Rill for the patch.
530
531 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
532
533 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
534
535 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
536 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
537 was enabled.
538
539 35. Improve JIT code generation for greedy plus quantifier.
540
541 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
542 1 to "aa" instead of to an empty string. The bug affected repeated groups
543 that could potentially match an empty string.
544
545 37. Optimizing single character iterators in JIT.
546
547 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
548 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
549 codepoints that are too big for the mode are faulted, and in a UTF mode,
550 disallowed codepoints are also faulted.
551
552 39. If PCRE was compiled with UTF support, in three places in the DFA
553 matcher there was code that should only have been obeyed in UTF mode, but
554 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
555 processing when bytes with values greater than 127 were present. In 16-bit
556 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
557 both cases the values are those that cannot be the first data item in a UTF
558 character. The three items that might have provoked this were recursions,
559 possessively repeated groups, and atomic groups.
560
561 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
562 and pcregrep, because some OS require shared objects to be explicitly
563 passed to ld, causing the link step to fail if they are not.
564
565 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
566 mode, patterns that started with \h* or \R* might be incorrectly matched.
567
568
569 Version 8.30 04-February-2012
570 -----------------------------
571
572 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
573 name is defined in ctype.h.
574
575 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
576 only in quite long subpatterns.
577
578 3. Removed the function pcre_info(), which has been obsolete and deprecated
579 since it was replaced by pcre_fullinfo() in February 2000.
580
581 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
582 match a (*MARK), and the match failed at the start of the subject, a
583 reference to memory before the start of the subject could occur. This bug
584 was introduced by fix 17 of release 8.21.
585
586 5. A reference to an unset group with zero minimum repetition was giving
587 totally wrong answers (in non-JavaScript-compatibility mode). For example,
588 /(another)?(\1?)test/ matched against "hello world test". This bug was
589 introduced in release 8.13.
590
591 6. Add support for 16-bit character strings (a large amount of work involving
592 many changes and refactorings).
593
594 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
595 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
596 from a file.
597
598 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
599 rounding is not applied in this particular case).
600
601 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
602 if they appear, or are escaped, in patterns.
603
604 10. Get rid of a number of -Wunused-but-set-variable warnings.
605
606 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
607 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
608 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
609 also returns the mark "x". This bug applied to capturing parentheses,
610 non-capturing parentheses, and atomic parentheses. It also applied to some
611 assertions.
612
613 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
614 information out of configure.ac instead of relying on pcre.h.generic, which
615 is not stored in the repository.
616
617 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
618 -lreadline.
619
620 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
621
622 15. Applied Graycode's patch to put the top-level frame on the stack rather
623 than the heap when not using the stack for recursion. This gives a
624 performance improvement in many cases when recursion is not deep.
625
626 16. Experimental code added to "pcretest -C" to output the stack frame size.
627
628
629 Version 8.21 12-Dec-2011
630 ------------------------
631
632 1. Updating the JIT compiler.
633
634 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
635 are added as well.
636
637 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
638 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
639 calling _pcre_jit_exec. Some extra comments are added.
640
641 4. (*MARK) settings inside atomic groups that do not contain any capturing
642 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
643 was introduced by change 18 for 8.20.
644
645 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
646 ECMA-262 standard.
647
648 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
649 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
650 This bug was probably introduced by change 9 of 8.13.
651
652 7. While fixing 6 above, I noticed that a number of other items were being
653 incorrectly rejected as "not fixed length". This arose partly because newer
654 opcodes had not been added to the fixed-length checking code. I have (a)
655 corrected the bug and added tests for these items, and (b) arranged for an
656 error to occur if an unknown opcode is encountered while checking for fixed
657 length instead of just assuming "not fixed length". The items that were
658 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
659 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
660 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
661
662 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
663 being incorrectly compiled and would have given unpredicatble results.
664
665 9. A possessively repeated subpattern with minimum repeat count greater than
666 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
667 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
668 the first (A) could occur when it should not.
669
670 10. Add a cast and remove a redundant test from the code.
671
672 11. JIT should use pcre_malloc/pcre_free for allocation.
673
674 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
675 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
676 is anything other than /usr, -L is still shown).
677
678 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
679
680 14. Perl does not support \N without a following name in a [] class; PCRE now
681 also gives an error.
682
683 15. If a forward reference was repeated with an upper limit of around 2000,
684 it caused the error "internal error: overran compiling workspace". The
685 maximum number of forward references (including repeats) was limited by the
686 internal workspace, and dependent on the LINK_SIZE. The code has been
687 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
688 the default depends on LINK_SIZE. There is a new upper limit (for safety)
689 of around 200,000 forward references. While doing this, I also speeded up
690 the filling in of repeated forward references.
691
692 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
693 incorrectly expecting the subject to contain another "a" after the start.
694
695 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
696 in the match, the SKIP should be ignored. This was not happening; instead
697 the SKIP was being treated as NOMATCH. For patterns such as
698 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
699 tested.
700
701 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
702 now much more compatible with Perl, in particular in cases where the result
703 is a non-match for a non-anchored pattern. For example, if
704 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
705 "m", where previously it did not return a name. A side effect of this
706 change is that for partial matches, the last encountered mark name is
707 returned, as for non matches. A number of tests that were previously not
708 Perl-compatible have been moved into the Perl-compatible test files. The
709 refactoring has had the pleasing side effect of removing one argument from
710 the match() function, thus reducing its stack requirements.
711
712 19. If the /S+ option was used in pcretest to study a pattern using JIT,
713 subsequent uses of /S (without +) incorrectly behaved like /S+.
714
715 21. Retrieve executable code size support for the JIT compiler and fixing
716 some warnings.
717
718 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
719 not work when the shorter character appeared right at the end of the
720 subject string.
721
722 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
723 systems.
724
725 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
726 output it when the /M option is used in pcretest.
727
728 25. The CheckMan script was not being included in the distribution. Also, added
729 an explicit "perl" to run Perl scripts from the PrepareRelease script
730 because this is reportedly needed in Windows.
731
732 26. If study data was being save in a file and studying had not found a set of
733 "starts with" bytes for the pattern, the data written to the file (though
734 never used) was taken from uninitialized memory and so caused valgrind to
735 complain.
736
737 27. Updated RunTest.bat as provided by Sheri Pierce.
738
739 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
740
741 29. Computation of memory usage for the table of capturing group names was
742 giving an unnecessarily large value.
743
744
745 Version 8.20 21-Oct-2011
746 ------------------------
747
748 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
749 a POSIX class. After further experiments with Perl, which convinced me that
750 Perl has bugs and confusions, a closing square bracket is no longer allowed
751 in a POSIX name. This bug also affected patterns with classes that started
752 with full stops.
753
754 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
755 captured substring, but while checking the failing first alternative,
756 substring 1 is temporarily captured. If the output vector supplied to
757 pcre_exec() was not big enough for this capture, the yield of the function
758 was still zero ("insufficient space for captured substrings"). This cannot
759 be totally fixed without adding another stack variable, which seems a lot
760 of expense for a edge case. However, I have improved the situation in cases
761 such as /(a)(b)x|abc/ matched against "abc", where the return code
762 indicates that fewer than the maximum number of slots in the ovector have
763 been set.
764
765 3. Related to (2) above: when there are more back references in a pattern than
766 slots in the output vector, pcre_exec() uses temporary memory during
767 matching, and copies in the captures as far as possible afterwards. It was
768 using the entire output vector, but this conflicts with the specification
769 that only 2/3 is used for passing back captured substrings. Now it uses
770 only the first 2/3, for compatibility. This is, of course, another edge
771 case.
772
773 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
774 main code base, and can be used by building with --enable-jit. When this is
775 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
776 runtime --no-jit option is given.
777
778 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
779 ovector, the return from the function was zero, implying that there were
780 other matches that did not fit. The correct "exactly full" value is now
781 returned.
782
783 6. If a subpattern that was called recursively or as a subroutine contained
784 (*PRUNE) or any other control that caused it to give a non-standard return,
785 invalid errors such as "Error -26 (nested recursion at the same subject
786 position)" or even infinite loops could occur.
787
788 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
789 computing the minimum length on reaching *ACCEPT, and so ended up with the
790 wrong value of 1 rather than 0. Further investigation indicates that
791 computing a minimum subject length in the presence of *ACCEPT is difficult
792 (think back references, subroutine calls), and so I have changed the code
793 so that no minimum is registered for a pattern that contains *ACCEPT.
794
795 8. If (*THEN) was present in the first (true) branch of a conditional group,
796 it was not handled as intended. [But see 16 below.]
797
798 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
799 Sheri Pierce.
800
801 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
802 the first byte in a match must be "a".
803
804 11. Change 17 for 8.13 increased the recursion depth for patterns like
805 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
806 pattern contains any instances of (*THEN). If it does not, the old
807 optimizations are restored. It would be nice to do this on a per-group
808 basis, but at the moment that is not feasible.
809
810 12. In some environments, the output of pcretest -C is CRLF terminated. This
811 broke RunTest's code that checks for the link size. A single white space
812 character after the value is now allowed for.
813
814 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
815 For "fr", it uses the Windows-specific input and output files.
816
817 14. If (*THEN) appeared in a group that was called recursively or as a
818 subroutine, it did not work as intended. [But see next item.]
819
820 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
821 pattern fragments (but not containing any | characters). If A and B are
822 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
823 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
824 D. In other words, Perl considers parentheses that do not contain any |
825 characters to be part of a surrounding alternative, whereas PCRE was
826 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
827 differently. PCRE now behaves in the same way as Perl, except in the case
828 of subroutine/recursion calls such as (?1) which have in any case always
829 been different (but PCRE had them first :-).
830
831 16. Related to 15 above: Perl does not treat the | in a conditional group as
832 creating alternatives. Such a group is treated in the same way as an
833 ordinary group without any | characters when processing (*THEN). PCRE has
834 been changed to match Perl's behaviour.
835
836 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
837 RunGrepTest script failed.
838
839 18. Change 22 for version 13 caused atomic groups to use more stack. This is
840 inevitable for groups that contain captures, but it can lead to a lot of
841 stack use in large patterns. The old behaviour has been restored for atomic
842 groups that do not contain any capturing parentheses.
843
844 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
845 suppress the check for a minimum subject length at run time. (If it was
846 given to pcre_exec() or pcre_dfa_exec() it did work.)
847
848 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
849 fail to work when decoding hex characters in data strings in EBCDIC
850 environments.
851
852 21. It appears that in at least one Mac OS environment, the isxdigit() function
853 is implemented as a macro that evaluates to its argument more than once,
854 contravening the C 90 Standard (I haven't checked a later standard). There
855 was an instance in pcretest which caused it to go wrong when processing
856 \x{...} escapes in subject strings. The has been rewritten to avoid using
857 things like p++ in the argument of isxdigit().
858
859
860 Version 8.13 16-Aug-2011
861 ------------------------
862
863 1. The Unicode data tables have been updated to Unicode 6.0.0.
864
865 2. Two minor typos in pcre_internal.h have been fixed.
866
867 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
868 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
869 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
870
871 4. There were a number of related bugs in the code for matching backrefences
872 caselessly in UTF-8 mode when codes for the characters concerned were
873 different numbers of bytes. For example, U+023A and U+2C65 are an upper
874 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
875 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
876 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
877 2-byte code at the end of the subject (it thought there wasn't enough data
878 left).
879
880 5. Comprehensive information about what went wrong is now returned by
881 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
882 as the output vector has at least 2 elements. The offset of the start of
883 the failing character and a reason code are placed in the vector.
884
885 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
886 now returned is for the first byte of the failing character, instead of the
887 last byte inspected. This is an incompatible change, but I hope it is small
888 enough not to be a problem. It makes the returned offset consistent with
889 pcre_exec() and pcre_dfa_exec().
890
891 7. pcretest now gives a text phrase as well as the error number when
892 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
893 failure, the offset and reason code are output.
894
895 8. When \R was used with a maximizing quantifier it failed to skip backwards
896 over a \r\n pair if the subsequent match failed. Instead, it just skipped
897 back over a single character (\n). This seems wrong (because it treated the
898 two characters as a single entity when going forwards), conflicts with the
899 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
900 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
901 has been changed.
902
903 9. Some internal refactoring has changed the processing so that the handling
904 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
905 time (the PCRE_DOTALL option was changed this way some time ago: version
906 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
907 which was always a bit of a fudge. It also means that there is one less
908 argument for the match() function, which reduces its stack requirements
909 slightly. This change also fixes an incompatibility with Perl: the pattern
910 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
911
912 10. More internal refactoring has drastically reduced the number of recursive
913 calls to match() for possessively repeated groups such as (abc)++ when
914 using pcre_exec().
915
916 11. While implementing 10, a number of bugs in the handling of groups were
917 discovered and fixed:
918
919 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
920 (a|)*(?1) gave a compile-time internal error.
921 ((a|)+)+ did not notice that the outer group could match an empty string.
922 (^a|^)+ was not marked as anchored.
923 (.*a|.*)+ was not marked as matching at start or after a newline.
924
925 12. Yet more internal refactoring has removed another argument from the match()
926 function. Special calls to this function are now indicated by setting a
927 value in a variable in the "match data" data block.
928
929 13. Be more explicit in pcre_study() instead of relying on "default" for
930 opcodes that mean there is no starting character; this means that when new
931 ones are added and accidentally left out of pcre_study(), testing should
932 pick them up.
933
934 14. The -s option of pcretest has been documented for ages as being an old
935 synonym of -m (show memory usage). I have changed it to mean "force study
936 for every regex", that is, assume /S for every regex. This is similar to -i
937 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
938 using it. It makes it easier to run collections of tests with and without
939 study enabled, and thereby test pcre_study() more easily. All the standard
940 tests are now run with and without -s (but some patterns can be marked as
941 "never study" - see 20 below).
942
943 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
944 restoration of the capturing data to the outer values was not happening
945 correctly.
946
947 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
948 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
949 pattern had matched an empty string, and so incorrectly returned a no
950 match.
951
952 17. There was optimizing code for the last branch of non-capturing parentheses,
953 and also for the obeyed branch of a conditional subexpression, which used
954 tail recursion to cut down on stack usage. Unfortunately, now that there is
955 the possibility of (*THEN) occurring in these branches, tail recursion is
956 no longer possible because the return has to be checked for (*THEN). These
957 two optimizations have therefore been removed. [But see 8.20/11 above.]
958
959 18. If a pattern containing \R was studied, it was assumed that \R always
960 matched two bytes, thus causing the minimum subject length to be
961 incorrectly computed because \R can also match just one byte.
962
963 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
964 was incorrectly computed.
965
966 20. If /S is present twice on a test pattern in pcretest input, it now
967 *disables* studying, thereby overriding the use of -s on the command line
968 (see 14 above). This is necessary for one or two tests to keep the output
969 identical in both cases.
970
971 21. When (*ACCEPT) was used in an assertion that matched an empty string and
972 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
973
974 22. When an atomic group that contained a capturing parenthesis was
975 successfully matched, but the branch in which it appeared failed, the
976 capturing was not being forgotten if a higher numbered group was later
977 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
978 group 1 to "a", when in fact it should be unset. This applied to multi-
979 branched capturing and non-capturing groups, repeated or not, and also to
980 positive assertions (capturing in negative assertions does not happen
981 in PCRE) and also to nested atomic groups.
982
983 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
984 subject after a captured substring, to make it easier to tell which of a
985 number of identical substrings has been captured.
986
987 24. The way atomic groups are processed by pcre_exec() has been changed so that
988 if they are repeated, backtracking one repetition now resets captured
989 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
990 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
991 "aaa". Previously, it would have been "a". As part of this code
992 refactoring, the way recursive calls are handled has also been changed.
993
994 25. If an assertion condition captured any substrings, they were not passed
995 back unless some other capturing happened later. For example, if
996 (?(?=(a))a) was matched against "a", no capturing was returned.
997
998 26. When studying a pattern that contained subroutine calls or assertions,
999 the code for finding the minimum length of a possible match was handling
1000 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1001 group 1 called group 2 while simultaneously a separate group 2 called group
1002 1). A stack overflow occurred in this case. I have fixed this by limiting
1003 the recursion depth to 10.
1004
1005 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1006 Fortmann. This supports explicit test numbers on the command line, and has
1007 argument validation and error reporting.
1008
1009 28. An instance of \X with an unlimited repeat could fail if at any point the
1010 first character it looked at was a mark character.
1011
1012 29. Some minor code refactoring concerning Unicode properties and scripts
1013 should reduce the stack requirement of match() slightly.
1014
1015 30. Added the '=' option to pcretest to check the setting of unused capturing
1016 slots at the end of the pattern, which are documented as being -1, but are
1017 not included in the return count.
1018
1019 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1020 compiled something random. Now it gives a compile-time error (as does
1021 Perl).
1022
1023 32. A *MARK encountered during the processing of a positive assertion is now
1024 recorded and passed back (compatible with Perl).
1025
1026 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1027 had alternative anchored branches, the search for a second match in a line
1028 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1029 matched the line "0102" twice. The same bug affected patterns that started
1030 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1031 twice.
1032
1033 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1034 does, and because of capturing effects, quantifying parenthesized
1035 assertions may at times be useful. Quantifiers are now allowed for
1036 parenthesized assertions.
1037
1038 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1039
1040 36. \g was being checked for fancy things in a character class, when it should
1041 just be a literal "g".
1042
1043 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1044 appearance of a nested POSIX class supersedes an apparent external class.
1045 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1046 unescaped square brackets may also appear as part of class names. For
1047 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1048 more like Perl. (But see 8.20/1 above.)
1049
1050 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1051 was because it thought it was \N{name}, which is not supported).
1052
1053 39. Add minix to OS list not supporting the -S option in pcretest.
1054
1055 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1056 cannot analyze patterns in sufficient detail to catch mutual recursions
1057 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1058 subgroup is called recursively as a subpattern for a second time at the
1059 same position in the subject string. In previous releases this might have
1060 been caught by the recursion limit, or it might have run out of stack.
1061
1062 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1063 happen only once. PCRE was, however incorrectly giving a compile time error
1064 "recursive call could loop indefinitely" because it cannot analyze the
1065 pattern in sufficient detail. The compile time test no longer happens when
1066 PCRE is compiling a conditional subpattern, but actual runaway loops are
1067 now caught at runtime (see 40 above).
1068
1069 42. It seems that Perl allows any characters other than a closing parenthesis
1070 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1071 has been changed to be the same.
1072
1073 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1074 as not to get warnings when autogen.sh is called. Also changed
1075 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1076
1077 44. To help people who use pcregrep to scan files containing exceedingly long
1078 lines, the following changes have been made:
1079
1080 (a) The default value of the buffer size parameter has been increased from
1081 8K to 20K. (The actual buffer used is three times this size.)
1082
1083 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1084 PCRE is built.
1085
1086 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1087 to be set at run time.
1088
1089 (d) Numerical values in pcregrep options can be followed by K or M, for
1090 example --buffer-size=50K.
1091
1092 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1093 given and the return code is set to 2.
1094
1095 45. Add a pointer to the latest mark to the callout data block.
1096
1097 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1098 partial match of an empty string instead of no match. This was specific to
1099 the use of ".".
1100
1101 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1102 complete match instead of a partial match. This bug was dependent on both
1103 the PCRE_UTF8 and PCRE_DOTALL options being set.
1104
1105 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1106 starting byte set, because \b was not being ignored.
1107
1108
1109 Version 8.12 15-Jan-2011
1110 ------------------------
1111
1112 1. Fixed some typos in the markup of the man pages, and wrote a script that
1113 checks for such things as part of the documentation building process.
1114
1115 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1116 --match-limit and --recursion-limit options (added for 8.11). In
1117 particular, this made one of the standard tests fail. (The integer value
1118 went into the wrong half of a long int.)
1119
1120 3. If the --colour option was given to pcregrep with -v (invert match), it
1121 did strange things, either producing crazy output, or crashing. It should,
1122 of course, ignore a request for colour when reporting lines that do not
1123 match.
