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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1370 - (show annotations)
Wed Oct 9 10:18:26 2013 UTC (6 years, 3 months ago) by ph10
File size: 253936 byte(s)
Add \o{} and tidy up \x{} handling. Minor update to RunTest.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.34 xx-xxxx-201x
5 --------------------------
6
7 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8 executable memory. Patch inspired by Carsten Klein.
9
10 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11 this macro is never tested and has no effect, because the work to support
12 coverage involves only compiling and linking options and special targets in
13 the Makefile. The comment in config.h implied that defining the macro would
14 enable coverage support, which is totally false. There was also support for
15 setting this macro in the CMake files (my fault, I just copied it from
16 configure). SUPPORT_GCOV has now been removed.
17
18 3. Make a small performance improvement in strlen16() and strlen32() in
19 pcretest.
20
21 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23 these cases). There was also one in pcretest.c.
24
25 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
26
27 6. In UTF mode, the code for checking whether a group could match an empty
28 string (which is used for indefinitely repeated groups to allow for
29 breaking an infinite loop) was broken when the group contained a repeated
30 negated single-character class with a character that occupied more than one
31 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
32 UTF-8 mode). The effect was undefined: the group might or might not be
33 deemed as matching an empty string, or the program might have crashed.
34
35 7. The code for checking whether a group could match an empty string was not
36 recognizing that \h, \H, \v, \V, and \R must match a character.
37
38 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
39 an empty string. If it can, pcretest shows this in its information output.
40
41 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
42 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
43 matched by pcre_exec() without using JIT:
44
45 (a) If the rest of the pattern did not match after a maximal run of
46 grapheme clusters, the code for backing up to try with fewer of them
47 did not always back up over a full grapheme when characters that do not
48 have the modifier quality were involved, e.g. Hangul syllables.
49
50 (b) If the match point in a subject started with modifier character, and
51 there was no match, the code could incorrectly back up beyond the match
52 point, and potentially beyond the first character in the subject,
53 leading to a segfault or an incorrect match result.
54
55 10. A conditional group with an assertion condition could lead to PCRE
56 recording an incorrect first data item for a match if no other first data
57 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58 first data item, and therefore matched "ca" after "c" instead of at the
59 start.
60
61 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
62 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
63
64 12. The source of pcregrep now includes z/OS-specific code so that it can be
65 compiled for z/OS as part of the special z/OS distribution.
66
67 13. Added the -T and -TM options to pcretest.
68
69 14. The code in pcre_compile.c for creating the table of named capturing groups
70 has been refactored. Instead of creating the table dynamically during the
71 actual compiling pass, the information is remembered during the pre-compile
72 pass (on the stack unless there are more than 20 named groups, in which
73 case malloc() is used) and the whole table is created before the actual
74 compile happens. This has simplified the code (it is now nearly 150 lines
75 shorter) and prepared the way for better handling of references to groups
76 with duplicate names.
77
78 15. A back reference to a named subpattern when there is more than one of the
79 same name now checks them in the order in which they appear in the pattern.
80 The first one that is set is used for the reference. Previously only the
81 first one was inspected. This change makes PCRE more compatible with Perl.
82
83 16. Unicode character properties were updated from Unicode 6.3.0.
84
85 17. The compile-time code for auto-possessification has been refactored, based
86 on a patch by Zoltan Herczeg. It now happens after instead of during
87 compilation. The code is cleaner, and more cases are handled. The option
88 PCRE_NO_AUTO_POSSESSIFY is added for testing purposes, and the -O and /O
89 options in pcretest are provided to set it.
90
91 18. The character VT has been added to the set of characters that match \s and
92 are generally treated as white space, following this same change in Perl
93 5.18. There is now no difference between "Perl space" and "POSIX space".
94
95 19. The code for checking named groups as conditions, either for being set or
96 for being recursed, has been refactored (this is related to 14 and 15
97 above). Processing unduplicated named groups should now be as fast at
98 numerical groups, and processing duplicated groups should be faster than
99 before.
100
101 20. Two patches to the CMake build system, by Alexander Barkov:
102
103 (1) Replace the "source" command by "." in CMakeLists.txt because
104 "source" is a bash-ism.
105
106 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
107 without these the CMake build does not work on Solaris.
108
109 21. Perl has changed its handling of \8 and \9. If there is no previously
110 encountered capturing group of those numbers, they are treated as the
111 literal characters 8 and 9 instead of a binary zero followed by the
112 literals. PCRE now does the same.
113
114 22. Following Perl, added \o{} to specify codepoints in octal, making it
115 possible to specify values greater than 0777 and also making them
116 unambiguous.
117
118 23. Perl now gives an error for missing closing braces after \x{... instead of
119 treating the string as literal. PCRE now does the same.
120
121 24. RunTest used to grumble if an inappropriate test was selected explicitly,
122 but just skip it when running all tests. This make it awkward to run ranges
123 of tests when one of them was inappropriate. Now it just skips any
124 inappropriate tests, as it always did when running all tests.
125
126
127
128 Version 8.33 28-May-2013
129 ------------------------
130
131 1. Added 'U' to some constants that are compared to unsigned integers, to
132 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
133 variables that are added to signed variables, to ensure the result is
134 signed and can be negated.
135
136 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
137 CMake config files.
138
139 3. Revise the creation of config.h.generic so that all boolean macros are
140 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
141 overriding via -D on the command line possible.
142
143 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
144 to unsigned int is reported to make a quite noticeable speed difference in
145 a specific Windows environment. Testing on Linux did also appear to show
146 some benefit (and it is clearly not harmful). Also fixed the definition of
147 Xop which should be unsigned.
148
149 5. Related to (4), changing the definition of the intermediate variable cc
150 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
151 improvements.
152
153 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
154 unnecessary spaces.
155
156 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
157 and later.
158
159 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
160
161 9. Optimizing fast_forward_start_bits in JIT.
162
163 10. Adding support for callouts in JIT, and fixing some issues revealed
164 during this work. Namely:
165
166 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
167
168 (b) Minimum length was not checked before the matching is started.
169
170 11. The value of capture_last that is passed to callouts was incorrect in some
171 cases when there was a capture on one path that was subsequently abandoned
172 after a backtrack. Also, the capture_last value is now reset after a
173 recursion, since all captures are also reset in this case.