1124
1125 4. Another pcregrep bug caused similar problems if --colour was specified with
1126 -M (multiline) and the pattern match finished with a line ending.
1127
1128 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1129 matched in multiline mode, the following line was shown as part of the
1130 match. This seems wrong, so I have changed it.
1131
1132 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1133 the check for further matches in the same line (so they could be coloured)
1134 to overrun the end of the current line. If another match was found, it was
1135 incorrectly shown (and then shown again when found in the next line).
1136
1137 7. If pcregrep was compiled under Windows, there was a reference to the
1138 function pcregrep_exit() before it was defined. I am assuming this was
1139 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1140 reported by a user. I've moved the definition above the reference.
1141
1142
1143 Version 8.11 10-Dec-2010
1144 ------------------------
1145
1146 1. (*THEN) was not working properly if there were untried alternatives prior
1147 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1148 backtracked to try for "b" instead of moving to the next alternative branch
1149 at the same level (in this case, to look for "c"). The Perl documentation
1150 is clear that when (*THEN) is backtracked onto, it goes to the "next
1151 alternative in the innermost enclosing group".
1152
1153 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1154 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1155 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1156 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1157 (*THEN).
1158
1159 3. If \s appeared in a character class, it removed the VT character from
1160 the class, even if it had been included by some previous item, for example
1161 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1162 of \s, but is part of the POSIX "space" class.)
1163
1164 4. A partial match never returns an empty string (because you can always
1165 match an empty string at the end of the subject); however the checking for
1166 an empty string was starting at the "start of match" point. This has been
1167 changed to the "earliest inspected character" point, because the returned
1168 data for a partial match starts at this character. This means that, for
1169 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1170 (previously it gave "no match").
1171
1172 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1173 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1174 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1175 has an implication that the given string is incomplete (because a partial
1176 match is preferred over a full match). For this reason, these items now
1177 give a partial match in this situation. [Aside: previously, the one case
1178 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1179 match rather than a full match, which was wrong by the old rules, but is
1180 now correct.]
1181
1182 6. There was a bug in the handling of #-introduced comments, recognized when
1183 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1184 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1185 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1186 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1187 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1188 places in pcre_compile().
1189
1190 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1191 comments when looking ahead for named forward references to subpatterns,
1192 the only newline sequence it recognized was NL. It now handles newlines
1193 according to the set newline convention.
1194
1195 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1196 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1197 cater for a lack of strerror(). These oversights have been fixed.
1198
1199 9. Added --match-limit and --recursion-limit to pcregrep.
1200
1201 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1202
1203 11. When the -o option was used, pcregrep was setting a return code of 1, even
1204 when matches were found, and --line-buffered was not being honoured.
1205
1206 12. Added an optional parentheses number to the -o and --only-matching options
1207 of pcregrep.
1208
1209 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1210 can match an empty string. The code to do it in pcretest and pcredemo
1211 needed fixing:
1212
1213 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1214 only one byte after an empty string match just before CRLF (this case
1215 just got forgotten; "any" and "anycrlf" were OK).
1216
1217 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1218 mode when an empty string match preceded an ASCII character followed by
1219 a non-ASCII character. (The code for advancing by one character rather
1220 than one byte was nonsense.)
1221
1222 (c) The pcredemo.c sample program did not have any code at all to handle
1223 the cases when CRLF is a valid newline sequence.
1224
1225 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1226 as a starting offset was within the subject string. There is now a new
1227 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1228 negative or greater than the length of the string. In order to test this,
1229 pcretest is extended to allow the setting of negative starting offsets.
1230
1231 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1232 starting offset points to the beginning of a UTF-8 character was
1233 unnecessarily clumsy. I tidied it up.
1234
1235 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1236 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1237
1238 17. Nobody had reported that the --include_dir option, which was added in
1239 release 7.7 should have been called --include-dir (hyphen, not underscore)
1240 for compatibility with GNU grep. I have changed it to --include-dir, but
1241 left --include_dir as an undocumented synonym, and the same for
1242 --exclude-dir, though that is not available in GNU grep, at least as of
1243 release 2.5.4.
1244
1245 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1246 characters from a string of bytes) have been redefined so as not to use
1247 loops, in order to improve performance in some environments. At the same
1248 time, I abstracted some of the common code into auxiliary macros to save
1249 repetition (this should not affect the compiled code).
1250
1251 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1252 compile-time error is now given if \c is not followed by an ASCII
1253 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1254 different, and any byte value is allowed.)
1255
1256 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1257 START_OPTIMIZE option, which is now allowed at compile time - but just
1258 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1259 to pcregrep and other applications that have no direct access to PCRE
1260 options. The new /Y option in pcretest sets this option when calling
1261 pcre_compile().
1262
1263 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1264 back references. Groups containing recursive back references were forced to
1265 be atomic by that change, but in the case of named groups, the amount of
1266 memory required was incorrectly computed, leading to "Failed: internal
1267 error: code overflow". This has been fixed.
1268
1269 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1270 pcretest.c, to avoid build problems in some Borland environments.
1271
1272
1273 Version 8.10 25-Jun-2010
1274 ------------------------
1275
1276 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1277 THEN.
1278
1279 2. (*ACCEPT) was not working when inside an atomic group.
1280
1281 3. Inside a character class, \B is treated as a literal by default, but
1282 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1283 causes the error). The code is unchanged, but I tidied the documentation.
1284
1285 4. Inside a character class, PCRE always treated \R and \X as literals,
1286 whereas Perl faults them if its -w option is set. I have changed PCRE so
1287 that it faults them when PCRE_EXTRA is set.
1288
1289 5. Added support for \N, which always matches any character other than
1290 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1291
1292 6. When compiling pcregrep with newer versions of gcc which may have
1293 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1294 declared with attribute warn_unused_result" were given. Just casting the
1295 result to (void) does not stop the warnings; a more elaborate fudge is
1296 needed. I've used a macro to implement this.
1297
1298 7. Minor change to pcretest.c to avoid a compiler warning.
1299
1300 8. Added four artifical Unicode properties to help with an option to make
1301 \s etc use properties (see next item). The new properties are: Xan
1302 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1303
1304 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1305 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1306 this option. Modified pcretest to add /W to test this facility. Added
1307 REG_UCP to make it available via the POSIX interface.
1308
1309 10. Added --line-buffered to pcregrep.
1310
1311 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1312 studied, and the match started with a letter with a code point greater than
1313 127 whose first byte was different to the first byte of the other case of
1314 the letter, the other case of this starting letter was not recognized
1315 (#976).
1316
1317 12. If a pattern that was studied started with a repeated Unicode property
1318 test, for example, \p{Nd}+, there was the theoretical possibility of
1319 setting up an incorrect bitmap of starting bytes, but fortunately it could
1320 not have actually happened in practice until change 8 above was made (it
1321 added property types that matched character-matching opcodes).
1322
1323 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1324 possible starting bytes for non-anchored patterns.
1325
1326 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1327 \R, and also a number of cases that involve Unicode properties, both
1328 explicit and implicit when PCRE_UCP is set.
1329
1330 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1331 input, it could crash or give wrong results if characters with values
1332 greater than 0xc0 were present in the subject string. (Detail: it assumed
1333 UTF-8 input when processing these items.)
1334
1335 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1336 size_t is 64-bit (#991).
1337
1338 17. Added a check for running out of memory when PCRE is compiled with
1339 --disable-stack-for-recursion (#990).
1340
1341 18. If the last data line in a file for pcretest does not have a newline on
1342 the end, a newline was missing in the output.
1343
1344 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1345 less than 128) in its various bitmaps. However, there is a facility for
1346 generating tables according to the current locale when PCRE is compiled. It
1347 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1348 characters, are recognized by isspace() and therefore were getting set in
1349 these tables, and indeed these tables seem to approximate to ISO 8859. This
1350 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1351 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1352 which of course cannot start UTF-8 characters. I have changed the code so
1353 that only real ASCII characters (less than 128) and the correct starting
1354 bytes for UTF-8 encodings are set for characters greater than 127 when in
1355 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1356 altogether.)
1357
1358 20. Added the /T option to pcretest so as to be able to run tests with non-
1359 standard character tables, thus making it possible to include the tests
1360 used for 19 above in the standard set of tests.
1361
1362 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1363 reference to a subpattern the other side of a comment that contains an
1364 opening parenthesis caused either an internal compiling error, or a
1365 reference to the wrong subpattern.
1366
1367
1368 Version 8.02 19-Mar-2010
1369 ------------------------
1370
1371 1. The Unicode data tables have been updated to Unicode 5.2.0.
1372
1373 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1374 configured.
1375
1376 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1377 original author of that file, following a query about its status.
1378
1379 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1380 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1381
1382 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1383 quantifier applied to a forward-referencing subroutine call, could compile
1384 incorrect code or give the error "internal error: previously-checked
1385 referenced subpattern not found".
1386
1387 6. Both MS Visual Studio and Symbian OS have problems with initializing
1388 variables to point to external functions. For these systems, therefore,
1389 pcre_malloc etc. are now initialized to local functions that call the
1390 relevant global functions.
1391
1392 7. There were two entries missing in the vectors called coptable and poptable
1393 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1394 I've fixed the data, and added a kludgy way of testing at compile time that
1395 the lengths are correct (equal to the number of opcodes).
1396
1397 8. Following on from 7, I added a similar kludge to check the length of the
1398 eint vector in pcreposix.c.
1399
1400 9. Error texts for pcre_compile() are held as one long string to avoid too
1401 much relocation at load time. To find a text, the string is searched,
1402 counting zeros. There was no check for running off the end of the string,
1403 which could happen if a new error number was added without updating the
1404 string.
1405
1406 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1407
1408 11. \K was not working if it appeared in an atomic group or in a group that
1409 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1410 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1411 the assertion is positive, but not if it is negative.
1412
1413 12. Change 11 fortuitously reduced the size of the stack frame used in the
1414 "match()" function of pcre_exec.c by one pointer. Forthcoming
1415 implementation of support for (*MARK) will need an extra pointer on the
1416 stack; I have reserved it now, so that the stack frame size does not
1417 decrease.
1418
1419 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1420 item in branch that calls a recursion is a subroutine call - as in the
1421 second branch in the above example - was incorrectly given the compile-
1422 time error "recursive call could loop indefinitely" because pcre_compile()
1423 was not correctly checking the subroutine for matching a non-empty string.
1424
1425 14. The checks for overrunning compiling workspace could trigger after an
1426 overrun had occurred. This is a "should never occur" error, but it can be
1427 triggered by pathological patterns such as hundreds of nested parentheses.
1428 The checks now trigger 100 bytes before the end of the workspace.
1429
1430 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1431
1432
1433 Version 8.01 19-Jan-2010
1434 ------------------------
1435
1436 1. If a pattern contained a conditional subpattern with only one branch (in
1437 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1438 computed the wrong minimum data length (which is of course zero for such
1439 subpatterns). This could cause incorrect "no match" results.
1440
1441 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1442 the pattern is reset in the first branch, pcre_compile() failed with
1443 "internal error: code overflow at offset...". This happened only when
1444 the reset was to the original external option setting. (An optimization
1445 abstracts leading options settings into an external setting, which was the
1446 cause of this.)
1447
1448 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1449 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1450 assertion pattern did not match (meaning that the assertion was true), it
1451 was incorrectly treated as false if the SKIP had been reached during the
1452 matching. This also applied to assertions used as conditions.
1453
1454 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1455 assertion subpattern, including such a pattern used as a condition,
1456 unpredictable results occurred, instead of the error return
1457 PCRE_ERROR_DFA_UITEM.
1458
1459 5. The C++ GlobalReplace function was not working like Perl for the special
1460 situation when an empty string is matched. It now does the fancy magic
1461 stuff that is necessary.
1462
1463 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1464 removed. (These were left over from very, very early versions of PCRE.)
1465
1466 7. Some cosmetic changes to the code to make life easier when compiling it
1467 as part of something else:
1468
1469 (a) Change DEBUG to PCRE_DEBUG.
1470
1471 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1472 called "current" as "current_branch", to prevent a collision with the
1473 Linux macro when compiled as a kernel module.
1474
1475 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1476 prevent a collision with the Linux macro when compiled as a kernel
1477 module.
1478
1479 8. In pcre_compile() there are some checks for integer overflows that used to
1480 cast potentially large values to (double). This has been changed to that
1481 when building, a check for int64_t is made, and if it is found, it is used
1482 instead, thus avoiding the use of floating point arithmetic. (There is no
1483 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1484 double.
1485
1486 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1487 2005 (difference between two addresses compared to an unsigned value).
1488
1489 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1490 custom one, because of the following reported problem in Windows:
1491
1492 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1493 under Win32.
1494 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1495 therefore missing the function definition.
1496 - The compiler thus generates a "C" signature for the test function.
1497 - The linker fails to find the "C" function.
1498 - PCRE fails to configure if asked to do so against libbz2.
1499
1500 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1501 messages were output:
1502
1503 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1504 rerunning libtoolize, to keep the correct libtool macros in-tree.
1505 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1506
1507 I have done both of these things.
1508
1509 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1510 most of the time, it *can* run out if it is given a pattern that contains a
1511 runaway infinite recursion. I updated the discussion in the pcrestack man
1512 page.
1513
1514 13. Now that we have gone to the x.xx style of version numbers, the minor
1515 version may start with zero. Using 08 or 09 is a bad idea because users
1516 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1517 interpreted as invalid octal numbers. I've updated the previous comment in
1518 configure.ac, and also added a check that gives an error if 08 or 09 are
1519 used.
1520
1521 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1522 causing partial matching to fail when the end of the subject matched \W
1523 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1524
1525 15. There were some discrepancies between the declarations in pcre_internal.h
1526 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1527 their definitions. The declarations used "const uschar *" and the
1528 definitions used USPTR. Even though USPTR is normally defined as "const
1529 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1530 reported that: "This difference in casting confuses some C++ compilers, for
1531 example, SunCC recognizes above declarations as different functions and
1532 generates broken code for hbpcre." I have changed the declarations to use
1533 USPTR.
1534
1535 16. GNU libtool is named differently on some systems. The autogen.sh script now
1536 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1537 (FreeBSD).
1538
1539 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1540 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1541 comment: "Figure out how to create a longlong from a string: strtoll and
1542 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1543 instance, but it only takes 2 args instead of 3!"
1544
1545 18. A subtle bug concerned with back references has been fixed by a change of
1546 specification, with a corresponding code fix. A pattern such as
1547 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1548 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1549 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1550 same bug. Such groups have to be quantified to be useful, or contained
1551 inside another quantified group. (If there's no repetition, the reference
1552 can never match.) The problem arises because, having left the group and
1553 moved on to the rest of the pattern, a later failure that backtracks into
1554 the group uses the captured value from the final iteration of the group
1555 rather than the correct earlier one. I have fixed this in PCRE by forcing
1556 any group that contains a reference to itself to be an atomic group; that
1557 is, there cannot be any backtracking into it once it has completed. This is
1558 similar to recursive and subroutine calls.
1559
1560
1561 Version 8.00 19-Oct-09
1562 ----------------------
1563
1564 1. The table for translating pcre_compile() error codes into POSIX error codes
1565 was out-of-date, and there was no check on the pcre_compile() error code
1566 being within the table. This could lead to an OK return being given in
1567 error.
1568
1569 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1570 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1571 in a Windows environment.
1572
1573 3. The pcregrep --count option prints the count for each file even when it is
1574 zero, as does GNU grep. However, pcregrep was also printing all files when
1575 --files-with-matches was added. Now, when both options are given, it prints
1576 counts only for those files that have at least one match. (GNU grep just
1577 prints the file name in this circumstance, but including the count seems
1578 more useful - otherwise, why use --count?) Also ensured that the
1579 combination -clh just lists non-zero counts, with no names.
1580
1581 4. The long form of the pcregrep -F option was incorrectly implemented as
1582 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1583 but it seems right to fix it, and I didn't think it was worth preserving
1584 the old behaviour.
1585
1586 5. The command line items --regex=pattern and --regexp=pattern were not
1587 recognized by pcregrep, which required --regex pattern or --regexp pattern
1588 (with a space rather than an '='). The man page documented the '=' forms,
1589 which are compatible with GNU grep; these now work.
1590
1591 6. No libpcreposix.pc file was created for pkg-config; there was just
1592 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1593
1594 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1595 when UCP support is not needed, by modifying the Python script that
1596 generates it from Unicode data files. This should not matter if the module
1597 is correctly used as a library, but I received one complaint about 50K of
1598 unwanted data. My guess is that the person linked everything into his
1599 program rather than using a library. Anyway, it does no harm.
1600
1601 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1602 was a minimum greater than 1 for a wide character in a possessive
1603 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1604 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1605 character. Chaos in the form of incorrect output or a compiling loop could
1606 result.
1607
1608 9. The restrictions on what a pattern can contain when partial matching is
1609 requested for pcre_exec() have been removed. All patterns can now be
1610 partially matched by this function. In addition, if there are at least two
1611 slots in the offset vector, the offset of the earliest inspected character
1612 for the match and the offset of the end of the subject are set in them when
1613 PCRE_ERROR_PARTIAL is returned.
1614
1615 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1616 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1617 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1618 and may be more useful for multi-segment matching.
1619
1620 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1621 used to be given if ever the end of the subject was reached; now it is
1622 given only if matching could not proceed because another character was
1623 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1624 string "Z", which now yields "no match" instead of "partial match". In the
1625 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1626 final character ended with (*FAIL).
1627
1628 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1629 if the pattern had a "must contain" character that was already found in the
1630 earlier partial match, unless partial matching was again requested. For
1631 example, with the pattern /dog.(body)?/, the "must contain" character is
1632 "g". If the first part-match was for the string "dog", restarting with
1633 "sbody" failed. This bug has been fixed.
1634
1635 13. The string returned by pcre_dfa_exec() after a partial match has been
1636 changed so that it starts at the first inspected character rather than the
1637 first character of the match. This makes a difference only if the pattern
1638 starts with a lookbehind assertion or \b or \B (\K is not supported by
1639 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1640 matching functions compatible, and I think it's the right thing to do.
1641
1642 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1643 so that the demonstration program is easily available in environments where
1644 PCRE has not been installed from source.
1645
1646 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1647 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1648 library.
1649
1650 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1651 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1652 is not the first non-POSIX option to be added. Clearly some people find
1653 these options useful.
1654
1655 17. If a caller to the POSIX matching function regexec() passes a non-zero
1656 value for nmatch with a NULL value for pmatch, the value of
1657 nmatch is forced to zero.
1658
1659 18. RunGrepTest did not have a test for the availability of the -u option of
1660 the diff command, as RunTest does. It now checks in the same way as
1661 RunTest, and also checks for the -b option.
1662
1663 19. If an odd number of negated classes containing just a single character
1664 interposed, within parentheses, between a forward reference to a named
1665 subpattern and the definition of the subpattern, compilation crashed with
1666 an internal error, complaining that it could not find the referenced
1667 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1668 [The bug was that it was starting one character too far in when skipping
1669 over the character class, thus treating the ] as data rather than
1670 terminating the class. This meant it could skip too much.]
1671
1672 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1673 /g option in pcretest when the pattern contains \K, which makes it possible
1674 to have an empty string match not at the start, even when the pattern is
1675 anchored. Updated pcretest and pcredemo to use this option.
1676
1677 21. If the maximum number of capturing subpatterns in a recursion was greater
1678 than the maximum at the outer level, the higher number was returned, but
1679 with unset values at the outer level. The correct (outer level) value is
1680 now given.
1681
1682 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1683 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1684 make it do so. The string so far is captured, making this feature
1685 compatible with Perl.
1686
1687 23. The tests have been re-organized, adding tests 11 and 12, to make it
1688 possible to check the Perl 5.10 features against Perl 5.10.
1689
1690 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1691 pattern matches a fixed length string. PCRE did not allow this; now it
1692 does. Neither allows recursion.
1693
1694 25. I finally figured out how to implement a request to provide the minimum
1695 length of subject string that was needed in order to match a given pattern.
1696 (It was back references and recursion that I had previously got hung up
1697 on.) This code has now been added to pcre_study(); it finds a lower bound
1698 to the length of subject needed. It is not necessarily the greatest lower
1699 bound, but using it to avoid searching strings that are too short does give
1700 some useful speed-ups. The value is available to calling programs via
1701 pcre_fullinfo().
1702
1703 26. While implementing 25, I discovered to my embarrassment that pcretest had
1704 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1705 study optimizations had never been tested with that matching function.
1706 Oops. What is worse, even when it was passed study data, there was a bug in
1707 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1708 were also very few tests of studied patterns with pcre_dfa_exec().
1709
1710 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1711 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1712 on the other side of the coin, they are no longer allowed to have different
1713 names, because these cannot be distinguished in PCRE, and this has caused
1714 confusion. (This is a difference from Perl.)
1715
1716 28. When duplicate subpattern names are present (necessarily with different
1717 numbers, as required by 27 above), and a test is made by name in a
1718 conditional pattern, either for a subpattern having been matched, or for
1719 recursion in such a pattern, all the associated numbered subpatterns are
1720 tested, and the overall condition is true if the condition is true for any
1721 one of them. This is the way Perl works, and is also more like the way
1722 testing by number works.
1723
1724
1725 Version 7.9 11-Apr-09
1726 ---------------------
1727
1728 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1729 (pcretest), all targets were linked against these libraries. This included
1730 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1731 libraries. This caused unwanted dependencies to be created. This problem
1732 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1733 pcretest is linked with readline.
1734
1735 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1736 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1737 moved outside it again, because FALSE and TRUE are already defined in AIX,
1738 but BOOL is not.
1739
1740 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1741 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1742
1743 4. The pcregrep documentation said spaces were inserted as well as colons (or
1744 hyphens) following file names and line numbers when outputting matching
1745 lines. This is not true; no spaces are inserted. I have also clarified the
1746 wording for the --colour (or --color) option.
1747
1748 5. In pcregrep, when --colour was used with -o, the list of matching strings
1749 was not coloured; this is different to GNU grep, so I have changed it to be
1750 the same.
1751
1752 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1753 each matching line was coloured. Now it goes on to look for further matches
1754 of any of the test patterns, which is the same behaviour as GNU grep.
1755
1756 7. A pattern that could match an empty string could cause pcregrep to loop; it
1757 doesn't make sense to accept an empty string match in pcregrep, so I have
1758 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1759 seems to be how GNU grep behaves. [But see later change 40 for release
1760 8.33.]
1761
1762 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1763 start or after a newline", because the conditional assertion was not being
1764 correctly handled. The rule now is that both the assertion and what follows
1765 in the first alternative must satisfy the test.
1766
1767 9. If auto-callout was enabled in a pattern with a conditional group whose
1768 condition was an assertion, PCRE could crash during matching, both with
1769 pcre_exec() and pcre_dfa_exec().
1770
1771 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1772 used for matching.
1773
1774 11. Unicode property support in character classes was not working for
1775 characters (bytes) greater than 127 when not in UTF-8 mode.
1776
1777 12. Added the -M command line option to pcretest.
1778
1779 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1780
1781 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1782
1783 16. Added comments and documentation about mis-use of no_arg in the C++
1784 wrapper.
1785
1786 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1787 from Martin Jerabek that uses macro names for all relevant character and
1788 string constants.
1789
1790 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1791 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1792 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1793 these, but not everybody uses configure.
1794
1795 19. A conditional group that had only one branch was not being correctly
1796 recognized as an item that could match an empty string. This meant that an
1797 enclosing group might also not be so recognized, causing infinite looping
1798 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1799 with the subject "ab", where knowledge that the repeated group can match
1800 nothing is needed in order to break the loop.
1801
1802 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1803 exec(), but without supplying a callout function, matching went wrong.
1804
1805 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1806 leak if the size of the offset vector was greater than 30. When the vector
1807 is smaller, the saved offsets during recursion go onto a local stack
1808 vector, but for larger vectors malloc() is used. It was failing to free
1809 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1810 error, in fact).
1811
1812 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1813 heapframe that is used only when UTF-8 support is enabled. This caused no
1814 problem, but was untidy.
1815
1816 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1817 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1818 included within another project.
1819
1820 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1821 slightly modified by me:
1822
1823 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1824 not building pcregrep.
1825
1826 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1827 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1828
1829 25. Forward references, both numeric and by name, in patterns that made use of
1830 duplicate group numbers, could behave incorrectly or give incorrect errors,
1831 because when scanning forward to find the reference group, PCRE was not
1832 taking into account the duplicate group numbers. A pattern such as
1833 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1834
1835 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1836 the feature of a custom pointer more persuasive (as requested by a user).
1837
1838 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1839 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1840 pre-defined.
1841
1842 28. Added support for (*UTF8) at the start of a pattern.
1843
1844 29. Arrange for flags added by the "release type" setting in CMake to be shown
1845 in the configuration summary.
1846
1847
1848 Version 7.8 05-Sep-08
1849 ---------------------
1850
1851 1. Replaced UCP searching code with optimized version as implemented for Ad
1852 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1853 stage table and inline lookup instead of a function, giving speed ups of 2
1854 to 5 times on some simple patterns that I tested. Permission was given to
1855 distribute the MultiStage2.py script that generates the tables (it's not in
1856 the tarball, but is in the Subversion repository).
1857
1858 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1859 scripts.
1860
1861 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1862 a group with a zero qualifier. The result of the study could be incorrect,
1863 or the function might crash, depending on the pattern.
1864
1865 4. Caseless matching was not working for non-ASCII characters in back
1866 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1867 It now works when Unicode Property Support is available.
1868
1869 5. In pcretest, an escape such as \x{de} in the data was always generating
1870 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1871 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1872 truncation.
1873
1874 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1875
1876 7. Added two (int) casts to pcregrep when printing the difference of two
1877 pointers, in case they are 64-bit values.
1878
1879 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1880 test 2 if it fails.
1881
1882 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1883 and a #define of that name to empty if it is not externally set. This is to
1884 allow users of MSVC to set it if necessary.
1885
1886 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1887 the convenience functions in the pcre_get.c source file.
1888
1889 11. An option change at the start of a pattern that had top-level alternatives
1890 could cause overwriting and/or a crash. This command provoked a crash in
1891 some environments:
1892
1893 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1894
1895 This potential security problem was recorded as CVE-2008-2371.
1896
1897 12. For a pattern where the match had to start at the beginning or immediately
1898 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1899 pcre_dfa_exec() could read past the end of the passed subject if there was
1900 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1901 pcretest so that it places the subject at the end of its malloc-ed buffer.
1902
1903 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1904 exec() might read past the end of the data buffer in UTF-8 mode.
1905
1906 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1907 the data contained the byte 0x85 as part of a UTF-8 character within its
1908 first line. This applied both to normal and DFA matching.
1909
1910 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1911 /^[^d]*?$/8 failed to match "abc".
1912
1913 16. Added a missing copyright notice to pcrecpp_internal.h.
1914
1915 17. Make it more clear in the documentation that values returned from
1916 pcre_exec() in ovector are byte offsets, not character counts.
1917
1918 18. Tidied a few places to stop certain compilers from issuing warnings.
1919
1920 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1921 supplied by Stefan Weber. I made a further small update for 7.8 because
1922 there is a change of source arrangements: the pcre_searchfuncs.c module is
1923 replaced by pcre_ucd.c.
1924
1925
1926 Version 7.7 07-May-08
1927 ---------------------
1928
1929 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1930 a string to a long long, pretend we don't even have a long long." This is
1931 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1932
1933 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1934 pre-7.6 versions, which defined a global no_arg variable instead of putting
1935 it in the RE class. (See also #8 below.)
1936
1937 3. Remove a line of dead code, identified by coverity and reported by Nuno
1938 Lopes.
1939
1940 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1941
1942 (1) The include/exclude patterns were being applied to the whole pathnames
1943 of files, instead of just to the final components.
1944
1945 (2) If there was more than one level of directory, the subdirectories were
1946 skipped unless they satisfied the include/exclude conditions. This is
1947 inconsistent with GNU grep (and could even be seen as contrary to the
1948 pcregrep specification - which I improved to make it absolutely clear).
1949 The action now is always to scan all levels of directory, and just
1950 apply the include/exclude patterns to regular files.
1951
1952 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1953 --exclude_dir in the tests to avoid scanning .svn directories.
1954
1955 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1956 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1957 doesn't support NULs in patterns.
1958
1959 7. Added some missing "const"s to declarations of static tables in
1960 pcre_compile.c and pcre_dfa_exec.c.
1961
1962 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1963 caused by fix #2 above. (Subsequently also a second patch to fix the
1964 first patch. And a third patch - this was a messy problem.)
1965
1966 9. Applied Craig's patch to remove the use of push_back().
1967
1968 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1969 matching function regexec().
1970
1971 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1972 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1973 references. PCRE supports relative numbers with this syntax (I don't think
1974 Oniguruma does).
1975
1976 12. Previously, a group with a zero repeat such as (...){0} was completely
1977 omitted from the compiled regex. However, this means that if the group
1978 was called as a subroutine from elsewhere in the pattern, things went wrong
1979 (an internal error was given). Such groups are now left in the compiled
1980 pattern, with a new opcode that causes them to be skipped at execution
1981 time.
1982
1983 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1984 to the way PCRE behaves:
1985
1986 (a) A lone ] character is dis-allowed (Perl treats it as data).
1987
1988 (b) A back reference to an unmatched subpattern matches an empty string
1989 (Perl fails the current match path).
1990
1991 (c) A data ] in a character class must be notated as \] because if the
1992 first data character in a class is ], it defines an empty class. (In
1993 Perl it is not possible to have an empty class.) The empty class []
1994 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1995 The negative empty class [^] matches any one character, independently
1996 of the DOTALL setting.
1997
1998 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1999 non-existent subpattern following a character class starting with ']' and
2000 containing () gave an internal compiling error instead of "reference to
2001 non-existent subpattern". Fortunately, when the pattern did exist, the
2002 compiled code was correct. (When scanning forwards to check for the
2003 existence of the subpattern, it was treating the data ']' as terminating
2004 the class, so got the count wrong. When actually compiling, the reference
2005 was subsequently set up correctly.)
2006
2007 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2008 it was being rejected as not supported by pcre_dfa_exec(), even though
2009 other assertions are supported. I have made pcre_dfa_exec() support
2010 (*FAIL).
2011
2012 16. The implementation of 13c above involved the invention of a new opcode,
2013 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2014 cannot be changed at match time, I realized I could make a small
2015 improvement to matching performance by compiling OP_ALLANY instead of
2016 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2017 on the OP_ANY path.
2018
2019 17. Compiling pcretest on Windows with readline support failed without the
2020 following two fixes: (1) Make the unistd.h include conditional on
2021 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2022
2023 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2024 ncurses library to be included for pcretest when ReadLine support is
2025 requested, but also to allow for it to be overridden. This patch came from
2026 Daniel Bergström.
2027
2028 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2029 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2030 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2031 spotting this.
2032
2033
2034 Version 7.6 28-Jan-08
2035 ---------------------
2036
2037 1. A character class containing a very large number of characters with
2038 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2039 overflow.
2040
2041 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2042 HAVE_LONG_LONG is not defined.
2043
2044 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2045 bring them up to date and include new features. This patch includes:
2046
2047 - Fixed PH's badly added libz and libbz2 support.
2048 - Fixed a problem with static linking.
2049 - Added pcredemo. [But later removed - see 7 below.]
2050 - Fixed dftables problem and added an option.
2051 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2052 HAVE_LONG_LONG.
2053 - Added readline support for pcretest.
2054 - Added an listing of the option settings after cmake has run.
2055
2056 4. A user submitted a patch to Makefile that makes it easy to create
2057 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2058 Makefile.am that cause it to include this special target, without
2059 affecting anything else. Note that the same mingw target plus all
2060 the other distribution libraries and programs are now supported
2061 when configuring with CMake (see 6 below) instead of with
2062 Configure/Make.
2063
2064 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2065 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2066 exported in the Windows port". It has not yet been confirmed that the patch
2067 solves the problem, but it does no harm.
2068
2069 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2070 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2071 with CMake, and also correct the comment about stack recursion.
2072
2073 7. Remove the automatic building of pcredemo from the ./configure system and
2074 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2075 of a program that users should build themselves after PCRE is installed, so
2076 building it automatically is not really right. What is more, it gave
2077 trouble in some build environments.
2078
2079 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2080
2081
2082 Version 7.5 10-Jan-08
2083 ---------------------
2084
2085 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2086 values in parens when parsing an RE using the C++ wrapper."
2087
2088 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2089 Characters greater than 255 were excluded from the class instead of being
2090 included.
2091
2092 3. The same bug as (2) above applied to negated POSIX classes such as
2093 [:^space:].
2094
2095 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2096 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2097 I have changed it.
2098
2099 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2100 first named subpattern) and a construct such as (?&a) would reference the
2101 first named subpattern whose name started with "a" (in other words, the
2102 length check was missing). Both these problems are fixed. "Subpattern name
2103 expected" is now given for (?&) (a zero-length name), and this patch also
2104 makes it give the same error for \k'' (previously it complained that that
2105 was a reference to a non-existent subpattern).
2106
2107 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2108 this is right because (?- can be followed by option settings as well as by
2109 digits. I have, however, made the messages clearer.
2110
2111 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2112 than the number used in the conditional) now cause a compile-time error.
2113 This is actually not compatible with Perl, which accepts such patterns, but
2114 treats the conditional as always being FALSE (as PCRE used to), but it
2115 seems to me that giving a diagnostic is better.
2116
2117 8. Change "alphameric" to the more common word "alphanumeric" in comments
2118 and messages.
2119
2120 9. Fix two occurrences of "backslash" in comments that should have been
2121 "backspace".
2122
2123 10. Remove two redundant lines of code that can never be obeyed (their function
2124 was moved elsewhere).
2125
2126 11. The program that makes PCRE's Unicode character property table had a bug
2127 which caused it to generate incorrect table entries for sequences of
2128 characters that have the same character type, but are in different scripts.
2129 It amalgamated them into a single range, with the script of the first of
2130 them. In other words, some characters were in the wrong script. There were
2131 thirteen such cases, affecting characters in the following ranges:
2132
2133 U+002b0 - U+002c1
2134 U+0060c - U+0060d
2135 U+0061e - U+00612
2136 U+0064b - U+0065e
2137 U+0074d - U+0076d
2138 U+01800 - U+01805
2139 U+01d00 - U+01d77
2140 U+01d9b - U+01dbf
2141 U+0200b - U+0200f
2142 U+030fc - U+030fe
2143 U+03260 - U+0327f
2144 U+0fb46 - U+0fbb1
2145 U+10450 - U+1049d
2146
2147 12. The -o option (show only the matching part of a line) for pcregrep was not
2148 compatible with GNU grep in that, if there was more than one match in a
2149 line, it showed only the first of them. It now behaves in the same way as
2150 GNU grep.
2151
2152 13. If the -o and -v options were combined for pcregrep, it printed a blank
2153 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2154 does the same. The return code can be used to tell if there were any
2155 non-matching lines.
2156
2157 14. Added --file-offsets and --line-offsets to pcregrep.
2158
2159 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2160 infinitely looping recursion. The bug was that positive lookaheads were not
2161 being skipped when checking for a possible empty match (negative lookaheads
2162 and both kinds of lookbehind were skipped).
2163
2164 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2165 inclusion of <windows.h> to before rather than after the definition of
2166 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2167
2168 17. Specifying a possessive quantifier with a specific limit for a Unicode
2169 character property caused pcre_compile() to compile bad code, which led at
2170 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2171 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2172 caused the error; without that there was no problem.
2173
2174 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2175
2176 19. Added --enable-pcretest-libreadline.
2177
2178 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2179 RE::GlobalReplace(). As a result, the number of replacements returned was
2180 double what it should be. I removed one of the increments, but Craig sent a
2181 later patch that removed the other one (the right fix) and added unit tests
2182 that check the return values (which was not done before).
2183
2184 21. Several CMake things:
2185
2186 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2187 the names libpcre and libpcreposix, not just pcre and pcreposix.
2188
2189 (2) The above change means that pcretest and pcregrep are now correctly
2190 linked with the newly-built libraries, not previously installed ones.
2191
2192 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2193
2194 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2195 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2196 UTF-8 newline character). The key issue is that the pattern starts .*;
2197 this means that the match must be either at the beginning, or after a
2198 newline. The bug was in the code for advancing after a failed match and
2199 checking that the new position followed a newline. It was not taking
2200 account of UTF-8 characters correctly.
2201
2202 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2203 character classes. PCRE was not treating the sequence [:...:] as a
2204 character class unless the ... were all letters. Perl, however, seems to
2205 allow any characters between [: and :], though of course it rejects as
2206 unknown any "names" that contain non-letters, because all the known class
2207 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2208 for example, whereas PCRE did not - it did not recognize a POSIX character
2209 class. This seemed a bit dangerous, so the code has been changed to be
2210 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2211 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2212 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2213 Perl does, and where it didn't before.
2214
2215 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2216 Windows environments %n is disabled by default.
2217
2218
2219 Version 7.4 21-Sep-07
2220 ---------------------
2221
2222 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2223 means that a class such as [\s] counted as "explicit reference to CR or
2224 LF". That isn't really right - the whole point of the change was to try to
2225 help when there was an actual mention of one of the two characters. So now
2226 the change happens only if \r or \n (or a literal CR or LF) character is
2227 encountered.
2228
2229 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2230 of both had grown to the point where there were only 3 bits left.
2231 Fortunately, there was spare space in the data structure, and so I have
2232 moved the internal flags into a new 16-bit field to free up more option
2233 bits.
2234
2235 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2236 but did not set the internal JCHANGED flag - either of these is enough to
2237 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2238 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2239 start sets both bits.
2240
2241 4. Added options (at build time, compile time, exec time) to change \R from
2242 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2243
2244 5. doc/pcresyntax.html was missing from the distribution.
2245
2246 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2247 compatibility, even though it is no longer used.
2248
2249 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2250 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2251 windows.h file is present (where different names are used). [This was
2252 reversed later after testing - see 16 below.]
2253
2254 8. Changed all #include <config.h> to #include "config.h". There were also
2255 some further <pcre.h> cases that I changed to "pcre.h".
2256
2257 9. When pcregrep was used with the --colour option, it missed the line ending
2258 sequence off the lines that it output.
2259
2260 10. It was pointed out to me that arrays of string pointers cause lots of
2261 relocations when a shared library is dynamically loaded. A technique of
2262 using a single long string with a table of offsets can drastically reduce
2263 these. I have refactored PCRE in four places to do this. The result is
2264 dramatic:
2265
2266 Originally: 290
2267 After changing UCP table: 187
2268 After changing error message table: 43
2269 After changing table of "verbs" 36
2270 After changing table of Posix names 22
2271
2272 Thanks to the folks working on Gregex for glib for this insight.
2273
2274 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2275 unicode-properties was also set.
2276
2277 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2278
2279 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2280 checked only for CRLF.
2281
2282 14. Added casts to pcretest.c to avoid compiler warnings.
2283
2284 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2285
2286 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2287 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2288 entirely. This removes changes made in 7 above.
2289
2290 17. The CMake files have been updated, and there is now more information about
2291 building with CMake in the NON-UNIX-USE document.
2292
2293
2294 Version 7.3 28-Aug-07
2295 ---------------------
2296
2297 1. In the rejigging of the build system that eventually resulted in 7.1, the
2298 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2299 brackets there is not right, since it causes compilers to look for an
2300 installed pcre.h, not the version that is in the source that is being
2301 compiled (which of course may be different). I have changed it back to:
2302
2303 #include "pcre.h"
2304
2305 I have a vague recollection that the change was concerned with compiling in
2306 different directories, but in the new build system, that is taken care of
2307 by the VPATH setting the Makefile.