174
175 12. The interpreter no longer returns the "too many substrings" error in the
176 case when an overflowing capture is in a branch that is subsequently
177 abandoned after a backtrack.
178
179 13. In the pathological case when an offset vector of size 2 is used, pcretest
180 now prints out the matched string after a yield of 0 or 1.
181
182 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
183 Only supported by the JIT compiler at the moment.
184
185 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
186
187 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
188 offset of the starting point of the matching process, provided the offsets
189 vector is large enough.
190
191 17. The \A escape now records a lookbehind value of 1, though its execution
192 does not actually inspect the previous character. This is to ensure that,
193 in partial multi-segment matching, at least one character from the old
194 segment is retained when a new segment is processed. Otherwise, if there
195 are no lookbehinds in the pattern, \A might match incorrectly at the start
196 of a new segment.
197
198 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
199
200 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
201 gives some modest performance improvement in 8-bit mode.
202
203 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
204 be expressed in certain programming languages using Universal Character
205 Names.
206
207 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
208 which points out that "non characters" are not "characters that may not
209 appear in Unicode strings" but rather "characters that are reserved for
210 internal use and have only local meaning".
211
212 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
213 there was a conditional group that depended on an assertion, if the
214 assertion was false, the callout that immediately followed the alternation
215 in the condition was skipped when pcre_exec() was used for matching.
216
217 23. Allow an explicit callout to be inserted before an assertion that is the
218 condition for a conditional group, for compatibility with automatic
219 callouts, which always insert a callout at this point.
220
221 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
222 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
223
224 25. (*PRUNE) is now supported by the JIT compiler.
225
226 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
227
228 27. Fix the case where there are two or more SKIPs with arguments that may be
229 ignored.
230
231 28. (*SKIP) is now supported by the JIT compiler.
232
233 29. (*THEN) is now supported by the JIT compiler.
234
235 30. Update RunTest with additional test selector options.
236
237 31. The way PCRE handles backtracking verbs has been changed in two ways.
238
239 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
240 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
241 backtracking. In some cases this makes it more Perl-compatible, but Perl's
242 rather obscure rules do not always do the same thing.
243
244 (2) Previously, backtracking verbs were confined within assertions. This is
245 no longer the case for positive assertions, except for (*ACCEPT). Again,
246 this sometimes improves Perl compatibility, and sometimes does not.
247
248 32. A number of tests that were in test 2 because Perl did things differently
249 have been moved to test 1, because either Perl or PCRE has changed, and
250 these tests are now compatible.
251
252 32. Backtracking control verbs are now handled in the same way in JIT and
253 interpreter.
254
255 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
256 contained a forward subroutine reference caused a compile error.
257
258 34. Auto-detect and optimize limited repetitions in JIT.
259
260 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
261 blocking (*UTF) etc.
262
263 36. In the interpreter, maximizing pattern repetitions for characters and
264 character types now use tail recursion, which reduces stack usage.
265
266 37. The value of the max lookbehind was not correctly preserved if a compiled
267 and saved regex was reloaded on a host of different endianness.
268
269 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
270 of the compiled pattern block, expand the flags field from 16 to 32 bits
271 because it was almost full.
272
273 39. Try madvise first before posix_madvise.
274
275 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
276 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
277 original change locked out all matches of empty strings. This has been
278 changed so that one match of an empty string per line is recognized.
279 Subsequent searches on the same line (for colouring or for --only-matching,
280 for example) do not recognize empty strings.
281
282 41. Applied a user patch to fix a number of spelling mistakes in comments.
283
284 42. Data lines longer than 65536 caused pcretest to crash.
285
286 43. Clarified the data type for length and startoffset arguments for pcre_exec
287 and pcre_dfa_exec in the function-specific man pages, where they were
288 explicitly stated to be in bytes, never having been updated. I also added
289 some clarification to the pcreapi man page.
290
291 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
292 a segmentation fault.
293
294
295 Version 8.32 30-November-2012
296 -----------------------------
297
298 1. Improved JIT compiler optimizations for first character search and single
299 character iterators.
300
301 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
302 Patch by Daniel Richard G.
303
304 3. Single character iterator optimizations in the JIT compiler.
305
306 4. Improved JIT compiler optimizations for character ranges.
307
308 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
309 Reported by Giuseppe D'Angelo.
310
311 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
312 of a line, was being set incorrectly in cases where .* appeared inside
313 atomic brackets at the start of a pattern, or where there was a subsequent
314 *PRUNE or *SKIP.
315
316 7. Improved instruction cache flush for POWER/PowerPC.
317 Patch by Daniel Richard G.
318
319 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
320 grep:
321
322 (a) There is now no limit to the number of patterns to be matched.
323
324 (b) An error is given if a pattern is too long.
325
326 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
327 are now supported.
328
329 (d) --exclude-from and --include-from (multiple use) have been added.
330
331 (e) Exclusions and inclusions now apply to all files and directories, not
332 just to those obtained from scanning a directory recursively.
333
334 (f) Multiple uses of -f and --file-list are now supported.
335
336 (g) In a Windows environment, the default for -d has been changed from
337 "read" (the GNU grep default) to "skip", because otherwise the presence
338 of a directory in the file list provokes an error.
339
340 (h) The documentation has been revised and clarified in places.
341
342 9. Improve the matching speed of capturing brackets.
343
344 10. Changed the meaning of \X so that it now matches a Unicode extended
345 grapheme cluster.
346
347 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
348 out POSIX threads when JIT support is configured.
349
350 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
351
352 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
353 structure could go wrong in environments where size_t is not the same size
354 as int.
355
356 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
357 set.
358
359 15. The EBCDIC support had decayed; later updates to the code had included
360 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
361 general tidy up of EBCDIC-related issues, and the documentation was also
362 not quite right. There is now a test that can be run on ASCII systems to
363 check some of the EBCDIC-related things (but is it not a full test).
364
365 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
366 in a small tidy to the code.
367
368 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
369
370 18. If the --only-matching (-o) option in pcregrep is specified multiple
371 times, each one causes appropriate output. For example, -o1 -o2 outputs the
372 substrings matched by the 1st and 2nd capturing parentheses. A separating
373 string can be specified by --om-separator (default empty).
374
375 19. Improving the first n character searches.
376
377 20. Turn case lists for horizontal and vertical white space into macros so that
378 they are defined only once.
379
380 21. This set of changes together give more compatible Unicode case-folding
381 behaviour for characters that have more than one other case when UCP
382 support is available.
383
384 (a) The Unicode property table now has offsets into a new table of sets of
385 three or more characters that are case-equivalent. The MultiStage2.py
386 script that generates these tables (the pcre_ucd.c file) now scans
387 CaseFolding.txt instead of UnicodeData.txt for character case
388 information.
389
390 (b) The code for adding characters or ranges of characters to a character
391 class has been abstracted into a generalized function that also handles
392 case-independence. In UTF-mode with UCP support, this uses the new data
393 to handle characters with more than one other case.
394
395 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
396 whose other case is greater than 256 are now correctly matched
397 caselessly. Previously, the high codepoint matched the low one, but not
398 vice versa.
399
400 (d) The processing of \h, \H, \v, and \ in character classes now makes use
401 of the new class addition function, using character lists defined as
402 macros alongside the case definitions of 20 above.
403
404 (e) Caseless back references now work with characters that have more than
405 one other case.
406
407 (f) General caseless matching of characters with more than one other case
408 is supported.
409
410 22. Unicode character properties were updated from Unicode 6.2.0
411
412 23. Improved CMake support under Windows. Patch by Daniel Richard G.
413
414 24. Add support for 32-bit character strings, and UTF-32
415
416 25. Major JIT compiler update (code refactoring and bugfixing).
417 Experimental Sparc 32 support is added.
418
419 26. Applied a modified version of Daniel Richard G's patch to create
420 pcre.h.generic and config.h.generic by "make" instead of in the
421 PrepareRelease script.
422
423 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
424 pcre_compile.c when checking for a zero character.
425
426 28. Introducing a native interface for JIT. Through this interface, the compiled
427 machine code can be directly executed. The purpose of this interface is to
428 provide fast pattern matching, so several sanity checks are not performed.
429 However, feature tests are still performed. The new interface provides
430 1.4x speedup compared to the old one.
431
432 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
433 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
434 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
435
436 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
437 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
438 the "old" RFC 2279). Instead, it ended up passing a negative length to
439 pcre_exec().
440
441 31. Add support for GCC's visibility feature to hide internal functions.
442
443 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
444 "unknown -C option" after outputting 0 or 1.
445
446 33. There is now support for generating a code coverage report for the test
447 suite in environments where gcc is the compiler and lcov is installed. This
448 is mainly for the benefit of the developers.
449
450 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
451 unaddressable using valgrind annotations, allowing valgrind to detect
452 invalid memory accesses. This is mainly for the benefit of the developers.
453
454 25. (*UTF) can now be used to start a pattern in any of the three libraries.
455
456 26. Give configure error if --enable-cpp but no C++ compiler found.
457
458
459 Version 8.31 06-July-2012
460 -------------------------
461
462 1. Fixing a wrong JIT test case and some compiler warnings.
463
464 2. Removed a bashism from the RunTest script.
465
466 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
467 to unsigned type, result still unsigned" that was given by an MS compiler
468 on encountering the code "-sizeof(xxx)".
469
470 4. Partial matching support is added to the JIT compiler.
471
472 5. Fixed several bugs concerned with partial matching of items that consist
473 of more than one character:
474
475 (a) /^(..)\1/ did not partially match "aba" because checking references was
476 done on an "all or nothing" basis. This also applied to repeated
477 references.
478
479 (b) \R did not give a hard partial match if \r was found at the end of the
480 subject.
481
482 (c) \X did not give a hard partial match after matching one or more
483 characters at the end of the subject.
484
485 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
486 a partial match for the string "\r".
487
488 (e) When newline was set to CRLF, the metacharacter "." did not recognize
489 a partial match for a CR character at the end of the subject string.
490
491 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
492 running pcretest, the text "(JIT)" added to the output whenever JIT is
493 actually used to run the match.
494
495 7. Individual JIT compile options can be set in pcretest by following -s+[+]
496 or /S+[+] with a digit between 1 and 7.
497
498 8. OP_NOT now supports any UTF character not just single-byte ones.
499
500 9. (*MARK) control verb is now supported by the JIT compiler.
501
502 10. The command "./RunTest list" lists the available tests without actually
503 running any of them. (Because I keep forgetting what they all are.)
504
505 11. Add PCRE_INFO_MAXLOOKBEHIND.
506
507 12. Applied a (slightly modified) user-supplied patch that improves performance
508 when the heap is used for recursion (compiled with --disable-stack-for-
509 recursion). Instead of malloc and free for each heap frame each time a
510 logical recursion happens, frames are retained on a chain and re-used where
511 possible. This sometimes gives as much as 30% improvement.
512
513 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
514 call.
515
516 14. As documented, (*COMMIT) is now confined to within a positive assertion.
517
518 15. It is now possible to link pcretest with libedit as an alternative to
519 libreadline.
520
521 16. (*COMMIT) control verb is now supported by the JIT compiler.
522
523 17. The Unicode data tables have been updated to Unicode 6.1.0.
524
525 18. Added --file-list option to pcregrep.
526
527 19. Added binary file support to pcregrep, including the -a, --binary-files,
528 -I, and --text options.
529
530 20. The madvise function is renamed for posix_madvise for QNX compatibility
531 reasons. Fixed by Giuseppe D'Angelo.
532
533 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
534
535 22. Changed the option for creating symbolic links for 16-bit man pages from
536 -s to -sf so that re-installing does not cause issues.
537
538 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
539
540 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
541 matches in certain environments (the workspace was not being correctly
542 retained). Also added to pcre_dfa_exec() a simple plausibility check on
543 some of the workspace data at the beginning of a restart.
544
545 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
546 was not doing so when it should - probably a typo introduced by SVN 528
547 (change 8.10/14).
548
549 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
550 \w+ when the character tables indicated that \x{c4} was a word character.
551 There were several related cases, all because the tests for doing a table
552 lookup were testing for characters less than 127 instead of 255.