2308
2309 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2310 when the subject happened to end in the byte 0x85 (e.g. if the last
2311 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2312 characters but of course it shouldn't be taken as a newline when it is part
2313 of another character. The bug was that, for an unlimited repeat of . in
2314 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2315 characters when looking for a newline.
2316
2317 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2318
2319 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2320 in debug output.
2321
2322 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2323 long printing in the pcrecpp unittest when running under MinGW.
2324
2325 6. ESC_K was left out of the EBCDIC table.
2326
2327 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2328 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2329 limit also applies to "virtual nesting" when a pattern is recursive, and in
2330 this case 1000 isn't so big. I have been able to remove this limit at the
2331 expense of backing off one optimization in certain circumstances. Normally,
2332 when pcre_exec() would call its internal match() function recursively and
2333 immediately return the result unconditionally, it uses a "tail recursion"
2334 feature to save stack. However, when a subpattern that can match an empty
2335 string has an unlimited repetition quantifier, it no longer makes this
2336 optimization. That gives it a stack frame in which to save the data for
2337 checking that an empty string has been matched. Previously this was taken
2338 from the 1000-entry workspace that had been reserved. So now there is no
2339 explicit limit, but more stack is used.
2340
2341 8. Applied Daniel's patches to solve problems with the import/export magic
2342 syntax that is required for Windows, and which was going wrong for the
2343 pcreposix and pcrecpp parts of the library. These were overlooked when this
2344 problem was solved for the main library.
2345
2346 9. There were some crude static tests to avoid integer overflow when computing
2347 the size of patterns that contain repeated groups with explicit upper
2348 limits. As the maximum quantifier is 65535, the maximum group length was
2349 set at 30,000 so that the product of these two numbers did not overflow a
2350 32-bit integer. However, it turns out that people want to use groups that
2351 are longer than 30,000 bytes (though not repeat them that many times).
2352 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2353 made it possible to implement the integer overflow checks in a much more
2354 dynamic way, which I have now done. The artificial limitation on group
2355 length has been removed - we now have only the limit on the total length of
2356 the compiled pattern, which depends on the LINK_SIZE setting.
2357
2358 10. Fixed a bug in the documentation for get/copy named substring when
2359 duplicate names are permitted. If none of the named substrings are set, the
2360 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2361 empty string.
2362
2363 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2364 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2365 because the ] is interpreted as the first data character and the
2366 terminating ] is not found. PCRE has been made compatible with Perl in this
2367 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2368 cause memory overwriting.
2369
2370 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2371 string has been matched (to stop an infinite loop). It was not recognizing
2372 a conditional subpattern that could match an empty string if that
2373 subpattern was within another subpattern. For example, it looped when
2374 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2375 condition was not nested. This bug has been fixed.
2376
2377 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2378 past the start of the subject in the presence of bytes with the top bit
2379 set, for example "\x8aBCD".
2380
2381 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2382 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2383
2384 14. Optimized (?!) to (*FAIL).
2385
2386 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2387 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2388 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2389 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2390 does: it's just the validity check that is more restrictive.
2391
2392 16. Inserted checks for integer overflows during escape sequence (backslash)
2393 processing, and also fixed erroneous offset values for syntax errors during
2394 backslash processing.
2395
2396 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2397 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2398
2399 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2400 caused an overrun.
2401
2402 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2403 something other than just ASCII characters) inside a group that had an
2404 unlimited repeat caused a loop at compile time (while checking to see
2405 whether the group could match an empty string).
2406
2407 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2408 [\P{Any}] did so. (Error in the code for printing property names.)
2409
2410 21. An orphan \E inside a character class could cause a crash.
2411
2412 22. A repeated capturing bracket such as (A)? could cause a wild memory
2413 reference during compilation.
2414
2415 23. There are several functions in pcre_compile() that scan along a compiled
2416 expression for various reasons (e.g. to see if it's fixed length for look
2417 behind). There were bugs in these functions when a repeated \p or \P was
2418 present in the pattern. These operators have additional parameters compared
2419 with \d, etc, and these were not being taken into account when moving along
2420 the compiled data. Specifically:
2421
2422 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2423 length.
2424
2425 (b) An item such as \pL+ within a repeated group could cause crashes or
2426 loops.
2427
2428 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2429 "reference to non-existent subpattern" error.
2430
2431 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2432
2433 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2434 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2435
2436 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2437
2438 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2439 character were causing crashes (broken optimization).
2440
2441 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2442 \p or \P) caused a compile-time loop.
2443
2444 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2445 break. For example, the unstudied pattern [\r\n]A does not match the string
2446 "\r\nA" because change 7.0/46 below moves the current point on by two
2447 characters after failing to match at the start. However, the pattern \nA
2448 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2449 the same is true. There doesn't seem any very clean way out of this, but
2450 what I have chosen to do makes the common cases work: PCRE now takes note
2451 of whether there can be an explicit match for \r or \n anywhere in the
2452 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2453 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2454 pattern has explicit CR or LF references.
2455
2456 29. Added (*CR) etc for changing newline setting at start of pattern.
2457
2458
2459 Version 7.2 19-Jun-07
2460 ---------------------
2461
2462 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2463 which is apparently normally available under Windows.
2464
2465 2. Re-jig the pcregrep tests with different newline settings in an attempt
2466 to make them independent of the local environment's newline setting.
2467
2468 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2469
2470 4. Some of the "internals" tests were previously cut out when the link size
2471 was not 2, because the output contained actual offsets. The recent new
2472 "Z" feature of pcretest means that these can be cut out, making the tests
2473 usable with all link sizes.
2474
2475 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2476 stack recursion. This gives a massive performance boost under BSD, but just
2477 a small improvement under Linux. However, it saves one field in the frame
2478 in all cases.
2479
2480 6. Added more features from the forthcoming Perl 5.10:
2481
2482 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2483 recursion call. It refers to the nth most recently opened parentheses.
2484
2485 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2486 to be opened parentheses.
2487
2488 (c) Conditions that refer to capturing parentheses can be specified
2489 relatively, for example, (?(-2)... or (?(+3)...
2490
2491 (d) \K resets the start of the current match so that everything before
2492 is not part of it.
2493
2494 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2495
2496 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2497 reference syntax.
2498
2499 (g) (?| introduces a group in which the numbering of parentheses in each
2500 alternative starts with the same number.
2501
2502 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2503
2504 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2505 PCRE_INFO_JCHANGED.
2506
2507 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2508 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2509 for detecting groups that can match an empty string.
2510
2511 9. A pattern with a very large number of alternatives (more than several
2512 hundred) was running out of internal workspace during the pre-compile
2513 phase, where pcre_compile() figures out how much memory will be needed. A
2514 bit of new cunning has reduced the workspace needed for groups with
2515 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2516 workspace instead of running out of the 4096 that are available.
2517
2518 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2519
2520 11. Applied patch from Google to remove an optimization that didn't quite work.
2521 The report of the bug said:
2522
2523 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2524 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2525 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2526
2527 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2528 it matched the wrong number of bytes.
2529
2530
2531 Version 7.1 24-Apr-07
2532 ---------------------
2533
2534 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2535 that is more "standard", making use of automake and other Autotools. There
2536 is some re-arrangement of the files and adjustment of comments consequent
2537 on this.
2538
2539 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2540 for recursive directory scanning broke on some systems because the files
2541 are not scanned in any specific order and on different systems the order
2542 was different. A call to "sort" has been inserted into RunGrepTest for the
2543 approprate test as a short-term fix. In the longer term there may be an
2544 alternative.
2545
2546 3. I had an email from Eric Raymond about problems translating some of PCRE's
2547 man pages to HTML (despite the fact that I distribute HTML pages, some
2548 people do their own conversions for various reasons). The problems
2549 concerned the use of low-level troff macros .br and .in. I have therefore
2550 removed all such uses from the man pages (some were redundant, some could
2551 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2552 HTML has been updated to handle .nf/.fi and to complain if it encounters
2553 .br or .in.
2554
2555 4. Updated comments in configure.ac that get placed in config.h.in and also
2556 arranged for config.h to be included in the distribution, with the name
2557 config.h.generic, for the benefit of those who have to compile without
2558 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2559
2560 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2561 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2562 makevp.bat for the current PCRE, using the additional files
2563 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2564
2565 6. A Windows user reported a minor discrepancy with test 2, which turned out
2566 to be caused by a trailing space on an input line that had got lost in his
2567 copy. The trailing space was an accident, so I've just removed it.
2568
2569 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2570 that is needed.
2571
2572 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2573 as "const" (a) because they are and (b) because it helps the PHP
2574 maintainers who have recently made a script to detect big data structures
2575 in the php code that should be moved to the .rodata section. I remembered
2576 to update Builducptable as well, so it won't revert if ucptable.h is ever
2577 re-created.
2578
2579 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2580 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2581 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2582 support is not required. This saves 1.5-2K of code, which is important in
2583 some applications.
2584
2585 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2586 so as not to refer to the tables, even though these functions will never be
2587 called when UTF-8 support is disabled. Otherwise there are problems with a
2588 shared library.
2589
2590 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2591
2592 (a) It was defining its arguments as char * instead of void *.
2593
2594 (b) It was assuming that all moves were upwards in memory; this was true
2595 a long time ago when I wrote it, but is no longer the case.
2596
2597 The emulated memove() is provided for those environments that have neither
2598 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2599 is clearly not the case, as these two bugs were recently reported.
2600
2601 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2602 and Detrail to create the HTML documentation, the .txt form of the man
2603 pages, and it removes trailing spaces from listed files. It also creates
2604 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2605 case, it wraps all the #defines with #ifndefs. This script should be run
2606 before "make dist".
2607
2608 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2609 with Unicode property support.
2610
2611 (a) For a maximizing quantifier, if the two different cases of the
2612 character were of different lengths in their UTF-8 codings (there are
2613 some cases like this - I found 11), and the matching function had to
2614 back up over a mixture of the two cases, it incorrectly assumed they
2615 were both the same length.
2616
2617 (b) When PCRE was configured to use the heap rather than the stack for
2618 recursion during matching, it was not correctly preserving the data for
2619 the other case of a UTF-8 character when checking ahead for a match
2620 while processing a minimizing repeat. If the check also involved
2621 matching a wide character, but failed, corruption could cause an
2622 erroneous result when trying to check for a repeat of the original
2623 character.
2624
2625 13. Some tidying changes to the testing mechanism:
2626
2627 (a) The RunTest script now detects the internal link size and whether there
2628 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2629 values substituted by "configure". (The RunGrepTest script already did
2630 this for UTF-8.) The configure.ac script no longer substitutes the
2631 relevant variables.
2632
2633 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2634 with length and offset values. This means that the output is different
2635 for different internal link sizes. Test 2 is skipped for link sizes
2636 other than 2 because of this, bypassing the problem. Unfortunately,
2637 there was also a test in test 3 (the locale tests) that used /B and
2638 failed for link sizes other than 2. Rather than cut the whole test out,
2639 I have added a new /Z option to pcretest that replaces the length and
2640 offset values with spaces. This is now used to make test 3 independent
2641 of link size. (Test 2 will be tidied up later.)
2642
2643 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2644 segmentation fault instead of returning the appropriate error message.
2645
2646 15. In multiline mode when the newline sequence was set to "any", the pattern
2647 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2648 This doesn't seem right; it now treats the CRLF combination as the line
2649 ending, and so does not match in that case. It's only a pattern such as ^$
2650 that would hit this one: something like ^ABC$ would have failed after \r
2651 and then tried again after \r\n.
2652
2653 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2654 in an attempt to make files that differ only in their line terminators
2655 compare equal. This works on Linux.
2656
2657 17. Under certain error circumstances pcregrep might try to free random memory
2658 as it exited. This is now fixed, thanks to valgrind.
2659
2660 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2661 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2662 was because its rules for how to advance for /g after matching an empty
2663 string at the end of a line did not allow for this case. They now check for
2664 it specially.
2665
2666 20. pcretest is supposed to handle patterns and data of any length, by
2667 extending its buffers when necessary. It was getting this wrong when the
2668 buffer for a data line had to be extended.
2669
2670 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2671 CRLF as a newline sequence.
2672
2673 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2674 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2675 I have nevertheless tidied it up.
2676
2677 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2678
2679 24. Added a man page for pcre-config.
2680
2681
2682 Version 7.0 19-Dec-06
2683 ---------------------
2684
2685 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2686 moving to gcc 4.1.1.
2687
2688 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2689 sys/time.h, which is documented as needed for this function. It doesn't
2690 seem to matter on Linux, but it showed up on some releases of OS X.
2691
2692 3. It seems that there are systems where bytes whose values are greater than
2693 127 match isprint() in the "C" locale. The "C" locale should be the
2694 default when a C program starts up. In most systems, only ASCII printing
2695 characters match isprint(). This difference caused the output from pcretest
2696 to vary, making some of the tests fail. I have changed pcretest so that:
2697
2698 (a) When it is outputting text in the compiled version of a pattern, bytes
2699 other than 32-126 are always shown as hex escapes.
2700
2701 (b) When it is outputting text that is a matched part of a subject string,
2702 it does the same, unless a different locale has been set for the match
2703 (using the /L modifier). In this case, it uses isprint() to decide.
2704
2705 4. Fixed a major bug that caused incorrect computation of the amount of memory
2706 required for a compiled pattern when options that changed within the
2707 pattern affected the logic of the preliminary scan that determines the
2708 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2709 that the computed length was too small. The symptoms of this bug were
2710 either the PCRE error "internal error: code overflow" from pcre_compile(),
2711 or a glibc crash with a message such as "pcretest: free(): invalid next
2712 size (fast)". Examples of patterns that provoked this bug (shown in
2713 pcretest format) are:
2714
2715 /(?-x: )/x
2716 /(?x)(?-x: \s*#\s*)/
2717 /((?i)[\x{c0}])/8
2718 /(?i:[\x{c0}])/8
2719
2720 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2721 is now done differently.
2722
2723 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2724 wrapper classes; (b) implement a new function in the C++ scanner that is
2725 more efficient than the old way of doing things because it avoids levels of
2726 recursion in the regex matching; (c) add a paragraph to the documentation
2727 for the FullMatch() function.
2728
2729 6. The escape sequence \n was being treated as whatever was defined as
2730 "newline". Not only was this contrary to the documentation, which states
2731 that \n is character 10 (hex 0A), but it also went horribly wrong when
2732 "newline" was defined as CRLF. This has been fixed.
2733
2734 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2735 was being set to -1 for the "end of line" case (supposedly a value that no
2736 character can have). Though this value is never used (the check for end of
2737 line is "zero bytes in current character"), it caused compiler complaints.
2738 I've changed it to 0xffffffff.
2739
2740 8. In pcre_version.c, the version string was being built by a sequence of
2741 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2742 string (as it is for production releases) called a macro with an empty
2743 argument. The C standard says the result of this is undefined. The gcc
2744 compiler treats it as an empty string (which was what was wanted) but it is
2745 reported that Visual C gives an error. The source has been hacked around to
2746 avoid this problem.
2747
2748 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2749 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2750 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2751 of them did).
2752
2753 10. Originally, pcretest opened its input and output without "b"; then I was
2754 told that "b" was needed in some environments, so it was added for release
2755 5.0 to both the input and output. (It makes no difference on Unix-like
2756 systems.) Later I was told that it is wrong for the input on Windows. I've
2757 now abstracted the modes into two macros, to make it easier to fiddle with
2758 them, and removed "b" from the input mode under Windows.
2759
2760 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2761
2762 12. Added -help and --help to pcretest as an official way of being reminded
2763 of the options.
2764
2765 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2766 and pcrecpp.cc because they annoy compilers at high warning levels.
2767
2768 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2769
2770 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2771 scripts are not C programs :-) and which was not noticed because it works
2772 on Linux.
2773
2774 16. pcretest is supposed to handle any length of pattern and data line (as one
2775 line or as a continued sequence of lines) by extending its input buffer if
2776 necessary. This feature was broken for very long pattern lines, leading to
2777 a string of junk being passed to pcre_compile() if the pattern was longer
2778 than about 50K.
2779
2780 17. I have done a major re-factoring of the way pcre_compile() computes the
2781 amount of memory needed for a compiled pattern. Previously, there was code
2782 that made a preliminary scan of the pattern in order to do this. That was
2783 OK when PCRE was new, but as the facilities have expanded, it has become
2784 harder and harder to keep it in step with the real compile phase, and there
2785 have been a number of bugs (see for example, 4 above). I have now found a
2786 cunning way of running the real compile function in a "fake" mode that
2787 enables it to compute how much memory it would need, while actually only
2788 ever using a few hundred bytes of working memory and without too many
2789 tests of the mode. This should make future maintenance and development
2790 easier. A side effect of this work is that the limit of 200 on the nesting
2791 depth of parentheses has been removed (though this was never a serious
2792 limitation, I suspect). However, there is a downside: pcre_compile() now
2793 runs more slowly than before (30% or more, depending on the pattern). I
2794 hope this isn't a big issue. There is no effect on runtime performance.
2795
2796 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2797 newline (only possible for the last line of a file) and it was a
2798 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2799
2800 19. Added additional timing features to pcretest. (1) The -tm option now times
2801 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2802 separate command line item, by a number that specifies the number of
2803 repeats to use when timing. The default is 50000; this gives better
2804 precision, but takes uncomfortably long for very large patterns.
2805
2806 20. Extended pcre_study() to be more clever in cases where a branch of a
2807 subpattern has no definite first character. For example, (a*|b*)[cd] would
2808 previously give no result from pcre_study(). Now it recognizes that the
2809 first character must be a, b, c, or d.
2810
2811 21. There was an incorrect error "recursive call could loop indefinitely" if
2812 a subpattern (or the entire pattern) that was being tested for matching an
2813 empty string contained only one non-empty item after a nested subpattern.
2814 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2815 incorrectly, because the \d was being skipped in the check.
2816
2817 22. The pcretest program now has a new pattern option /B and a command line
2818 option -b, which is equivalent to adding /B to every pattern. This causes
2819 it to show the compiled bytecode, without the additional information that
2820 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2821 is the same as /B/I).
2822
2823 23. A new optimization is now able automatically to treat some sequences such
2824 as a*b as a*+b. More specifically, if something simple (such as a character
2825 or a simple class like \d) has an unlimited quantifier, and is followed by
2826 something that cannot possibly match the quantified thing, the quantifier
2827 is automatically "possessified".
2828
2829 24. A recursive reference to a subpattern whose number was greater than 39
2830 went wrong under certain circumstances in UTF-8 mode. This bug could also
2831 have affected the operation of pcre_study().
2832
2833 25. Realized that a little bit of performance could be had by replacing
2834 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2835
2836 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2837
2838 27. Possessive quantifiers such as a++ were previously implemented by turning
2839 them into atomic groups such as ($>a+). Now they have their own opcodes,
2840 which improves performance. This includes the automatically created ones
2841 from 23 above.
2842
2843 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2844 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2845 the first matched character to be a colon. This applied both to named and
2846 numbered groups.
2847
2848 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2849
2850 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2851 building PCRE on Windows easier, so I have included it in the distribution.
2852
2853 31. There is now a check in pcretest against a ridiculously large number being
2854 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2855 loop, the loop is abandoned.
2856
2857 32. Forward references to subpatterns in conditions such as (?(2)...) where
2858 subpattern 2 is defined later cause pcre_compile() to search forwards in
2859 the pattern for the relevant set of parentheses. This search went wrong
2860 when there were unescaped parentheses in a character class, parentheses
2861 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2862
2863 33. "Subroutine" calls and backreferences were previously restricted to
2864 referencing subpatterns earlier in the regex. This restriction has now
2865 been removed.