553
554 27. If a pattern contains capturing parentheses that are not used in a match,
555 their slots in the ovector are set to -1. For those that are higher than
556 any matched groups, this happens at the end of processing. In the case when
557 there were back references that the ovector was too small to contain
558 (causing temporary malloc'd memory to be used during matching), and the
559 highest capturing number was not used, memory off the end of the ovector
560 was incorrectly being set to -1. (It was using the size of the temporary
561 memory instead of the true size.)
562
563 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
564 ovector size, it uses memory at the end of the block that it has got.
565
566 29. Check for an overlong MARK name and give an error at compile time. The
567 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
568
569 30. JIT compiler update.
570
571 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
572 Rill for the patch.
573
574 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
575
576 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
577
578 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
579 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
580 was enabled.
581
582 35. Improve JIT code generation for greedy plus quantifier.
583
584 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
585 1 to "aa" instead of to an empty string. The bug affected repeated groups
586 that could potentially match an empty string.
587
588 37. Optimizing single character iterators in JIT.
589
590 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
591 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
592 codepoints that are too big for the mode are faulted, and in a UTF mode,
593 disallowed codepoints are also faulted.
594
595 39. If PCRE was compiled with UTF support, in three places in the DFA
596 matcher there was code that should only have been obeyed in UTF mode, but
597 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
598 processing when bytes with values greater than 127 were present. In 16-bit
599 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
600 both cases the values are those that cannot be the first data item in a UTF
601 character. The three items that might have provoked this were recursions,
602 possessively repeated groups, and atomic groups.
603
604 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
605 and pcregrep, because some OS require shared objects to be explicitly
606 passed to ld, causing the link step to fail if they are not.
607
608 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
609 mode, patterns that started with \h* or \R* might be incorrectly matched.
610
611
612 Version 8.30 04-February-2012
613 -----------------------------
614
615 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
616 name is defined in ctype.h.
617
618 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
619 only in quite long subpatterns.
620
621 3. Removed the function pcre_info(), which has been obsolete and deprecated
622 since it was replaced by pcre_fullinfo() in February 2000.
623
624 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
625 match a (*MARK), and the match failed at the start of the subject, a
626 reference to memory before the start of the subject could occur. This bug
627 was introduced by fix 17 of release 8.21.
628
629 5. A reference to an unset group with zero minimum repetition was giving
630 totally wrong answers (in non-JavaScript-compatibility mode). For example,
631 /(another)?(\1?)test/ matched against "hello world test". This bug was
632 introduced in release 8.13.
633
634 6. Add support for 16-bit character strings (a large amount of work involving
635 many changes and refactorings).
636
637 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
638 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
639 from a file.
640
641 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
642 rounding is not applied in this particular case).
643
644 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
645 if they appear, or are escaped, in patterns.
646
647 10. Get rid of a number of -Wunused-but-set-variable warnings.
648
649 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
650 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
651 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
652 also returns the mark "x". This bug applied to capturing parentheses,
653 non-capturing parentheses, and atomic parentheses. It also applied to some
654 assertions.
655
656 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
657 information out of configure.ac instead of relying on pcre.h.generic, which
658 is not stored in the repository.
659
660 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
661 -lreadline.
662
663 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
664
665 15. Applied Graycode's patch to put the top-level frame on the stack rather
666 than the heap when not using the stack for recursion. This gives a
667 performance improvement in many cases when recursion is not deep.
668
669 16. Experimental code added to "pcretest -C" to output the stack frame size.
670
671
672 Version 8.21 12-Dec-2011
673 ------------------------
674
675 1. Updating the JIT compiler.
676
677 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
678 are added as well.
679
680 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
681 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
682 calling _pcre_jit_exec. Some extra comments are added.
683
684 4. (*MARK) settings inside atomic groups that do not contain any capturing
685 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
686 was introduced by change 18 for 8.20.
687
688 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
689 ECMA-262 standard.
690
691 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
692 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
693 This bug was probably introduced by change 9 of 8.13.
694
695 7. While fixing 6 above, I noticed that a number of other items were being
696 incorrectly rejected as "not fixed length". This arose partly because newer
697 opcodes had not been added to the fixed-length checking code. I have (a)
698 corrected the bug and added tests for these items, and (b) arranged for an
699 error to occur if an unknown opcode is encountered while checking for fixed
700 length instead of just assuming "not fixed length". The items that were
701 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
702 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
703 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
704
705 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
706 being incorrectly compiled and would have given unpredicatble results.
707
708 9. A possessively repeated subpattern with minimum repeat count greater than
709 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
710 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
711 the first (A) could occur when it should not.
712
713 10. Add a cast and remove a redundant test from the code.
714
715 11. JIT should use pcre_malloc/pcre_free for allocation.
716
717 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
718 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
719 is anything other than /usr, -L is still shown).
720
721 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
722
723 14. Perl does not support \N without a following name in a [] class; PCRE now
724 also gives an error.
725
726 15. If a forward reference was repeated with an upper limit of around 2000,
727 it caused the error "internal error: overran compiling workspace". The
728 maximum number of forward references (including repeats) was limited by the
729 internal workspace, and dependent on the LINK_SIZE. The code has been
730 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
731 the default depends on LINK_SIZE. There is a new upper limit (for safety)
732 of around 200,000 forward references. While doing this, I also speeded up
733 the filling in of repeated forward references.
734
735 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
736 incorrectly expecting the subject to contain another "a" after the start.
737
738 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
739 in the match, the SKIP should be ignored. This was not happening; instead
740 the SKIP was being treated as NOMATCH. For patterns such as
741 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
742 tested.
743
744 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
745 now much more compatible with Perl, in particular in cases where the result
746 is a non-match for a non-anchored pattern. For example, if
747 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
748 "m", where previously it did not return a name. A side effect of this
749 change is that for partial matches, the last encountered mark name is
750 returned, as for non matches. A number of tests that were previously not
751 Perl-compatible have been moved into the Perl-compatible test files. The
752 refactoring has had the pleasing side effect of removing one argument from
753 the match() function, thus reducing its stack requirements.
754
755 19. If the /S+ option was used in pcretest to study a pattern using JIT,
756 subsequent uses of /S (without +) incorrectly behaved like /S+.
757
758 21. Retrieve executable code size support for the JIT compiler and fixing
759 some warnings.