2866
2867 34. Added a number of extra features that are going to be in Perl 5.10. On the
2868 whole, these are just syntactic alternatives for features that PCRE had
2869 previously implemented using the Python syntax or my own invention. The
2870 other formats are all retained for compatibility.
2871
2872 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2873 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2874 also .NET compatible.
2875
2876 (b) A recursion or subroutine call to a named group can now be defined as
2877 (?&name) as well as (?P>name).
2878
2879 (c) A backreference to a named group can now be defined as \k<name> or
2880 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2881 5.10, are also .NET compatible.
2882
2883 (d) A conditional reference to a named group can now use the syntax
2884 (?(<name>) or (?('name') as well as (?(name).
2885
2886 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2887 groups (named and numbered) that are never evaluated inline, but can be
2888 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2889 is always false. There may be only one alternative in such a group.
2890
2891 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2892 as the simple (?(R). The condition is true only if the most recent
2893 recursion is that of the given number or name. It does not search out
2894 through the entire recursion stack.
2895
2896 (g) The escape \gN or \g{N} has been added, where N is a positive or
2897 negative number, specifying an absolute or relative reference.
2898
2899 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2900 some "unreachable code" warnings.
2901
2902 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2903 things, this adds five new scripts.
2904
2905 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2906 There were also incompatibilities regarding the handling of \Q..\E inside
2907 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2908 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2909
2910 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2911 matches an empty string, and forcibly breaks the loop. There were bugs in
2912 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2913 against aaaa the result was just "a" rather than "aaaa", for example. Two
2914 separate and independent bugs (that affected different cases) have been
2915 fixed.
2916
2917 39. Refactored the code to abolish the use of different opcodes for small
2918 capturing bracket numbers. This is a tidy that I avoided doing when I
2919 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2920 The new approach is not only tidier, it makes it possible to reduce the
2921 memory needed to fix the previous bug (38).
2922
2923 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2924 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2925 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2926 mode.
2927
2928 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2929 report.
2930
2931 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2932 copy construction and assignment in the C++ wrapper.
2933
2934 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2935 couple of bugs that could have given wrong results in the "--newline=crlf"
2936 case.
2937
2938 44. Added a number of casts and did some reorganization of signed/unsigned int
2939 variables following suggestions from Dair Grant. Also renamed the variable
2940 "this" as "item" because it is a C++ keyword.
2941
2942 45. Arranged for dftables to add
2943
2944 #include "pcre_internal.h"
2945
2946 to pcre_chartables.c because without it, gcc 4.x may remove the array
2947 definition from the final binary if PCRE is built into a static library and
2948 dead code stripping is activated.
2949
2950 46. For an unanchored pattern, if a match attempt fails at the start of a
2951 newline sequence, and the newline setting is CRLF or ANY, and the next two
2952 characters are CRLF, advance by two characters instead of one.
2953
2954
2955 Version 6.7 04-Jul-06
2956 ---------------------
2957
2958 1. In order to handle tests when input lines are enormously long, pcretest has
2959 been re-factored so that it automatically extends its buffers when
2960 necessary. The code is crude, but this _is_ just a test program. The
2961 default size has been increased from 32K to 50K.
2962
2963 2. The code in pcre_study() was using the value of the re argument before
2964 testing it for NULL. (Of course, in any sensible call of the function, it
2965 won't be NULL.)
2966
2967 3. The memmove() emulation function in pcre_internal.h, which is used on
2968 systems that lack both memmove() and bcopy() - that is, hardly ever -
2969 was missing a "static" storage class specifier.
2970
2971 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2972 containing an extended class (one that cannot be represented by a bitmap
2973 because it contains high-valued characters or Unicode property items, e.g.
2974 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2975 pattern, but PCRE should not loop if you do not (it no longer does).
2976 [Detail: two cases were found: (a) a repeated subpattern containing an
2977 extended class; (b) a recursive reference to a subpattern that followed a
2978 previous extended class. It wasn't skipping over the extended class
2979 correctly when UTF-8 mode was not set.]
2980
2981 5. A negated single-character class was not being recognized as fixed-length
2982 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2983 compile error "lookbehind assertion is not fixed length".
2984
2985 6. The RunPerlTest auxiliary script was showing an unexpected difference
2986 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2987 write a Perl script that can interpret lines of an input file either as
2988 byte characters or as UTF-8, which is what "perltest" was being required to
2989 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2990 can't do is switch easily at run time between having the "use utf8;" pragma
2991 or not. In the end, I fudged it by using the RunPerlTest script to insert
2992 "use utf8;" explicitly for the UTF-8 tests.
2993
2994 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2995 the end of the subject string, contrary to the documentation and to what
2996 Perl does. This was true of both matching functions. Now it matches only at
2997 the start of the subject and immediately after *internal* newlines.
2998
2999 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3000 a pointer to an int instead of a pointer to an unsigned long int. This
3001 caused problems on 64-bit systems.
3002
3003 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3004 instance of the 'standard' template library not being so standard".
3005
3006 10. There was no check on the number of named subpatterns nor the maximum
3007 length of a subpattern name. The product of these values is used to compute
3008 the size of the memory block for a compiled pattern. By supplying a very
3009 long subpattern name and a large number of named subpatterns, the size
3010 computation could be caused to overflow. This is now prevented by limiting
3011 the length of names to 32 characters, and the number of named subpatterns
3012 to 10,000.
3013
3014 11. Subpatterns that are repeated with specific counts have to be replicated in
3015 the compiled pattern. The size of memory for this was computed from the
3016 length of the subpattern and the repeat count. The latter is limited to
3017 65535, but there was no limit on the former, meaning that integer overflow
3018 could in principle occur. The compiled length of a repeated subpattern is
3019 now limited to 30,000 bytes in order to prevent this.
3020
3021 12. Added the optional facility to have named substrings with the same name.
3022
3023 13. Added the ability to use a named substring as a condition, using the
3024 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3025 are numbers (not recommended). Forward references are permitted.
3026
3027 14. Added forward references in named backreferences (if you see what I mean).
3028
3029 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3030 pattern could run off the end of the subject. For example, the pattern
3031 "(?s)(.{1,5})"8 did this with the subject "ab".
3032
3033 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3034 PCRE_CASELESS was set when matching characters that were quantified with ?
3035 or *.
3036
3037 17. A character class other than a single negated character that had a minimum
3038 but no maximum quantifier - for example [ab]{6,} - was not handled
3039 correctly by pce_dfa_exec(). It would match only one character.
3040
3041 18. A valid (though odd) pattern that looked like a POSIX character
3042 class but used an invalid character after [ (for example [[,abc,]]) caused
3043 pcre_compile() to give the error "Failed: internal error: code overflow" or
3044 in some cases to crash with a glibc free() error. This could even happen if
3045 the pattern terminated after [[ but there just happened to be a sequence of
3046 letters, a binary zero, and a closing ] in the memory that followed.
3047
3048 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3049 over the years. Originally (before any Unicode support), just the bottom 8
3050 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3051 output from "man perlunicode" includes this:
3052
3053 The regular expression compiler produces polymorphic opcodes. That
3054 is, the pattern adapts to the data and automatically switches to
3055 the Unicode character scheme when presented with Unicode data--or
3056 instead uses a traditional byte scheme when presented with byte
3057 data.
3058
3059 Sadly, a wide octal escape does not cause a switch, and in a string with
3060 no other multibyte characters, these octal escapes are treated as before.
3061 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3062 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3063 Unicode string.
3064
3065 I have not perpetrated such confusion in PCRE. Up till now, it took just
3066 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3067 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3068 translate to the appropriate multibyte character.
3069
3070 29. Applied some refactoring to reduce the number of warnings from Microsoft
3071 and Borland compilers. This has included removing the fudge introduced
3072 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3073 a warning about an unused variable.
3074
3075 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3076 characters since release 4.0, because Perl (from release 5.004) does not.
3077 [Or at least, is documented not to: some releases seem to be in conflict
3078 with the documentation.] However, when a pattern was studied with
3079 pcre_study() and all its branches started with \s, PCRE still included VT
3080 as a possible starting character. Of course, this did no harm; it just
3081 caused an unnecessary match attempt.
3082
3083 22. Removed a now-redundant internal flag bit that recorded the fact that case
3084 dependency changed within the pattern. This was once needed for "required
3085 byte" processing, but is no longer used. This recovers a now-scarce options
3086 bit. Also moved the least significant internal flag bit to the most-
3087 significant bit of the word, which was not previously used (hangover from
3088 the days when it was an int rather than a uint) to free up another bit for
3089 the future.
3090
3091 23. Added support for CRLF line endings as well as CR and LF. As well as the
3092 default being selectable at build time, it can now be changed at runtime
3093 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3094 specify that it is scanning data with non-default line endings.
3095
3096 24. Changed the definition of CXXLINK to make it agree with the definition of
3097 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3098
3099 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3100 recursions. This makes a big different to stack usage for some patterns.
3101
3102 26. If a subpattern containing a named recursion or subroutine reference such
3103 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3104 the space required for the compiled pattern went wrong and gave too small a
3105 value. Depending on the environment, this could lead to "Failed: internal
3106 error: code overflow at offset 49" or "glibc detected double free or
3107 corruption" errors.
3108
3109 27. Applied patches from Google (a) to support the new newline modes and (b) to
3110 advance over multibyte UTF-8 characters in GlobalReplace.
3111
3112 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3113 difference for some implementation of PCRE in some Windows version.
3114
3115 29. Added some extra testing facilities to pcretest:
3116
3117 \q<number> in a data line sets the "match limit" value
3118 \Q<number> in a data line sets the "match recursion limt" value
3119 -S <number> sets the stack size, where <number> is in megabytes
3120
3121 The -S option isn't available for Windows.
3122
3123
3124 Version 6.6 06-Feb-06
3125 ---------------------
3126
3127 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3128 in pcreposix.h. I have copied the definition from pcre.h.
3129
3130 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3131 because pcre.h is no longer a built file.
3132
3133 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3134 not normally included in the compiled code.
3135
3136
3137 Version 6.5 01-Feb-06
3138 ---------------------
3139
3140 1. When using the partial match feature with pcre_dfa_exec(), it was not
3141 anchoring the second and subsequent partial matches at the new starting
3142 point. This could lead to incorrect results. For example, with the pattern
3143 /1234/, partially matching against "123" and then "a4" gave a match.
3144
3145 2. Changes to pcregrep:
3146
3147 (a) All non-match returns from pcre_exec() were being treated as failures
3148 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3149 error message is output. Some extra information is given for the
3150 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3151 probably the only errors that are likely to be caused by users (by
3152 specifying a regex that has nested indefinite repeats, for instance).
3153 If there are more than 20 of these errors, pcregrep is abandoned.
3154
3155 (b) A binary zero was treated as data while matching, but terminated the
3156 output line if it was written out. This has been fixed: binary zeroes
3157 are now no different to any other data bytes.
3158
3159 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3160 used to set a locale for matching. The --locale=xxxx long option has
3161 been added (no short equivalent) to specify a locale explicitly on the
3162 pcregrep command, overriding the environment variables.
3163
3164 (d) When -B was used with -n, some line numbers in the output were one less
3165 than they should have been.
3166
3167 (e) Added the -o (--only-matching) option.
3168
3169 (f) If -A or -C was used with -c (count only), some lines of context were
3170 accidentally printed for the final match.
3171
3172 (g) Added the -H (--with-filename) option.
3173
3174 (h) The combination of options -rh failed to suppress file names for files
3175 that were found from directory arguments.
3176
3177 (i) Added the -D (--devices) and -d (--directories) options.
3178
3179 (j) Added the -F (--fixed-strings) option.
3180
3181 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3182
3183 (l) Added the --colo(u)r option.
3184
3185 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3186 is not present by default.
3187
3188 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3189 items such as (?R) or (?1), when the recursion could match a number of
3190 alternatives. If it matched one of the alternatives, but subsequently,
3191 outside the recursion, there was a failure, the code tried to back up into
3192 the recursion. However, because of the way PCRE is implemented, this is not
3193 possible, and the result was an incorrect result from the match.
3194
3195 In order to prevent this happening, the specification of recursion has
3196 been changed so that all such subpatterns are automatically treated as
3197 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3198
3199 4. I had overlooked the fact that, in some locales, there are characters for
3200 which isalpha() is true but neither isupper() nor islower() are true. In
3201 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3202 and ordfeminine) are like this. This affected the treatment of \w and \W
3203 when they appeared in character classes, but not when they appeared outside
3204 a character class. The bit map for "word" characters is now created
3205 separately from the results of isalnum() instead of just taking it from the
3206 upper, lower, and digit maps. (Plus the underscore character, of course.)
3207
3208 5. The above bug also affected the handling of POSIX character classes such as
3209 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3210 permanent tables. Instead, the bit maps for such a class were previously
3211 created as the appropriate unions of the upper, lower, and digit bitmaps.
3212 Now they are created by subtraction from the [[:word:]] class, which has
3213 its own bitmap.
3214
3215 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3216 It is created by subtracting the vertical space characters (\x09, \x0a,
3217 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3218 subtraction was done in the overall bitmap for a character class, meaning
3219 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3220 be recognized. This bug has been fixed.
3221
3222 7. Patches from the folks at Google:
3223
3224 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3225 real life, but is still worth protecting against".
3226
3227 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3228 regular expressions".
3229
3230 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3231 have it.
3232
3233 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3234 "configure" and the latter not, in order to fix a problem somebody had
3235 with compiling the Arg class on HP-UX.
3236
3237 (e) Improve the error-handling of the C++ wrapper a little bit.
3238
3239 (f) New tests for checking recursion limiting.
3240
3241 8. The pcre_memmove() function, which is used only if the environment does not
3242 have a standard memmove() function (and is therefore rarely compiled),
3243 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3244 returning a result (though PCRE never actually uses the result).
3245
3246 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3247 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3248 returned instead of calling malloc() with an overflowing number that would
3249 most likely cause subsequent chaos.
3250
3251 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3252
3253 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3254 with this option is matched, the nmatch and pmatch options of regexec() are
3255 ignored.
3256
3257 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3258 provided in case anyone wants to the the POSIX interface with UTF-8
3259 strings.
3260
3261 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3262 C++ linking (needed for some HP-UX environments).
3263
3264 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3265 (unused parameter) and in the pcre_printint() function (omitted "default"
3266 switch label when the default is to do nothing).
3267
3268 15. Added some code to make it possible, when PCRE is compiled as a C++
3269 library, to replace subject pointers for pcre_exec() with a smart pointer
3270 class, thus making it possible to process discontinuous strings.
3271
3272 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3273 much the same function. They were added by different people who were trying
3274 to make PCRE easy to compile on non-Unix systems. It has been suggested
3275 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3276 for compiling on Windows systems. I have therefore replaced it with
3277 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3278 defaults to "extern" for C or "extern C" for C++, which works fine on
3279 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3280 SCOPE with something explicit in config.h. In addition:
3281
3282 (a) pcreposix.h still had just "extern" instead of either of these macros;
3283 I have replaced it with PCRE_DATA_SCOPE.
3284
3285 (b) Functions such as _pcre_xclass(), which are internal to the library,
3286 but external in the C sense, all had PCRE_EXPORT in their definitions.
3287 This is apparently wrong for the Windows case, so I have removed it.
3288 (It makes no difference on Unix-like systems.)
3289
3290 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3291 of recursive calls to match(). This is different to MATCH_LIMIT because
3292 that limits the total number of calls to match(), not all of which increase
3293 the depth of recursion. Limiting the recursion depth limits the amount of
3294 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3295 when PCRE is compiled, and changed at run time. A patch from Google adds
3296 this functionality to the C++ interface.
3297
3298 18. Changes to the handling of Unicode character properties:
3299
3300 (a) Updated the table to Unicode 4.1.0.
3301
3302 (b) Recognize characters that are not in the table as "Cn" (undefined).
3303
3304 (c) I revised the way the table is implemented to a much improved format
3305 which includes recognition of ranges. It now supports the ranges that
3306 are defined in UnicodeData.txt, and it also amalgamates other
3307 characters into ranges. This has reduced the number of entries in the
3308 table from around 16,000 to around 3,000, thus reducing its size
3309 considerably. I realized I did not need to use a tree structure after
3310 all - a binary chop search is just as efficient. Having reduced the
3311 number of entries, I extended their size from 6 bytes to 8 bytes to
3312 allow for more data.
3313
3314 (d) Added support for Unicode script names via properties such as \p{Han}.
3315
3316 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3317 matching that character.
3318
3319 20. When matching a repeated Unicode property with a minimum greater than zero,
3320 (for example \pL{2,}), PCRE could look past the end of the subject if it
3321 reached it while seeking the minimum number of characters. This could
3322 happen only if some of the characters were more than one byte long, because
3323 there is a check for at least the minimum number of bytes.
3324
3325 21. Refactored the implementation of \p and \P so as to be more general, to
3326 allow for more different types of property in future. This has changed the
3327 compiled form incompatibly. Anybody with saved compiled patterns that use
3328 \p or \P will have to recompile them.
3329
3330 22. Added "Any" and "L&" to the supported property types.
3331
3332 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3333 but give a compile time error if the value is greater than 0xff.
3334
3335 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3336 accidentally not being installed or uninstalled.
3337
3338 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3339 made were to insert the current release number. This seemed silly, because
3340 it made things harder for people building PCRE on systems that don't run
3341 "configure". I have turned pcre.h into a distributed file, no longer built
3342 by "configure", with the version identification directly included. There is
3343 no longer a pcre.h.in file.
3344
3345 However, this change necessitated a change to the pcre-config script as
3346 well. It is built from pcre-config.in, and one of the substitutions was the
3347 release number. I have updated configure.ac so that ./configure now finds
3348 the release number by grepping pcre.h.
3349
3350 26. Added the ability to run the tests under valgrind.
3351
3352
3353 Version 6.4 05-Sep-05
3354 ---------------------
3355
3356 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3357 "--" to be printed when multiple files were scanned, even when none of the
3358 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3359 consider it to be a bug, and have restored the previous behaviour.
3360
3361 2. A couple of code tidies to get rid of compiler warnings.
3362
3363 3. The pcretest program used to cheat by referring to symbols in the library
3364 whose names begin with _pcre_. These are internal symbols that are not
3365 really supposed to be visible externally, and in some environments it is
3366 possible to suppress them. The cheating is now confined to including
3367 certain files from the library's source, which is a bit cleaner.
3368
3369 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3370 file's purpose clearer.
3371
3372 5. Reorganized pcre_ucp_findchar().
3373
3374
3375 Version 6.3 15-Aug-05
3376 ---------------------
3377
3378 1. The file libpcre.pc.in did not have general read permission in the tarball.
3379
3380 2. There were some problems when building without C++ support:
3381
3382 (a) If C++ support was not built, "make install" and "make test" still
3383 tried to test it.
3384
3385 (b) There were problems when the value of CXX was explicitly set. Some
3386 changes have been made to try to fix these, and ...
3387
3388 (c) --disable-cpp can now be used to explicitly disable C++ support.
3389
3390 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3391 backslash in a target when C++ was disabled. This confuses some
3392 versions of "make", apparently. Using an intermediate variable solves
3393 this. (Same for CPP_LOBJ.)
3394
3395 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3396 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3397 necessary on certain architectures.
3398
3399 4. Added a setting of -export-symbols-regex to the link command to remove
3400 those symbols that are exported in the C sense, but actually are local
3401 within the library, and not documented. Their names all begin with
3402 "_pcre_". This is not a perfect job, because (a) we have to except some
3403 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3404 available (and never for static libraries). I have made a note to try to
3405 find a way round (a) in the future.
3406
3407
3408 Version 6.2 01-Aug-05
3409 ---------------------
3410
3411 1. There was no test for integer overflow of quantifier values. A construction
3412 such as {1111111111111111} would give undefined results. What is worse, if
3413 a minimum quantifier for a parenthesized subpattern overflowed and became
3414 negative, the calculation of the memory size went wrong. This could have
3415 led to memory overwriting.
3416
3417 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3418
3419 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3420 operating environments where this matters.