760
761 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
762 not work when the shorter character appeared right at the end of the
763 subject string.
764
765 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
766 systems.
767
768 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
769 output it when the /M option is used in pcretest.
770
771 25. The CheckMan script was not being included in the distribution. Also, added
772 an explicit "perl" to run Perl scripts from the PrepareRelease script
773 because this is reportedly needed in Windows.
774
775 26. If study data was being save in a file and studying had not found a set of
776 "starts with" bytes for the pattern, the data written to the file (though
777 never used) was taken from uninitialized memory and so caused valgrind to
778 complain.
779
780 27. Updated RunTest.bat as provided by Sheri Pierce.
781
782 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
783
784 29. Computation of memory usage for the table of capturing group names was
785 giving an unnecessarily large value.
786
787
788 Version 8.20 21-Oct-2011
789 ------------------------
790
791 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
792 a POSIX class. After further experiments with Perl, which convinced me that
793 Perl has bugs and confusions, a closing square bracket is no longer allowed
794 in a POSIX name. This bug also affected patterns with classes that started
795 with full stops.
796
797 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
798 captured substring, but while checking the failing first alternative,
799 substring 1 is temporarily captured. If the output vector supplied to
800 pcre_exec() was not big enough for this capture, the yield of the function
801 was still zero ("insufficient space for captured substrings"). This cannot
802 be totally fixed without adding another stack variable, which seems a lot
803 of expense for a edge case. However, I have improved the situation in cases
804 such as /(a)(b)x|abc/ matched against "abc", where the return code
805 indicates that fewer than the maximum number of slots in the ovector have
806 been set.
807
808 3. Related to (2) above: when there are more back references in a pattern than
809 slots in the output vector, pcre_exec() uses temporary memory during
810 matching, and copies in the captures as far as possible afterwards. It was
811 using the entire output vector, but this conflicts with the specification
812 that only 2/3 is used for passing back captured substrings. Now it uses
813 only the first 2/3, for compatibility. This is, of course, another edge
814 case.
815
816 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
817 main code base, and can be used by building with --enable-jit. When this is
818 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
819 runtime --no-jit option is given.
820
821 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
822 ovector, the return from the function was zero, implying that there were
823 other matches that did not fit. The correct "exactly full" value is now
824 returned.
825
826 6. If a subpattern that was called recursively or as a subroutine contained
827 (*PRUNE) or any other control that caused it to give a non-standard return,
828 invalid errors such as "Error -26 (nested recursion at the same subject
829 position)" or even infinite loops could occur.
830
831 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
832 computing the minimum length on reaching *ACCEPT, and so ended up with the
833 wrong value of 1 rather than 0. Further investigation indicates that
834 computing a minimum subject length in the presence of *ACCEPT is difficult
835 (think back references, subroutine calls), and so I have changed the code
836 so that no minimum is registered for a pattern that contains *ACCEPT.
837
838 8. If (*THEN) was present in the first (true) branch of a conditional group,
839 it was not handled as intended. [But see 16 below.]
840
841 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
842 Sheri Pierce.
843
844 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
845 the first byte in a match must be "a".
846
847 11. Change 17 for 8.13 increased the recursion depth for patterns like
848 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
849 pattern contains any instances of (*THEN). If it does not, the old
850 optimizations are restored. It would be nice to do this on a per-group
851 basis, but at the moment that is not feasible.
852
853 12. In some environments, the output of pcretest -C is CRLF terminated. This
854 broke RunTest's code that checks for the link size. A single white space
855 character after the value is now allowed for.
856
857 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
858 For "fr", it uses the Windows-specific input and output files.
859
860 14. If (*THEN) appeared in a group that was called recursively or as a
861 subroutine, it did not work as intended. [But see next item.]
862
863 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
864 pattern fragments (but not containing any | characters). If A and B are
865 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
866 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
867 D. In other words, Perl considers parentheses that do not contain any |
868 characters to be part of a surrounding alternative, whereas PCRE was
869 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
870 differently. PCRE now behaves in the same way as Perl, except in the case
871 of subroutine/recursion calls such as (?1) which have in any case always
872 been different (but PCRE had them first :-).
873
874 16. Related to 15 above: Perl does not treat the | in a conditional group as
875 creating alternatives. Such a group is treated in the same way as an
876 ordinary group without any | characters when processing (*THEN). PCRE has
877 been changed to match Perl's behaviour.
878
879 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
880 RunGrepTest script failed.
881
882 18. Change 22 for version 13 caused atomic groups to use more stack. This is
883 inevitable for groups that contain captures, but it can lead to a lot of
884 stack use in large patterns. The old behaviour has been restored for atomic
885 groups that do not contain any capturing parentheses.
886
887 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
888 suppress the check for a minimum subject length at run time. (If it was
889 given to pcre_exec() or pcre_dfa_exec() it did work.)
890
891 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
892 fail to work when decoding hex characters in data strings in EBCDIC
893 environments.
894
895 21. It appears that in at least one Mac OS environment, the isxdigit() function
896 is implemented as a macro that evaluates to its argument more than once,
897 contravening the C 90 Standard (I haven't checked a later standard). There
898 was an instance in pcretest which caused it to go wrong when processing
899 \x{...} escapes in subject strings. The has been rewritten to avoid using
900 things like p++ in the argument of isxdigit().
901
902
903 Version 8.13 16-Aug-2011
904 ------------------------
905
906 1. The Unicode data tables have been updated to Unicode 6.0.0.
907
908 2. Two minor typos in pcre_internal.h have been fixed.
909
910 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
911 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
912 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
913
914 4. There were a number of related bugs in the code for matching backrefences
915 caselessly in UTF-8 mode when codes for the characters concerned were
916 different numbers of bytes. For example, U+023A and U+2C65 are an upper
917 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
918 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
919 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
920 2-byte code at the end of the subject (it thought there wasn't enough data
921 left).
922
923 5. Comprehensive information about what went wrong is now returned by
924 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
925 as the output vector has at least 2 elements. The offset of the start of
926 the failing character and a reason code are placed in the vector.
927
928 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
929 now returned is for the first byte of the failing character, instead of the
930 last byte inspected. This is an incompatible change, but I hope it is small
931 enough not to be a problem. It makes the returned offset consistent with
932 pcre_exec() and pcre_dfa_exec().
933
934 7. pcretest now gives a text phrase as well as the error number when
935 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
936 failure, the offset and reason code are output.
937
938 8. When \R was used with a maximizing quantifier it failed to skip backwards
939 over a \r\n pair if the subsequent match failed. Instead, it just skipped
940 back over a single character (\n). This seems wrong (because it treated the
941 two characters as a single entity when going forwards), conflicts with the
942 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
943 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
944 has been changed.
945
946 9. Some internal refactoring has changed the processing so that the handling
947 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
948 time (the PCRE_DOTALL option was changed this way some time ago: version
949 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
950 which was always a bit of a fudge. It also means that there is one less
951 argument for the match() function, which reduces its stack requirements
952 slightly. This change also fixes an incompatibility with Perl: the pattern
953 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
954
955 10. More internal refactoring has drastically reduced the number of recursive
956 calls to match() for possessively repeated groups such as (abc)++ when
957 using pcre_exec().
958
959 11. While implementing 10, a number of bugs in the handling of groups were
960 discovered and fixed:
961
962 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
963 (a|)*(?1) gave a compile-time internal error.
964 ((a|)+)+ did not notice that the outer group could match an empty string.
965 (^a|^)+ was not marked as anchored.
966 (.*a|.*)+ was not marked as matching at start or after a newline.
967
968 12. Yet more internal refactoring has removed another argument from the match()
969 function. Special calls to this function are now indicated by setting a
970 value in a variable in the "match data" data block.
971
972 13. Be more explicit in pcre_study() instead of relying on "default" for
973 opcodes that mean there is no starting character; this means that when new
974 ones are added and accidentally left out of pcre_study(), testing should
975 pick them up.
976
977 14. The -s option of pcretest has been documented for ages as being an old
978 synonym of -m (show memory usage). I have changed it to mean "force study
979 for every regex", that is, assume /S for every regex. This is similar to -i
980 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
981 using it. It makes it easier to run collections of tests with and without
982 study enabled, and thereby test pcre_study() more easily. All the standard
983 tests are now run with and without -s (but some patterns can be marked as
984 "never study" - see 20 below).
985
986 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
987 restoration of the capturing data to the outer values was not happening
988 correctly.
989
990 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
991 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
992 pattern had matched an empty string, and so incorrectly returned a no
993 match.
994
995 17. There was optimizing code for the last branch of non-capturing parentheses,
996 and also for the obeyed branch of a conditional subexpression, which used
997 tail recursion to cut down on stack usage. Unfortunately, now that there is
998 the possibility of (*THEN) occurring in these branches, tail recursion is
999 no longer possible because the return has to be checked for (*THEN). These
1000 two optimizations have therefore been removed. [But see 8.20/11 above.]
1001
1002 18. If a pattern containing \R was studied, it was assumed that \R always
1003 matched two bytes, thus causing the minimum subject length to be
1004 incorrectly computed because \R can also match just one byte.
1005
1006 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1007 was incorrectly computed.
1008
1009 20. If /S is present twice on a test pattern in pcretest input, it now
1010 *disables* studying, thereby overriding the use of -s on the command line
1011 (see 14 above). This is necessary for one or two tests to keep the output
1012 identical in both cases.
1013
1014 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1015 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1016
1017 22. When an atomic group that contained a capturing parenthesis was
1018 successfully matched, but the branch in which it appeared failed, the
1019 capturing was not being forgotten if a higher numbered group was later
1020 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1021 group 1 to "a", when in fact it should be unset. This applied to multi-
1022 branched capturing and non-capturing groups, repeated or not, and also to
1023 positive assertions (capturing in negative assertions does not happen
1024 in PCRE) and also to nested atomic groups.
1025
1026 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1027 subject after a captured substring, to make it easier to tell which of a
1028 number of identical substrings has been captured.
1029
1030 24. The way atomic groups are processed by pcre_exec() has been changed so that
1031 if they are repeated, backtracking one repetition now resets captured
1032 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1033 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1034 "aaa". Previously, it would have been "a". As part of this code
1035 refactoring, the way recursive calls are handled has also been changed.
1036
1037 25. If an assertion condition captured any substrings, they were not passed
1038 back unless some other capturing happened later. For example, if
1039 (?(?=(a))a) was matched against "a", no capturing was returned.
1040
1041 26. When studying a pattern that contained subroutine calls or assertions,
1042 the code for finding the minimum length of a possible match was handling
1043 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1044 group 1 called group 2 while simultaneously a separate group 2 called group
1045 1). A stack overflow occurred in this case. I have fixed this by limiting
1046 the recursion depth to 10.
1047
1048 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1049 Fortmann. This supports explicit test numbers on the command line, and has
1050 argument validation and error reporting.
1051
1052 28. An instance of \X with an unlimited repeat could fail if at any point the
1053 first character it looked at was a mark character.
1054
1055 29. Some minor code refactoring concerning Unicode properties and scripts
1056 should reduce the stack requirement of match() slightly.
1057
1058 30. Added the '=' option to pcretest to check the setting of unused capturing
1059 slots at the end of the pattern, which are documented as being -1, but are
1060 not included in the return count.
1061
1062 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1063 compiled something random. Now it gives a compile-time error (as does
1064 Perl).
1065
1066 32. A *MARK encountered during the processing of a positive assertion is now
1067 recorded and passed back (compatible with Perl).
1068
1069 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1070 had alternative anchored branches, the search for a second match in a line
1071 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1072 matched the line "0102" twice. The same bug affected patterns that started
1073 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1074 twice.
1075
1076 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1077 does, and because of capturing effects, quantifying parenthesized
1078 assertions may at times be useful. Quantifiers are now allowed for
1079 parenthesized assertions.
1080
1081 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1082
1083 36. \g was being checked for fancy things in a character class, when it should
1084 just be a literal "g".
1085
1086 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1087 appearance of a nested POSIX class supersedes an apparent external class.
1088 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1089 unescaped square brackets may also appear as part of class names. For
1090 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1091 more like Perl. (But see 8.20/1 above.)
1092
1093 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1094 was because it thought it was \N{name}, which is not supported).
1095
1096 39. Add minix to OS list not supporting the -S option in pcretest.