3421
3422 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3423 PCRE options from within the C++ wrapper.
3424
3425 5. Named capturing subpatterns were not being correctly counted when a pattern
3426 was compiled. This caused two problems: (a) If there were more than 100
3427 such subpatterns, the calculation of the memory needed for the whole
3428 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3429 back references of the form \12, where the number was greater than 9, were
3430 not recognized as back references, even though there were sufficient
3431 previous subpatterns.
3432
3433 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3434 versions of gcc, e.g. 2.95.4.
3435
3436
3437 Version 6.1 21-Jun-05
3438 ---------------------
3439
3440 1. There was one reference to the variable "posix" in pcretest.c that was not
3441 surrounded by "#if !defined NOPOSIX".
3442
3443 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3444 the cross-check on the old pcre_info() function, for the benefit of the
3445 cut-down version of PCRE that is currently imported into Exim.
3446
3447 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3448 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3449 patterns that start (?i) but I don't think that matters. The use of (?i) is
3450 just an example; this all applies to the other options as well.
3451
3452 4. Since libtool seems to echo the compile commands it is issuing, the output
3453 from "make" can be reduced a bit by putting "@" in front of each libtool
3454 compile command.
3455
3456 5. Patch from the folks at Google for configure.in to be a bit more thorough
3457 in checking for a suitable C++ installation before trying to compile the
3458 C++ stuff. This should fix a reported problem when a compiler was present,
3459 but no suitable headers.
3460
3461 6. The man pages all had just "PCRE" as their title. I have changed them to
3462 be the relevant file name. I have also arranged that these names are
3463 retained in the file doc/pcre.txt, which is a concatenation in text format
3464 of all the man pages except the little individual ones for each function.
3465
3466 7. The NON-UNIX-USE file had not been updated for the different set of source
3467 files that come with release 6. I also added a few comments about the C++
3468 wrapper.
3469
3470
3471 Version 6.0 07-Jun-05
3472 ---------------------
3473
3474 1. Some minor internal re-organization to help with my DFA experiments.
3475
3476 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3477 didn't matter for the library itself when fully configured, but did matter
3478 when compiling without UCP support, or within Exim, where the ucp files are
3479 not imported.
3480
3481 3. Refactoring of the library code to split up the various functions into
3482 different source modules. The addition of the new DFA matching code (see
3483 below) to a single monolithic source would have made it really too
3484 unwieldy, quite apart from causing all the code to be include in a
3485 statically linked application, when only some functions are used. This is
3486 relevant even without the DFA addition now that patterns can be compiled in
3487 one application and matched in another.
3488
3489 The downside of splitting up is that there have to be some external
3490 functions and data tables that are used internally in different modules of
3491 the library but which are not part of the API. These have all had their
3492 names changed to start with "_pcre_" so that they are unlikely to clash
3493 with other external names.
3494
3495 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3496 a different (DFA) algorithm. Although it is slower than the original
3497 function, it does have some advantages for certain types of matching
3498 problem.
3499
3500 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3501 including restarting after a partial match.
3502
3503 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3504 defined when compiling for Windows was sent to me. I have put it into the
3505 code, though I have no means of testing or verifying it.
3506
3507 7. Added the pcre_refcount() auxiliary function.
3508
3509 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3510 match before or at the first newline in the subject string. In pcretest,
3511 the /f option on a pattern can be used to set this.
3512
3513 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3514 would behave wrongly. This has been present in PCRE since release 4.0.
3515
3516 10. A number of changes to the pcregrep command:
3517
3518 (a) Refactored how -x works; insert ^(...)$ instead of setting
3519 PCRE_ANCHORED and checking the length, in preparation for adding
3520 something similar for -w.
3521
3522 (b) Added the -w (match as a word) option.
3523
3524 (c) Refactored the way lines are read and buffered so as to have more
3525 than one at a time available.
3526
3527 (d) Implemented a pcregrep test script.
3528
3529 (e) Added the -M (multiline match) option. This allows patterns to match
3530 over several lines of the subject. The buffering ensures that at least
3531 8K, or the rest of the document (whichever is the shorter) is available
3532 for matching (and similarly the previous 8K for lookbehind assertions).
3533
3534 (f) Changed the --help output so that it now says
3535
3536 -w, --word-regex(p)
3537
3538 instead of two lines, one with "regex" and the other with "regexp"
3539 because that confused at least one person since the short forms are the
3540 same. (This required a bit of code, as the output is generated
3541 automatically from a table. It wasn't just a text change.)
3542
3543 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3544 option but starts with a hyphen. Could be a pattern or a path name
3545 starting with a hyphen, for instance.
3546
3547 (h) "-" can be given as a file name to represent stdin.
3548
3549 (i) When file names are being printed, "(standard input)" is used for
3550 the standard input, for compatibility with GNU grep. Previously
3551 "<stdin>" was used.
3552
3553 (j) The option --label=xxx can be used to supply a name to be used for
3554 stdin when file names are being printed. There is no short form.
3555
3556 (k) Re-factored the options decoding logic because we are going to add
3557 two more options that take data. Such options can now be given in four
3558 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3559
3560 (l) Added the -A, -B, and -C options for requesting that lines of context
3561 around matches be printed.
3562
3563 (m) Added the -L option to print the names of files that do not contain
3564 any matching lines, that is, the complement of -l.
3565
3566 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3567 continue to scan other files.
3568
3569 (o) The -s option was incorrectly implemented. For compatibility with other
3570 greps, it now suppresses the error message for a non-existent or non-
3571 accessible file (but not the return code). There is a new option called
3572 -q that suppresses the output of matching lines, which was what -s was
3573 previously doing.
3574
3575 (p) Added --include and --exclude options to specify files for inclusion
3576 and exclusion when recursing.
3577
3578 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3579 Hopefully, it now does.
3580
3581 12. Missing cast in pcre_study().
3582
3583 13. Added an "uninstall" target to the makefile.
3584
3585 14. Replaced "extern" in the function prototypes in Makefile.in with
3586 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3587 world, but is set differently for Windows.
3588
3589 15. Added a second compiling function called pcre_compile2(). The only
3590 difference is that it has an extra argument, which is a pointer to an
3591 integer error code. When there is a compile-time failure, this is set
3592 non-zero, in addition to the error test pointer being set to point to an
3593 error message. The new argument may be NULL if no error number is required
3594 (but then you may as well call pcre_compile(), which is now just a
3595 wrapper). This facility is provided because some applications need a
3596 numeric error indication, but it has also enabled me to tidy up the way
3597 compile-time errors are handled in the POSIX wrapper.
3598
3599 16. Added VPATH=.libs to the makefile; this should help when building with one
3600 prefix path and installing with another. (Or so I'm told by someone who
3601 knows more about this stuff than I do.)
3602
3603 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3604 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3605 match everything, including newlines. This is not POSIX-compatible, but
3606 somebody wanted the feature. From pcretest it can be activated by using
3607 both the P and the s flags.
3608
3609 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3610
3611 19. libpcre.pc was being incorrectly installed as executable.
3612
3613 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3614 it now also looks for '\r' so that it will work unmodified on Windows.
3615
3616 21. Added Google's contributed C++ wrapper to the distribution.
3617
3618 22. Added some untidy missing memory free() calls in pcretest, to keep
3619 Electric Fence happy when testing.
3620
3621
3622
3623 Version 5.0 13-Sep-04
3624 ---------------------
3625
3626 1. Internal change: literal characters are no longer packed up into items
3627 containing multiple characters in a single byte-string. Each character
3628 is now matched using a separate opcode. However, there may be more than one
3629 byte in the character in UTF-8 mode.
3630
3631 2. The pcre_callout_block structure has two new fields: pattern_position and
3632 next_item_length. These contain the offset in the pattern to the next match
3633 item, and its length, respectively.
3634
3635 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3636 insertion of callouts before each pattern item. Added the /C option to
3637 pcretest to make use of this.
3638
3639 4. On the advice of a Windows user, the lines
3640
3641 #if defined(_WIN32) || defined(WIN32)
3642 _setmode( _fileno( stdout ), 0x8000 );
3643 #endif /* defined(_WIN32) || defined(WIN32) */
3644
3645 have been added to the source of pcretest. This apparently does useful
3646 magic in relation to line terminators.
3647
3648 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3649 for the benefit of those environments where the "b" makes a difference.
3650
3651 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3652 to know about it. I have put a hack into configure.in that adds in code
3653 to set GCC=yes if CC=icc. This seems to end up at a point in the
3654 generated configure script that is early enough to affect the setting of
3655 compiler options, which is what is needed, but I have no means of testing
3656 whether it really works. (The user who reported this had patched the
3657 generated configure script, which of course I cannot do.)
3658
3659 LATER: After change 22 below (new libtool files), the configure script
3660 seems to know about icc (and also ecc). Therefore, I have commented out
3661 this hack in configure.in.
3662
3663 7. Added support for pkg-config (2 patches were sent in).
3664
3665 8. Negated POSIX character classes that used a combination of internal tables
3666 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3667 [[:^ascii]]. Typically, they would match almost any characters. The other
3668 POSIX classes were not broken in this way.
3669
3670 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3671 to find the match, as PCRE was deluded into thinking that the match had to
3672 start at the start point or following a newline. The same bug applied to
3673 patterns with negative forward assertions or any backward assertions
3674 preceding ".*" at the start, unless the pattern required a fixed first
3675 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3676
3677 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3678 starting at the last subject character, bytes beyond the end of the subject
3679 string were read.
3680
3681 11. Renamed the variable "class" as "classbits" to make life easier for C++
3682 users. (Previously there was a macro definition, but it apparently wasn't
3683 enough.)
3684
3685 12. Added the new field "tables" to the extra data so that tables can be passed
3686 in at exec time, or the internal tables can be re-selected. This allows
3687 a compiled regex to be saved and re-used at a later time by a different
3688 program that might have everything at different addresses.
3689
3690 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3691 -R library as well as a -L library.
3692
3693 14. The debugging options of pcretest (-d on the command line or D on a
3694 pattern) showed incorrect output for anything following an extended class
3695 that contained multibyte characters and which was followed by a quantifier.
3696
3697 15. Added optional support for general category Unicode character properties
3698 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3699 support. It adds about 90K to the size of the library. The meanings of the
3700 inbuilt class escapes such as \d and \s have NOT been changed.
3701
3702 16. Updated pcredemo.c to include calls to free() to release the memory for the
3703 compiled pattern.
3704
3705 17. The generated file chartables.c was being created in the source directory
3706 instead of in the building directory. This caused the build to fail if the
3707 source directory was different from the building directory, and was
3708 read-only.
3709
3710 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3711 file. No doubt somebody will tell me if they don't make sense... Also added
3712 Dan Mooney's comments about building on OpenVMS.
3713
3714 19. Added support for partial matching via the PCRE_PARTIAL option for
3715 pcre_exec() and the \P data escape in pcretest.
3716
3717 20. Extended pcretest with 3 new pattern features:
3718
3719 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3720 write the compiled pattern to the file whose name is "rest-of-line".
3721 This is a straight binary dump of the data, with the saved pointer to
3722 the character tables forced to be NULL. The study data, if any, is
3723 written too. After writing, pcretest reads a new pattern.
3724
3725 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3726 compiled pattern from the given file. There must not be any
3727 occurrences of "<" in the file name (pretty unlikely); if there are,
3728 pcretest will instead treat the initial "<" as a pattern delimiter.
3729 After reading in the pattern, pcretest goes on to read data lines as
3730 usual.
3731
3732 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3733 and 16-bit fields in a compiled pattern, to simulate a pattern that
3734 was compiled on a host of opposite endianness.
3735
3736 21. The pcre-exec() function can now cope with patterns that were compiled on
3737 hosts of opposite endianness, with this restriction:
3738
3739 As for any compiled expression that is saved and used later, the tables
3740 pointer field cannot be preserved; the extra_data field in the arguments
3741 to pcre_exec() should be used to pass in a tables address if a value
3742 other than the default internal tables were used at compile time.
3743
3744 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3745 now diagnosed as an error. Previously, most of the time, a negative number
3746 would have been treated as zero, but if in addition "ovector" was passed as
3747 NULL, a crash could occur.
3748
3749 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3750 new versions from the libtool 1.5 distribution (the last one is a copy of
3751 a file called libtool.m4). This seems to have fixed the need to patch
3752 "configure" to support Darwin 1.3 (which I used to do). However, I still
3753 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3754 workstation).
3755
3756 24. Changed the PCRE licence to be the more standard "BSD" licence.
3757
3758
3759 Version 4.5 01-Dec-03
3760 ---------------------
3761
3762 1. There has been some re-arrangement of the code for the match() function so
3763 that it can be compiled in a version that does not call itself recursively.
3764 Instead, it keeps those local variables that need separate instances for
3765 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3766 needs to "recurse". Keeping track of where control must go is done by means
3767 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3768 hide most of the details from the main code, and operates only if
3769 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3770 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3771 operating.
3772
3773 To make it easier for callers to provide specially tailored get/free
3774 functions for this usage, two new functions, pcre_stack_malloc, and
3775 pcre_stack_free, are used. They are always called in strict stacking order,
3776 and the size of block requested is always the same.
3777
3778 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3779 PCRE has been compiled to use the stack or the heap for recursion. The
3780 -C option of pcretest uses this to show which version is compiled.
3781
3782 A new data escape \S, is added to pcretest; it causes the amounts of store
3783 obtained and freed by both kinds of malloc/free at match time to be added
3784 to the output.
3785
3786 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3787 what's available on my current Linux desktop machine.
3788
3789 3. When matching a UTF-8 string, the test for a valid string at the start has
3790 been extended. If start_offset is not zero, PCRE now checks that it points
3791 to a byte that is the start of a UTF-8 character. If not, it returns
3792 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3793 this is necessary because there may be backward assertions in the pattern.
3794 When matching the same subject several times, it may save resources to use
3795 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3796
3797 4. The code for checking the validity of UTF-8 strings has been tightened so
3798 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3799 containing "overlong sequences".
3800
3801 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3802 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3803 should have been "&", but it just so happened that all the cases this let
3804 through by mistake were picked up later in the function.
3805
3806 6. I had used a variable called "isblank" - this is a C99 function, causing
3807 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3808
3809 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3810 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3811 so the version is shown; (c) stop "make test" echoing "./RunTest".
3812
3813 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3814
3815 9. The prototype for memmove() for systems that don't have it was using
3816 size_t, but the inclusion of the header that defines size_t was later. I've
3817 moved the #includes for the C headers earlier to avoid this.
3818
3819 10. Added some adjustments to the code to make it easier to compiler on certain
3820 special systems:
3821
3822 (a) Some "const" qualifiers were missing.
3823 (b) Added the macro EXPORT before all exported functions; by default this
3824 is defined to be empty.
3825 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3826 that it reads its output file name as an argument instead of writing
3827 to the standard output and assuming this can be redirected.
3828
3829 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3830 class containing characters with values greater than 255, PCRE compilation
3831 went into a loop.
3832
3833 12. A recursive reference to a subpattern that was within another subpattern
3834 that had a minimum quantifier of zero caused PCRE to crash. For example,
3835 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3836 recursion. If the recursively-called subpattern itself had a zero repeat,
3837 that was OK.
3838
3839 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3840 buffer into which it was copied (for escape processing) was still set at
3841 1024, so long lines caused crashes.
3842
3843 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3844 "internal error: code overflow...". This applied to any character class
3845 that was followed by a possessive quantifier.
3846
3847 15. Modified the Makefile to add libpcre.la as a prerequisite for
3848 libpcreposix.la because I was told this is needed for a parallel build to
3849 work.
3850
3851 16. If a pattern that contained .* following optional items at the start was
3852 studied, the wrong optimizing data was generated, leading to matching
3853 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3854 matching string must start with a or b or c. The correct conclusion for
3855 this pattern is that a match can start with any character.
3856
3857
3858 Version 4.4 13-Aug-03
3859 ---------------------
3860
3861 1. In UTF-8 mode, a character class containing characters with values between
3862 127 and 255 was not handled correctly if the compiled pattern was studied.
3863 In fixing this, I have also improved the studying algorithm for such
3864 classes (slightly).
3865
3866 2. Three internal functions had redundant arguments passed to them. Removal
3867 might give a very teeny performance improvement.
3868
3869 3. Documentation bug: the value of the capture_top field in a callout is *one
3870 more than* the number of the hightest numbered captured substring.
3871
3872 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3873 in incorrectly linking with a previously installed version. They now link
3874 explicitly with libpcre.la.
3875
3876 5. configure.in no longer needs to recognize Cygwin specially.
3877
3878 6. A problem in pcre.in for Windows platforms is fixed.
3879
3880 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3881 pcretest, it used to include the size of the study block as part of its
3882 output. Unfortunately, the structure contains a field that has a different
3883 size on different hardware architectures. This meant that the tests that
3884 showed this size failed. As the block is currently always of a fixed size,
3885 this information isn't actually particularly useful in pcretest output, so
3886 I have just removed it.
3887
3888 8. Three pre-processor statements accidentally did not start in column 1.
3889 Sadly, there are *still* compilers around that complain, even though
3890 standard C has not required this for well over a decade. Sigh.
3891
3892 9. In pcretest, the code for checking callouts passed small integers in the
3893 callout_data field, which is a void * field. However, some picky compilers
3894 complained about the casts involved for this on 64-bit systems. Now
3895 pcretest passes the address of the small integer instead, which should get
3896 rid of the warnings.
3897
3898 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3899 both compile and run time, and gives an error if an invalid UTF-8 sequence
3900 is found. There is a option for disabling this check in cases where the
3901 string is known to be correct and/or the maximum performance is wanted.
3902
3903 11. In response to a bug report, I changed one line in Makefile.in from
3904
3905 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3906 to
3907 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3908
3909 to look similar to other lines, but I have no way of telling whether this
3910 is the right thing to do, as I do not use Windows. No doubt I'll get told
3911 if it's wrong...
3912
3913
3914 Version 4.3 21-May-03
3915 ---------------------
3916
3917 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3918 Makefile.
3919
3920 2. Some refactoring to improve the quality of the code:
3921
3922 (i) The utf8_table... variables are now declared "const".
3923
3924 (ii) The code for \cx, which used the "case flipping" table to upper case
3925 lower case letters, now just substracts 32. This is ASCII-specific,
3926 but the whole concept of \cx is ASCII-specific, so it seems
3927 reasonable.
3928
3929 (iii) PCRE was using its character types table to recognize decimal and
3930 hexadecimal digits in the pattern. This is silly, because it handles
3931 only 0-9, a-f, and A-F, but the character types table is locale-
3932 specific, which means strange things might happen. A private
3933 table is now used for this - though it costs 256 bytes, a table is
3934 much faster than multiple explicit tests. Of course, the standard
3935 character types table is still used for matching digits in subject
3936 strings against \d.
3937
3938 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3939 ending in _t are). So I've renamed it as ESC_tee.
3940
3941 3. The first argument for regexec() in the POSIX wrapper should have been
3942 defined as "const".
3943
3944 4. Changed pcretest to use malloc() for its buffers so that they can be
3945 Electric Fenced for debugging.
3946
3947 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3948 to read one or more bytes before the start of the subject string. Often this
3949 had no effect on PCRE's behaviour, but in some circumstances it could
3950 provoke a segmentation fault.
3951
3952 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3953 to try to read one or more bytes before the start of the subject string.
3954
3955 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3956 UTF-8 support could misbehave in various ways if the subject string
3957 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3958 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3959 back over UTF-8 characters.)
3960
3961
3962 Version 4.2 14-Apr-03
3963 ---------------------
3964
3965 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3966
3967 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3968 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3969 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3970 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3971 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3972 and BUILD_EXEEXT
3973 Note: automatic setting of the BUILD variables is not yet working
3974 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3975 compile-time but not at link-time
3976 [LINK]: use for linking executables only
3977 make different versions for Windows and non-Windows
3978 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3979 libraries
3980 [LINK_FOR_BUILD]: new variable
3981 [OBJEXT]: use throughout
3982 [EXEEXT]: use throughout
3983 <winshared>: new target
3984 <wininstall>: new target
3985 <dftables.o>: use native compiler
3986 <dftables>: use native linker
3987 <install>: handle Windows platform correctly
3988 <clean>: ditto
3989 <check>: ditto
3990 copy DLL to top builddir before testing
3991
3992 As part of these changes, -no-undefined was removed again. This was reported
3993 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3994 in any case.
3995
3996 3. Some tidies to get rid of compiler warnings:
3997
3998 . In the match_data structure, match_limit was an unsigned long int, whereas
3999 match_call_count was an int. I've made them both unsigned long ints.
4000
4001 . In pcretest the fact that a const uschar * doesn't automatically cast to
4002 a void * provoked a warning.
4003
4004 . Turning on some more compiler warnings threw up some "shadow" variables
4005 and a few more missing casts.
4006
4007 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4008 option, a class that contained a single character with a value between 128
4009 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4010
4011 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4012 option, a class that contained several characters, but with at least one
4013 whose value was between 128 and 255 caused PCRE to crash.
4014
4015
4016 Version 4.1 12-Mar-03
4017 ---------------------
4018
4019 1. Compiling with gcc -pedantic found a couple of places where casts were
4020 needed, and a string in dftables.c that was longer than standard compilers are
4021 required to support.
4022
4023 2. Compiling with Sun's compiler found a few more places where the code could
4024 be tidied up in order to avoid warnings.
4025
4026 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4027 first of these names is deprecated in the latest Autoconf in favour of the name
4028 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4029 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4030 analogy I have changed it to CFLAGS_FOR_BUILD.
4031
4032 4. Added -no-undefined to the linking command in the Makefile, because this is
4033 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4034 linking step for the pcreposix library.
4035
4036 5. PCRE was failing to diagnose the case of two named groups with the same
4037 name.
4038
4039 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4040 literal character that is needed in the subject for a match, and scans along to
4041 ensure that it is present before embarking on the full matching process. This
4042 saves time in cases of nested unlimited repeats that are never going to match.
4043 Problem: the scan can take a lot of time if the subject is very long (e.g.
4044 megabytes), thus penalizing straightforward matches. It is now done only if the
4045 amount of subject to be scanned is less than 1000 bytes.
4046
4047 7. A lesser problem with the same optimization is that it was recording the
4048 first character of an anchored pattern as "needed", thus provoking a search
4049 right along the subject, even when the first match of the pattern was going to
4050 fail. The "needed" character is now not set for anchored patterns, unless it
4051 follows something in the pattern that is of non-fixed length. Thus, it still
4052 fulfils its original purpose of finding quick non-matches in cases of nested
4053 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4054
4055
4056 Version 4.0 17-Feb-03
4057 ---------------------
4058
4059 1. If a comment in an extended regex that started immediately after a meta-item
4060 extended to the end of string, PCRE compiled incorrect data. This could lead to
4061 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4062
4063 2. Moved to autoconf 2.53 and libtool 1.4.2.
4064
4065 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4066 the special perltest8 script is no longer needed - all the tests can be run
4067 from a single perltest script.
4068
4069 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4070 by \s. It has now been removed in PCRE. This means it isn't recognized as
4071 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4072 class [:space:] *does* include VT, thereby creating a mess.
4073
4074 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4075 space and tab.
4076
4077 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4078 its new features into the main test script, reducing the number of scripts.
4079
4080 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4081 were backward compatible, and made the (?i) apply to the whole pattern, as if
4082 /i were given. Now it behaves more logically, and applies the option setting
4083 only to what follows. PCRE has been changed to follow suit. However, if it
4084 finds options settings right at the start of the pattern, it extracts them into
4085 the global options, as before. Thus, they show up in the info data.
4086
4087 8. Added support for the \Q...\E escape sequence. Characters in between are
4088 treated as literals. This is slightly different from Perl in that $ and @ are
4089 also handled as literals inside the quotes. In Perl, they will cause variable
4090 interpolation. Note the following examples:
4091
4092 Pattern PCRE matches Perl matches
4093
4094 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4095 \Qabc\$xyz\E abc\$xyz abc\$xyz
4096 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4097
4098 For compatibility with Perl, \Q...\E sequences are recognized inside character
4099 classes as well as outside them.
4100
4101 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4102 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4103 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4104 signed/unsigned warnings.
4105
4106 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4107 option for pcretest, so I've replaced it by a simple function that does just
4108 that job.
4109
4110 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4111 "pcregrep -".
4112
4113 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4114 Java package. This provides some syntactic sugar for simple cases of what my
4115 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4116 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4117 item, you can use this simplified notation. Note that only makes sense with
4118 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4119 greediness, whatever the setting of the PCRE_UNGREEDY option.
4120
4121 13. A change of greediness default within a pattern was not taking effect at
4122 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4123 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4124 was abstracted outside.
4125
4126 14. PCRE now supports the \G assertion. It is true when the current matching
4127 position is at the start point of the match. This differs from \A when the
4128 starting offset is non-zero. Used with the /g option of pcretest (or similar
4129 code), it works in the same way as it does for Perl's /g option. If all
4130 alternatives of a regex begin with \G, the expression is anchored to the start
4131 match position, and the "anchored" flag is set in the compiled expression.
4132
4133 15. Some bugs concerning the handling of certain option changes within patterns
4134 have been fixed. These applied to options other than (?ims). For example,
4135 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4136 been the other way round. Some of this was related to change 7 above.
4137
4138 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4139 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4140 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4141 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4142
4143 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4144 mode. Unlike ".", it always matches newline, whatever the setting of
4145 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4146 assertions. Perl allows it, but it doesn't (in general) work because it can't
4147 calculate the length of the lookbehind. At least, that's the case for Perl
4148 5.8.0 - I've been told they are going to document that it doesn't work in
4149 future.
4150
4151 18. Added an error diagnosis for escapes that PCRE does not support: these are
4152 \L, \l, \N, \P, \p, \U, \u, and \X.
4153
4154 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4155 reading past the end of the pattern in cases such as /[abcd/.
4156
4157 20. PCRE was getting more memory than necessary for patterns with classes that
4158 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4159
4160 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4161 compiling PCRE for use with Virtual Pascal.
4162
4163 22. Small fix to the Makefile to make it work properly if the build is done
4164 outside the source tree.
4165
4166 23. Added a new extension: a condition to go with recursion. If a conditional
4167 subpattern starts with (?(R) the "true" branch is used if recursion has
4168 happened, whereas the "false" branch is used only at the top level.
4169
4170 24. When there was a very long string of literal characters (over 255 bytes
4171 without UTF support, over 250 bytes with UTF support), the computation of how
4172 much memory was required could be incorrect, leading to segfaults or other
4173 strange effects.
4174
4175 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4176 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4177 there was a subsequent back reference to those brackets. This meant that, for
4178 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4179 possible to check for precisely this case. All we can do is abandon the
4180 optimization if .* occurs inside capturing brackets when there are any back
4181 references whatsoever. (See below for a better fix that came later.)
4182
4183 26. The handling of the optimization for finding the first character of a
4184 non-anchored pattern, and for finding a character that is required later in the
4185 match were failing in some cases. This didn't break the matching; it just
4186 failed to optimize when it could. The way this is done has been re-implemented.
4187
4188 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4189
4190 28. Added a new feature that provides some of the functionality that Perl
4191 provides with (?{...}). The facility is termed a "callout". The way it is done
4192 in PCRE is for the caller to provide an optional function, by setting
4193 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4194 global variable. By default it is unset, which disables all calling out. To get
4195 the function called, the regex must include (?C) at appropriate points. This
4196 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4197 This provides a means of identifying different callout points. When PCRE
4198 reaches such a point in the regex, if pcre_callout has been set, the external
4199 function is called. It is provided with data in a structure called
4200 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4201 matching continues; if it returns a non-zero value, the match at the current
4202 point fails. However, backtracking will occur if possible. [This was changed
4203 later and other features added - see item 49 below.]
4204
4205 29. pcretest is upgraded to test the callout functionality. It provides a
4206 callout function that displays information. By default, it shows the start of
4207 the match and the current position in the text. There are some new data escapes
4208 to vary what happens:
4209
4210 \C+ in addition, show current contents of captured substrings
4211 \C- do not supply a callout function
4212 \C!n return 1 when callout number n is reached
4213 \C!n!m return 1 when callout number n is reached for the mth time
4214
4215 30. If pcregrep was called with the -l option and just a single file name, it
4216 output "<stdin>" if a match was found, instead of the file name.
4217
4218 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4219 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4220 pcre_exec(). This saves a malloc/free per call. The default value of
4221 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4222 when configuring.
4223
4224 32. The default maximum size of a compiled pattern is 64K. There have been a
4225 few cases of people hitting this limit. The code now uses macros to handle the
4226 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4227 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4228 configuring. Tests 2 and 5 work only with 2-byte links because they output
4229 debugging information about compiled patterns.
4230
4231 33. Internal code re-arrangements:
4232
4233 (a) Moved the debugging function for printing out a compiled regex into
4234 its own source file (printint.c) and used #include to pull it into
4235 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4236 separate copies.
4237
4238 (b) Defined the list of op-code names for debugging as a macro in
4239 internal.h so that it is next to the definition of the opcodes.
4240
4241 (c) Defined a table of op-code lengths for simpler skipping along compiled
4242 code. This is again a macro in internal.h so that it is next to the
4243 definition of the opcodes.
4244
4245 34. Added support for recursive calls to individual subpatterns, along the
4246 lines of Robin Houston's patch (but implemented somewhat differently).
4247
4248 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4249 allow it to read and process whole directories in Win32. This code was
4250 contributed by Lionel Fourquaux; it has not been tested by me.
4251
4252 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4253 used to name a group. Names consist of alphanumerics and underscores, and must
4254 be unique. Back references use the syntax (?P=name) and recursive calls use
4255 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4256 numbers. The function pcre_fullinfo() can be used after compilation to extract
4257 a name/number map. There are three relevant calls:
4258
4259 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4260 PCRE_INFO_NAMECOUNT yields the number of entries
4261 PCRE_INFO_NAMETABLE yields a pointer to the map.
4262
4263 The map is a vector of fixed-size entries. The size of each entry depends on
4264 the length of the longest name used. The first two bytes of each entry are the
4265 group number, most significant byte first. There follows the corresponding
4266 name, zero terminated. The names are in alphabetical order.
4267
4268 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4269 case instead of 255. Making it the same both with and without UTF-8 support
4270 means that the same test output works with both.
4271
4272 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4273 calling malloc() with a zero argument.
4274
4275 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4276 optimization. I've improved things by keeping a bitmap of backreferences with
4277 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4278 fact referenced, the optimization can be applied. It is unlikely that a
4279 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4280 the match to follow \n) will appear inside brackets with a number greater than
4281 31, but if it does, any back reference > 31 suppresses the optimization.
4282
4283 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4284 of disabling numbered capturing parentheses. Any opening parenthesis that is
4285 not followed by ? behaves as if it were followed by ?: but named parentheses
4286 can still be used for capturing (and they will acquire numbers in the usual
4287 way).
4288
4289 41. Redesigned the return codes from the match() function into yes/no/error so
4290 that errors can be passed back from deep inside the nested calls. A malloc
4291 failure while inside a recursive subpattern call now causes the
4292 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4293
4294 42. It is now possible to set a limit on the number of times the match()
4295 function is called in a call to pcre_exec(). This facility makes it possible to
4296 limit the amount of recursion and backtracking, though not in a directly
4297 obvious way, because the match() function is used in a number of different
4298 circumstances. The count starts from zero for each position in the subject
4299 string (for non-anchored patterns). The default limit is, for compatibility, a
4300 large number, namely 10 000 000. You can change this in two ways:
4301
4302 (a) When configuring PCRE before making, you can use --with-match-limit=n
4303 to set a default value for the compiled library.
4304
4305 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4306 a different value is set. See 45 below.
4307
4308 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4309
4310 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4311 of things that can be changed at compile time. The first argument specifies
4312 what is wanted and the second points to where the information is to be placed.
4313 The current list of available information is:
4314
4315 PCRE_CONFIG_UTF8
4316
4317 The output is an integer that is set to one if UTF-8 support is available;
4318 otherwise it is set to zero.
4319
4320 PCRE_CONFIG_NEWLINE
4321
4322 The output is an integer that it set to the value of the code that is used for
4323 newline. It is either LF (10) or CR (13).
4324
4325 PCRE_CONFIG_LINK_SIZE
4326
4327 The output is an integer that contains the number of bytes used for internal
4328 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4329
4330 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4331
4332 The output is an integer that contains the threshold above which the POSIX
4333 interface uses malloc() for output vectors. See item 31 above.
4334
4335 PCRE_CONFIG_MATCH_LIMIT
4336
4337 The output is an unsigned integer that contains the default limit of the number
4338 of match() calls in a pcre_exec() execution. See 42 above.
4339
4340 44. pcretest has been upgraded by the addition of the -C option. This causes it
4341 to extract all the available output from the new pcre_config() function, and to
4342 output it. The program then exits immediately.
4343
4344 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4345 order to support additional features. One way would have been to define
4346 pcre_exec2() (for example) with extra arguments, but this would not have been
4347 extensible, and would also have required all calls to the original function to
4348 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4349 is used for passing in "extra" data from pcre_study().
4350
4351 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4352 contains the following fields:
4353
4354 flags a bitmap indicating which of the following fields are set
4355 study_data opaque data from pcre_study()
4356 match_limit a way of specifying a limit on match() calls for a specific
4357 call to pcre_exec()
4358 callout_data data for callouts (see 49 below)
4359
4360 The flag bits are also defined in pcre.h, and are
4361
4362 PCRE_EXTRA_STUDY_DATA
4363 PCRE_EXTRA_MATCH_LIMIT
4364 PCRE_EXTRA_CALLOUT_DATA
4365
4366 The pcre_study() function now returns one of these new pcre_extra blocks, with
4367 the actual study data pointed to by the study_data field, and the
4368 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4369 before. That is, this change is entirely upwards-compatible and requires no
4370 change to existing code.
4371
4372 If you want to pass in additional data to pcre_exec(), you can either place it
4373 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4374 block.
4375
4376 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4377 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4378 times with different match limits, until it finds the minimum value needed for
4379 pcre_exec() to complete. The value is then output. This can be instructive; for
4380 most simple matches the number is quite small, but for pathological cases it
4381 gets very large very quickly.
4382
4383 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4384 returns the size of the data block pointed to by the study_data field in a
4385 pcre_extra block, that is, the value that was passed as the argument to
4386 pcre_malloc() when PCRE was getting memory in which to place the information
4387 created by pcre_study(). The fourth argument should point to a size_t variable.
4388 pcretest has been extended so that this information is shown after a successful
4389 pcre_study() call when information about the compiled regex is being displayed.
4390
4391 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4392 because what follows is always an absolute path. (Later: it turns out that this
4393 is more than cosmetic for MinGW, because it doesn't like empty path
4394 components.)
4395
4396 49. Some changes have been made to the callout feature (see 28 above):
4397
4398 (i) A callout function now has three choices for what it returns:
4399
4400 0 => success, carry on matching
4401 > 0 => failure at this point, but backtrack if possible
4402 < 0 => serious error, return this value from pcre_exec()
4403
4404 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4405 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4406 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4407 use by callout functions. It will never be used by PCRE itself.
4408
4409 (ii) The pcre_extra structure (see 45 above) has a void * field called
4410 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4411 pcre_callout_block structure has a field of the same name. The contents of
4412 the field passed in the pcre_extra structure are passed to the callout
4413 function in the corresponding field in the callout block. This makes it
4414 easier to use the same callout-containing regex from multiple threads. For
4415 testing, the pcretest program has a new data escape
4416
4417 \C*n pass the number n (may be negative) as callout_data
4418
4419 If the callout function in pcretest receives a non-zero value as
4420 callout_data, it returns that value.
4421
4422 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4423 there were some redundant $(CFLAGS) in commands that are now specified as
4424 $(LINK), which already includes $(CFLAGS).
4425
4426 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4427 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4428 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4429 one-byte characters throughout. Note that case-insensitive matching applies
4430 only to characters whose values are less than 256. PCRE doesn't support the
4431 notion of cases for higher-valued characters.
4432
4433 (i) A character class whose characters are all within 0-255 is handled as
4434 a bit map, and the map is inverted for negative classes. Previously, a
4435 character > 255 always failed to match such a class; however it should
4436 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4437
4438 (ii) A negated character class with a single character < 255 is coded as
4439 "not this character" (OP_NOT). This wasn't working properly when the test
4440 character was multibyte, either singly or repeated.
4441
4442 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4443 mode, for example: \x{100}{2,3}.
4444
4445 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4446 singly or repeated) now correctly test multibyte characters. However,
4447 PCRE doesn't recognize any characters with values greater than 255 as
4448 digits, spaces, or word characters. Such characters always match \D, \S,
4449 and \W, and never match \d, \s, or \w.
4450
4451 (v) Classes may now contain characters and character ranges with values
4452 greater than 255. For example: [ab\x{100}-\x{400}].
4453
4454 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4455 PCRE in UTF-8 mode.
4456
4457 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4458 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4459 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4460 value.)
4461
4462 53. The single man page has become too large. I have therefore split it up into
4463 a number of separate man pages. These also give rise to individual HTML pages;
4464 these are now put in a separate directory, and there is an index.html page that
4465 lists them all. Some hyperlinking between the pages has been installed.
4466
4467 54. Added convenience functions for handling named capturing parentheses.
4468
4469 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4470 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4471 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4472 are faulted.
4473
4474 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4475 calling configure. These values are used when compiling the dftables.c program
4476 which is run to generate the source of the default character tables. They
4477 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4478 you will need to set these values.
4479
4480 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4481
4482
4483 Version 3.9 02-Jan-02
4484 ---------------------
4485
4486 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4487
4488 2. If --disable-static was given, the building process failed when trying to
4489 build pcretest and pcregrep. (For some reason it was using libtool to compile
4490 them, which is not right, as they aren't part of the library.)
4491
4492
4493 Version 3.8 18-Dec-01
4494 ---------------------
4495
4496 1. The experimental UTF-8 code was completely screwed up. It was packing the
4497 bytes in the wrong order. How dumb can you get?
4498
4499
4500 Version 3.7 29-Oct-01
4501 ---------------------
4502
4503 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4504 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4505 this didn't happen under Solaris 8, where I normally test things.
4506
4507 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4508 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4509 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4510 this fix an hour or so after the initial 3.7 release.)
4511
4512
4513 Version 3.6 23-Oct-01
4514 ---------------------
4515
4516 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4517 offsets passed as NULL with zero offset count.
4518
4519 2. The config.guess and config.sub files had not been updated when I moved to
4520 the latest autoconf.
4521
4522
4523 Version 3.5 15-Aug-01
4524 ---------------------
4525
4526 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4527 had been forgotten.
4528
4529 2. By using declared but undefined structures, we can avoid using "void"
4530 definitions in pcre.h while keeping the internal definitions of the structures
4531 private.
4532
4533 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4534 user point of view, this means that both static and shared libraries are built
4535 by default, but this can be individually controlled. More of the work of
4536 handling this static/shared cases is now inside libtool instead of PCRE's make
4537 file.
4538
4539 4. The pcretest utility is now installed along with pcregrep because it is
4540 useful for users (to test regexs) and by doing this, it automatically gets
4541 relinked by libtool. The documentation has been turned into a man page, so
4542 there are now .1, .txt, and .html versions in /doc.
4543
4544 5. Upgrades to pcregrep:
4545 (i) Added long-form option names like gnu grep.
4546 (ii) Added --help to list all options with an explanatory phrase.
4547 (iii) Added -r, --recursive to recurse into sub-directories.
4548 (iv) Added -f, --file to read patterns from a file.