1097
1098 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1099 cannot analyze patterns in sufficient detail to catch mutual recursions
1100 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1101 subgroup is called recursively as a subpattern for a second time at the
1102 same position in the subject string. In previous releases this might have
1103 been caught by the recursion limit, or it might have run out of stack.
1104
1105 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1106 happen only once. PCRE was, however incorrectly giving a compile time error
1107 "recursive call could loop indefinitely" because it cannot analyze the
1108 pattern in sufficient detail. The compile time test no longer happens when
1109 PCRE is compiling a conditional subpattern, but actual runaway loops are
1110 now caught at runtime (see 40 above).
1111
1112 42. It seems that Perl allows any characters other than a closing parenthesis
1113 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1114 has been changed to be the same.
1115
1116 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1117 as not to get warnings when autogen.sh is called. Also changed
1118 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1119
1120 44. To help people who use pcregrep to scan files containing exceedingly long
1121 lines, the following changes have been made:
1122
1123 (a) The default value of the buffer size parameter has been increased from
1124 8K to 20K. (The actual buffer used is three times this size.)
1125
1126 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1127 PCRE is built.
1128
1129 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1130 to be set at run time.
1131
1132 (d) Numerical values in pcregrep options can be followed by K or M, for
1133 example --buffer-size=50K.
1134
1135 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1136 given and the return code is set to 2.
1137
1138 45. Add a pointer to the latest mark to the callout data block.
1139
1140 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1141 partial match of an empty string instead of no match. This was specific to
1142 the use of ".".
1143
1144 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1145 complete match instead of a partial match. This bug was dependent on both
1146 the PCRE_UTF8 and PCRE_DOTALL options being set.
1147
1148 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1149 starting byte set, because \b was not being ignored.
1150
1151
1152 Version 8.12 15-Jan-2011
1153 ------------------------
1154
1155 1. Fixed some typos in the markup of the man pages, and wrote a script that
1156 checks for such things as part of the documentation building process.
1157
1158 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1159 --match-limit and --recursion-limit options (added for 8.11). In
1160 particular, this made one of the standard tests fail. (The integer value
1161 went into the wrong half of a long int.)
1162
1163 3. If the --colour option was given to pcregrep with -v (invert match), it
1164 did strange things, either producing crazy output, or crashing. It should,
1165 of course, ignore a request for colour when reporting lines that do not
1166 match.
1167
1168 4. Another pcregrep bug caused similar problems if --colour was specified with
1169 -M (multiline) and the pattern match finished with a line ending.
1170
1171 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1172 matched in multiline mode, the following line was shown as part of the
1173 match. This seems wrong, so I have changed it.
1174
1175 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1176 the check for further matches in the same line (so they could be coloured)
1177 to overrun the end of the current line. If another match was found, it was
1178 incorrectly shown (and then shown again when found in the next line).
1179
1180 7. If pcregrep was compiled under Windows, there was a reference to the
1181 function pcregrep_exit() before it was defined. I am assuming this was
1182 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1183 reported by a user. I've moved the definition above the reference.
1184
1185
1186 Version 8.11 10-Dec-2010
1187 ------------------------
1188
1189 1. (*THEN) was not working properly if there were untried alternatives prior
1190 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1191 backtracked to try for "b" instead of moving to the next alternative branch
1192 at the same level (in this case, to look for "c"). The Perl documentation
1193 is clear that when (*THEN) is backtracked onto, it goes to the "next
1194 alternative in the innermost enclosing group".
1195
1196 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1197 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1198 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1199 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1200 (*THEN).
1201
1202 3. If \s appeared in a character class, it removed the VT character from
1203 the class, even if it had been included by some previous item, for example
1204 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1205 of \s, but is part of the POSIX "space" class.)
1206
1207 4. A partial match never returns an empty string (because you can always
1208 match an empty string at the end of the subject); however the checking for
1209 an empty string was starting at the "start of match" point. This has been
1210 changed to the "earliest inspected character" point, because the returned
1211 data for a partial match starts at this character. This means that, for
1212 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1213 (previously it gave "no match").
1214
1215 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1216 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1217 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1218 has an implication that the given string is incomplete (because a partial
1219 match is preferred over a full match). For this reason, these items now
1220 give a partial match in this situation. [Aside: previously, the one case
1221 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1222 match rather than a full match, which was wrong by the old rules, but is
1223 now correct.]
1224
1225 6. There was a bug in the handling of #-introduced comments, recognized when
1226 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1227 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1228 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1229 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1230 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1231 places in pcre_compile().
1232
1233 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1234 comments when looking ahead for named forward references to subpatterns,
1235 the only newline sequence it recognized was NL. It now handles newlines
1236 according to the set newline convention.
1237
1238 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1239 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1240 cater for a lack of strerror(). These oversights have been fixed.
1241
1242 9. Added --match-limit and --recursion-limit to pcregrep.
1243
1244 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1245
1246 11. When the -o option was used, pcregrep was setting a return code of 1, even
1247 when matches were found, and --line-buffered was not being honoured.
1248
1249 12. Added an optional parentheses number to the -o and --only-matching options
1250 of pcregrep.
1251
1252 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1253 can match an empty string. The code to do it in pcretest and pcredemo
1254 needed fixing:
1255
1256 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1257 only one byte after an empty string match just before CRLF (this case
1258 just got forgotten; "any" and "anycrlf" were OK).
1259
1260 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1261 mode when an empty string match preceded an ASCII character followed by
1262 a non-ASCII character. (The code for advancing by one character rather
1263 than one byte was nonsense.)
1264
1265 (c) The pcredemo.c sample program did not have any code at all to handle
1266 the cases when CRLF is a valid newline sequence.
1267
1268 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1269 as a starting offset was within the subject string. There is now a new
1270 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1271 negative or greater than the length of the string. In order to test this,
1272 pcretest is extended to allow the setting of negative starting offsets.
1273
1274 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1275 starting offset points to the beginning of a UTF-8 character was
1276 unnecessarily clumsy. I tidied it up.
1277
1278 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1279 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1280
1281 17. Nobody had reported that the --include_dir option, which was added in
1282 release 7.7 should have been called --include-dir (hyphen, not underscore)
1283 for compatibility with GNU grep. I have changed it to --include-dir, but
1284 left --include_dir as an undocumented synonym, and the same for
1285 --exclude-dir, though that is not available in GNU grep, at least as of
1286 release 2.5.4.