4549
4550 6. pcre_exec() was referring to its "code" argument before testing that
4551 argument for NULL (and giving an error if it was NULL).
4552
4553 7. Upgraded Makefile.in to allow for compiling in a different directory from
4554 the source directory.
4555
4556 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4557 options bits, the pointer it was passed was to an int instead of to an unsigned
4558 long int. This mattered only on 64-bit systems.
4559
4560 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4561 generated) instead of pcre.in, which it its source. Also made the same change
4562 in several of the .c files.
4563
4564 10. A new release of gcc defines printf() as a macro, which broke pcretest
4565 because it had an ifdef in the middle of a string argument for printf(). Fixed
4566 by using separate calls to printf().
4567
4568 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4569 script, to force use of CR or LF instead of \n in the source. On non-Unix
4570 systems, the value can be set in config.h.
4571
4572 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4573 absolute limit. Changed the text of the error message to make this clear, and
4574 likewise updated the man page.
4575
4576 13. The limit of 99 on the number of capturing subpatterns has been removed.
4577 The new limit is 65535, which I hope will not be a "real" limit.
4578
4579
4580 Version 3.4 22-Aug-00
4581 ---------------------
4582
4583 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4584
4585 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4586
4587
4588 Version 3.3 01-Aug-00
4589 ---------------------
4590
4591 1. If an octal character was given, but the value was greater than \377, it
4592 was not getting masked to the least significant bits, as documented. This could
4593 lead to crashes in some systems.
4594
4595 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4596 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4597
4598 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4599 These just pass their arguments on to (pcre_free)(), but they are provided
4600 because some uses of PCRE bind it to non-C systems that can call its functions,
4601 but cannot call free() or pcre_free() directly.
4602
4603 4. Add "make test" as a synonym for "make check". Corrected some comments in
4604 the Makefile.
4605
4606 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4607 Makefile.
4608
4609 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4610 command called pgrep for grepping around the active processes.
4611
4612 7. Added the beginnings of support for UTF-8 character strings.
4613
4614 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4615 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4616 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4617 out for the ar command.)
4618
4619
4620 Version 3.2 12-May-00
4621 ---------------------
4622
4623 This is purely a bug fixing release.
4624
4625 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4626 of ZA. This was just one example of several cases that could provoke this bug,
4627 which was introduced by change 9 of version 2.00. The code for breaking
4628 infinite loops after an iteration that matches an empty string was't working
4629 correctly.
4630
4631 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4632 when matched against abbab (for example). After matching an empty string, it
4633 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4634 caused it to match further down the string than it should.
4635
4636 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4637 was there because it doesn't seem to be needed, and it causes trouble on some
4638 systems, as it is not a Standard C header. It has been removed.
4639
4640 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4641 were reported on the Macintosh. The changes were from
4642
4643 while ((c = *(++ptr)) != 0 && c != '\n');
4644 to
4645 while ((c = *(++ptr)) != 0 && c != '\n') ;
4646
4647 Totally extraordinary, but if that's what it takes...
4648
4649 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4650 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4651 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4652 assumes the way PCRE uses memmove() (always moving upwards).
4653
4654 6. PCRE is being used in one environment where strchr() is not available. There
4655 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4656 faster code anyway.
4657
4658
4659 Version 3.1 09-Feb-00
4660 ---------------------
4661
4662 The only change in this release is the fixing of some bugs in Makefile.in for
4663 the "install" target:
4664
4665 (1) It was failing to install pcreposix.h.
4666
4667 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4668
4669
4670 Version 3.0 01-Feb-00
4671 ---------------------
4672
4673 1. Add support for the /+ modifier to perltest (to output $` like it does in
4674 pcretest).
4675
4676 2. Add support for the /g modifier to perltest.
4677
4678 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4679 matches null strings.
4680
4681 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4682 pattern. Perl treats empty patterns specially - it reuses the most recent
4683 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4684 effect.
4685
4686 5. The POSIX interface was broken in that it was just handing over the POSIX
4687 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4688 required a bigger vector, with some working space on the end. This means that
4689 the POSIX wrapper now has to get and free some memory, and copy the results.
4690
4691 6. Added some simple autoconf support, placing the test data and the
4692 documentation in separate directories, re-organizing some of the
4693 information files, and making it build pcre-config (a GNU standard). Also added
4694 libtool support for building PCRE as a shared library, which is now the
4695 default.
4696
4697 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4698 09 are not valid octal constants. Single digits will be used for minor values
4699 less than 10.
4700
4701 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4702 existing programs that set these in the POSIX interface can use PCRE without
4703 modification.
4704
4705 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4706 return all that pcre_info() returns, plus additional data. The pcre_info()
4707 function is retained for compatibility, but is considered to be obsolete.
4708
4709 10. Added experimental recursion feature (?R) to handle one common case that
4710 Perl 5.6 will be able to do with (?p{...}).
4711
4712 11. Added support for POSIX character classes like [:alpha:], which Perl is
4713 adopting.
4714
4715
4716 Version 2.08 31-Aug-99
4717 ----------------------
4718
4719 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4720 trying to match at the startoffset position, but instead was moving forward to
4721 the next newline as if a previous match had failed.
4722
4723 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4724 and could get into a loop if a null string was matched other than at the start
4725 of the subject.
4726
4727 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4728 be distinguished at compile time, and for completeness also added PCRE_DATE.
4729
4730 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4731 in GnuWin32 environments.
4732
4733
4734 Version 2.07 29-Jul-99
4735 ----------------------
4736
4737 1. The documentation is now supplied in plain text form and HTML as well as in
4738 the form of man page sources.
4739
4740 2. C++ compilers don't like assigning (void *) values to other pointer types.
4741 In particular this affects malloc(). Although there is no problem in Standard
4742 C, I've put in casts to keep C++ compilers happy.
4743
4744 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4745 should be (const char *).
4746
4747 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4748 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4749 However, I haven't made this a standard facility. The documentation doesn't
4750 mention it, and the Makefile doesn't support it.
4751
4752 5. The Makefile now contains an "install" target, with editable destinations at
4753 the top of the file. The pcretest program is not installed.
4754
4755 6. pgrep -V now gives the PCRE version number and date.
4756
4757 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4758 causing the entire string to be ignored, instead of just the last character.
4759
4760 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4761 non-matching string, it can take a very, very long time, even for strings of
4762 quite modest length, because of the nested recursion. PCRE now does better in
4763 some of these cases. It does this by remembering the last required literal
4764 character in the pattern, and pre-searching the subject to ensure it is present
4765 before running the real match. In other words, it applies a heuristic to detect
4766 some types of certain failure quickly, and in the above example, if presented
4767 with a string that has no trailing " it gives "no match" very quickly.
4768
4769 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4770 other alternatives are tried instead.
4771
4772
4773 Version 2.06 09-Jun-99
4774 ----------------------
4775
4776 1. Change pcretest's output for amount of store used to show just the code
4777 space, because the remainder (the data block) varies in size between 32-bit and
4778 64-bit systems.
4779
4780 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4781 start matching at. This allows lookbehinds to work when searching for multiple
4782 occurrences in a string.
4783
4784 3. Added additional options to pcretest for testing multiple occurrences:
4785
4786 /+ outputs the rest of the string that follows a match
4787 /g loops for multiple occurrences, using the new startoffset argument
4788 /G loops for multiple occurrences by passing an incremented pointer
4789
4790 4. PCRE wasn't doing the "first character" optimization for patterns starting
4791 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4792 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4793 the letter 'x'. On long subject strings, this gives a significant speed-up.
4794
4795
4796 Version 2.05 21-Apr-99
4797 ----------------------
4798
4799 1. Changed the type of magic_number from int to long int so that it works
4800 properly on 16-bit systems.
4801
4802 2. Fixed a bug which caused patterns starting with .* not to work correctly
4803 when the subject string contained newline characters. PCRE was assuming
4804 anchoring for such patterns in all cases, which is not correct because .* will
4805 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4806 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4807 must be retried after every newline in the subject.
4808
4809
4810 Version 2.04 18-Feb-99
4811 ----------------------
4812
4813 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4814 computation of the store needed to hold the pattern was incorrect (too large).
4815 If such patterns were nested a few deep, this could multiply and become a real
4816 problem.
4817
4818 2. Added /M option to pcretest to show the memory requirement of a specific
4819 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4820
4821 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4822 compiled in such a way that the backtracking after subsequent failure was
4823 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4824 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4825
4826
4827 Version 2.03 02-Feb-99
4828 ----------------------
4829
4830 1. Fixed typo and small mistake in man page.
4831
4832 2. Added 4th condition (GPL supersedes if conflict) and created separate
4833 LICENCE file containing the conditions.
4834
4835 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4836 Perl, that is the internal \ allows the delimiter to be included in the
4837 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4838 the final delimiter, add \ to the end of the pattern (to test the error).
4839
4840 4. Added the convenience functions for extracting substrings after a successful
4841 match. Updated pcretest to make it able to test these functions.
4842
4843
4844 Version 2.02 14-Jan-99
4845 ----------------------
4846
4847 1. Initialized the working variables associated with each extraction so that
4848 their saving and restoring doesn't refer to uninitialized store.
4849
4850 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4851 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4852 fix the problem.
4853
4854 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4855 calls, and wasn't printing the correct value for compiling calls. Increased the
4856 default value of LOOPREPEAT, and the number of significant figures in the
4857 times.
4858
4859 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4860
4861 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4862 a building problem on Windows NT with a FAT file system.
4863
4864
4865 Version 2.01 21-Oct-98
4866 ----------------------
4867
4868 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4869 to character tables built by pcre_maketables() in the current locale. If NULL
4870 is passed, the default tables are used.
4871
4872
4873 Version 2.00 24-Sep-98
4874 ----------------------
4875
4876 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4877 it any more.
4878
4879 2. Allow quantification of (?>) groups, and make it work correctly.
4880
4881 3. The first character computation wasn't working for (?>) groups.
4882
4883 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4884 end of the subject) and add 5.005's \z, which really does match only at the
4885 very end of the subject.
4886
4887 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4888
4889 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4890 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4891 localized options. All options to pcre_study() were also removed.
4892
4893 7. Add other new features from 5.005:
4894
4895 $(?<= positive lookbehind
4896 $(?<! negative lookbehind
4897 (?imsx-imsx) added the unsetting capability
4898 such a setting is global if at outer level; local otherwise
4899 (?imsx-imsx:) non-capturing groups with option setting
4900 (?(cond)re|re) conditional pattern matching
4901
4902 A backreference to itself in a repeated group matches the previous
4903 captured string.
4904
4905 8. General tidying up of studying (both automatic and via "study")
4906 consequential on the addition of new assertions.
4907
4908 9. As in 5.005, unlimited repeated groups that could match an empty substring
4909 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4910 runtime if any iteration does actually match an empty substring.
4911
4912 10. Include the RunTest script in the distribution.
4913
4914 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4915 discrepancies, some of which were old and were also with respect to 5.004. They
4916 have now been fixed.
4917
4918
4919 Version 1.09 28-Apr-98
4920 ----------------------
4921
4922 1. A negated single character class followed by a quantifier with a minimum
4923 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4924 program crashes, or just wrong answers. This did not apply to negated classes
4925 containing more than one character, or to minima other than one.
4926
4927
4928 Version 1.08 27-Mar-98
4929 ----------------------
4930
4931 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4932
4933 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4934 latter must appear before anything that relies on it in the pattern.
4935
4936
4937 Version 1.07 16-Feb-98
4938 ----------------------
4939
4940 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4941 repeat of a potentially empty string).
4942
4943
4944 Version 1.06 23-Jan-98
4945 ----------------------
4946
4947 1. Added Markus Oberhumer's little patches for C++.
4948
4949 2. Literal strings longer than 255 characters were broken.
4950
4951
4952 Version 1.05 23-Dec-97
4953 ----------------------
4954
4955 1. Negated character classes containing more than one character were failing if
4956 PCRE_CASELESS was set at run time.
4957
4958
4959 Version 1.04 19-Dec-97
4960 ----------------------
4961
4962 1. Corrected the man page, where some "const" qualifiers had been omitted.
4963
4964 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4965 input syntax.
4966
4967 3. Fixed memory leak which occurred when a regex with back references was
4968 matched with an offsets vector that wasn't big enough. The temporary memory
4969 that is used in this case wasn't being freed if the match failed.
4970
4971 4. Tidied pcretest to ensure it frees memory that it gets.
4972
4973 5. Temporary memory was being obtained in the case where the passed offsets
4974 vector was exactly big enough.
4975
4976 6. Corrected definition of offsetof() from change 5 below.
4977
4978 7. I had screwed up change 6 below and broken the rules for the use of
4979 setjmp(). Now fixed.
4980
4981
4982 Version 1.03 18-Dec-97
4983 ----------------------
4984
4985 1. A erroneous regex with a missing opening parenthesis was correctly
4986 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4987 on some systems.
4988
4989 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4990 it was reported that one broken compiler failed on the former because "code" is
4991 also an independent variable.
4992
4993 3. The erroneous regex a[]b caused an array overrun reference.
4994
4995 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4996 fail on data ending with that character. (It was going on too far, and checking
4997 the next character, typically a binary zero.) This was specific to the
4998 optimized code for single-character negative classes.
4999
5000 5. Added a contributed patch from the TIN world which does the following:
5001
5002 + Add an undef for memmove, in case the the system defines a macro for it.
5003
5004 + Add a definition of offsetof(), in case there isn't one. (I don't know
5005 the reason behind this - offsetof() is part of the ANSI standard - but
5006 it does no harm).
5007
5008 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5009 most of the places where whitespace preceded '#'. I have given up and
5010 allowed the remaining 2 cases to be at the margin.
5011
5012 + Rename some variables in pcre to eliminate shadowing. This seems very
5013 pedantic, but does no harm, of course.
5014
5015 6. Moved the call to setjmp() into its own function, to get rid of warnings
5016 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5017
5018 7. Constructs such as \d{8,} were compiling into the equivalent of
5019 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5020 outcome, but in this particular case used more store than had been allocated,
5021 which caused the bug to be discovered because it threw up an internal error.
5022
5023 8. The debugging code in both pcre and pcretest for outputting the compiled
5024 form of a regex was going wrong in the case of back references followed by
5025 curly-bracketed repeats.
5026
5027
5028 Version 1.02 12-Dec-97
5029 ----------------------
5030
5031 1. Typos in pcre.3 and comments in the source fixed.
5032
5033 2. Applied a contributed patch to get rid of places where it used to remove
5034 'const' from variables, and fixed some signed/unsigned and uninitialized
5035 variable warnings.
5036
5037 3. Added the "runtest" target to Makefile.
5038
5039 4. Set default compiler flag to -O2 rather than just -O.
5040
5041
5042 Version 1.01 19-Nov-97
5043 ----------------------
5044
5045 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5046 like /([ab]*)*/, that is, for classes with more than one character in them.
5047
5048 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5049 as /((?>a*))*/ (a PCRE_EXTRA facility).
5050
5051
5052 Version 1.00 18-Nov-97
5053 ----------------------
5054
5055 1. Added compile-time macros to support systems such as SunOS4 which don't have
5056 memmove() or strerror() but have other things that can be used instead.
5057
5058 2. Arranged that "make clean" removes the executables.
5059
5060
5061 Version 0.99 27-Oct-97
5062 ----------------------
5063
5064 1. Fixed bug in code for optimizing classes with only one character. It was
5065 initializing a 32-byte map regardless, which could cause it to run off the end
5066 of the memory it had got.
5067
5068 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5069
5070
5071 Version 0.98 22-Oct-97
5072 ----------------------
5073
5074 1. Fixed bug in code for handling temporary memory usage when there are more
5075 back references than supplied space in the ovector. This could cause segfaults.
5076
5077
5078 Version 0.97 21-Oct-97
5079 ----------------------
5080
5081 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5082
5083 2. Optimized negated single characters not to use a bit map.
5084
5085 3. Brought error texts together as macro definitions; clarified some of them;
5086 fixed one that was wrong - it said "range out of order" when it meant "invalid
5087 escape sequence".
5088
5089 4. Changed some char * arguments to const char *.
5090
5091 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5092
5093 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5094 pcretest.
5095
5096
5097 Version 0.96 16-Oct-97
5098 ----------------------
5099
5100 1. Added a simple "pgrep" utility to the distribution.
5101
5102 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5103 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5104 where "ddd" means "one or more decimal digits".
5105
5106 3. Fixed serious bug. If a pattern had a back reference, but the call to
5107 pcre_exec() didn't supply a large enough ovector to record the related
5108 identifying subpattern, the match always failed. PCRE now remembers the number
5109 of the largest back reference, and gets some temporary memory in which to save
5110 the offsets during matching if necessary, in order to ensure that
5111 backreferences always work.
5112
5113 4. Increased the compatibility with Perl in a number of ways:
5114
5115 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5116 to request this handling. The option can be set at compile or exec time.
5117
5118 (b) $ matches before a terminating newline by default; an option
5119 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5120 mode). The option can be set at compile or exec time.
5121
5122 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5123 the same as Perl's. If the decimal number it represents is less than 10
5124 or there aren't that many previous left capturing parentheses, an octal
5125 escape is read. Inside a character class, it's always an octal escape,
5126 even if it is a single digit.
5127
5128 (d) An escaped but undefined alphabetic character is taken as a literal,
5129 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5130 escapes.
5131
5132 (e) {0} is now permitted. (The previous item is removed from the compiled
5133 pattern).
5134
5135 5. Changed all the names of code files so that the basic parts are no longer
5136 than 10 characters, and abolished the teeny "globals.c" file.
5137
5138 6. Changed the handling of character classes; they are now done with a 32-byte
5139 bit map always.
5140
5141 7. Added the -d and /D options to pcretest to make it possible to look at the
5142 internals of compilation without having to recompile pcre.
5143
5144
5145 Version 0.95 23-Sep-97
5146 ----------------------
5147
5148 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5149 \x20 at the start of a run of normal characters. These were being treated as
5150 real characters, instead of the source characters being re-checked.
5151
5152
5153 Version 0.94 18-Sep-97
5154 ----------------------
5155
5156 1. The functions are now thread-safe, with the caveat that the global variables
5157 containing pointers to malloc() and free() or alternative functions are the
5158 same for all threads.
5159
5160 2. Get pcre_study() to generate a bitmap of initial characters for non-
5161 anchored patterns when this is possible, and use it if passed to pcre_exec().
5162
5163
5164 Version 0.93 15-Sep-97
5165 ----------------------
5166
5167 1. /(b)|(:+)/ was computing an incorrect first character.
5168
5169 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5170 but not actually doing anything yet.
5171
5172 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5173 as Perl does (e.g. [-az] or [az-]).
5174
5175 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5176 all possible positions.
5177
5178 5. Split up into different modules to avoid including unneeded functions in a
5179 compiled binary. However, compile and exec are still in one module. The "study"
5180 function is split off.
5181
5182 6. The character tables are now in a separate module whose source is generated
5183 by an auxiliary program - but can then be edited by hand if required. There are
5184 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5185 toupper() in the code.
5186
5187 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5188 make them global. Abolish the function for setting them, as the caller can now
5189 set them directly.
5190
5191
5192 Version 0.92 11-Sep-97
5193 ----------------------
5194
5195 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5196 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5197
5198 2. Caseless matching was not working in character classes if the characters in
5199 the pattern were in upper case.
5200
5201 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5202
5203 4. Make PCRE_ANCHORED public and accept as a compile option.
5204
5205 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5206 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5207 pass them.
5208
5209 6. Give an error if bad option bits passed at compile or run time.
5210
5211 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5212 pcretest to cause it to pass that flag.
5213
5214 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5215 options, and the first character, if set.
5216
5217 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5218
5219
5220 Version 0.91 10-Sep-97
5221 ----------------------
5222
5223 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5224 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5225
5226 2. PCRE was looping on encountering an indefinitely repeated back reference to
5227 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5228 Perl does - treats the match as successful.
5229
5230 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5