1287
1288 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1289 characters from a string of bytes) have been redefined so as not to use
1290 loops, in order to improve performance in some environments. At the same
1291 time, I abstracted some of the common code into auxiliary macros to save
1292 repetition (this should not affect the compiled code).
1293
1294 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1295 compile-time error is now given if \c is not followed by an ASCII
1296 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1297 different, and any byte value is allowed.)
1298
1299 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1300 START_OPTIMIZE option, which is now allowed at compile time - but just
1301 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1302 to pcregrep and other applications that have no direct access to PCRE
1303 options. The new /Y option in pcretest sets this option when calling
1304 pcre_compile().
1305
1306 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1307 back references. Groups containing recursive back references were forced to
1308 be atomic by that change, but in the case of named groups, the amount of
1309 memory required was incorrectly computed, leading to "Failed: internal
1310 error: code overflow". This has been fixed.
1311
1312 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1313 pcretest.c, to avoid build problems in some Borland environments.
1314
1315
1316 Version 8.10 25-Jun-2010
1317 ------------------------
1318
1319 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1320 THEN.
1321
1322 2. (*ACCEPT) was not working when inside an atomic group.
1323
1324 3. Inside a character class, \B is treated as a literal by default, but
1325 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1326 causes the error). The code is unchanged, but I tidied the documentation.
1327
1328 4. Inside a character class, PCRE always treated \R and \X as literals,
1329 whereas Perl faults them if its -w option is set. I have changed PCRE so
1330 that it faults them when PCRE_EXTRA is set.
1331
1332 5. Added support for \N, which always matches any character other than
1333 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1334
1335 6. When compiling pcregrep with newer versions of gcc which may have
1336 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1337 declared with attribute warn_unused_result" were given. Just casting the
1338 result to (void) does not stop the warnings; a more elaborate fudge is
1339 needed. I've used a macro to implement this.
1340
1341 7. Minor change to pcretest.c to avoid a compiler warning.
1342
1343 8. Added four artifical Unicode properties to help with an option to make
1344 \s etc use properties (see next item). The new properties are: Xan
1345 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1346
1347 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1348 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1349 this option. Modified pcretest to add /W to test this facility. Added
1350 REG_UCP to make it available via the POSIX interface.
1351
1352 10. Added --line-buffered to pcregrep.
1353
1354 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1355 studied, and the match started with a letter with a code point greater than
1356 127 whose first byte was different to the first byte of the other case of
1357 the letter, the other case of this starting letter was not recognized
1358 (#976).
1359
1360 12. If a pattern that was studied started with a repeated Unicode property
1361 test, for example, \p{Nd}+, there was the theoretical possibility of
1362 setting up an incorrect bitmap of starting bytes, but fortunately it could
1363 not have actually happened in practice until change 8 above was made (it
1364 added property types that matched character-matching opcodes).
1365
1366 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1367 possible starting bytes for non-anchored patterns.
1368
1369 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1370 \R, and also a number of cases that involve Unicode properties, both
1371 explicit and implicit when PCRE_UCP is set.
1372
1373 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1374 input, it could crash or give wrong results if characters with values
1375 greater than 0xc0 were present in the subject string. (Detail: it assumed
1376 UTF-8 input when processing these items.)
1377
1378 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1379 size_t is 64-bit (#991).
1380
1381 17. Added a check for running out of memory when PCRE is compiled with
1382 --disable-stack-for-recursion (#990).
1383
1384 18. If the last data line in a file for pcretest does not have a newline on
1385 the end, a newline was missing in the output.
1386
1387 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1388 less than 128) in its various bitmaps. However, there is a facility for
1389 generating tables according to the current locale when PCRE is compiled. It
1390 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1391 characters, are recognized by isspace() and therefore were getting set in
1392 these tables, and indeed these tables seem to approximate to ISO 8859. This
1393 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1394 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1395 which of course cannot start UTF-8 characters. I have changed the code so
1396 that only real ASCII characters (less than 128) and the correct starting
1397 bytes for UTF-8 encodings are set for characters greater than 127 when in
1398 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1399 altogether.)
1400
1401 20. Added the /T option to pcretest so as to be able to run tests with non-
1402 standard character tables, thus making it possible to include the tests
1403 used for 19 above in the standard set of tests.
1404
1405 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1406 reference to a subpattern the other side of a comment that contains an
1407 opening parenthesis caused either an internal compiling error, or a
1408 reference to the wrong subpattern.
1409
1410
1411 Version 8.02 19-Mar-2010
1412 ------------------------
1413
1414 1. The Unicode data tables have been updated to Unicode 5.2.0.
1415
1416 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1417 configured.
1418
1419 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1420 original author of that file, following a query about its status.
1421
1422 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1423 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1424
1425 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1426 quantifier applied to a forward-referencing subroutine call, could compile
1427 incorrect code or give the error "internal error: previously-checked
1428 referenced subpattern not found".
1429
1430 6. Both MS Visual Studio and Symbian OS have problems with initializing
1431 variables to point to external functions. For these systems, therefore,
1432 pcre_malloc etc. are now initialized to local functions that call the
1433 relevant global functions.
1434
1435 7. There were two entries missing in the vectors called coptable and poptable
1436 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1437 I've fixed the data, and added a kludgy way of testing at compile time that
1438 the lengths are correct (equal to the number of opcodes).
1439
1440 8. Following on from 7, I added a similar kludge to check the length of the
1441 eint vector in pcreposix.c.
1442
1443 9. Error texts for pcre_compile() are held as one long string to avoid too
1444 much relocation at load time. To find a text, the string is searched,
1445 counting zeros. There was no check for running off the end of the string,
1446 which could happen if a new error number was added without updating the
1447 string.
1448
1449 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1450
1451 11. \K was not working if it appeared in an atomic group or in a group that
1452 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1453 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1454 the assertion is positive, but not if it is negative.
1455
1456 12. Change 11 fortuitously reduced the size of the stack frame used in the
1457 "match()" function of pcre_exec.c by one pointer. Forthcoming
1458 implementation of support for (*MARK) will need an extra pointer on the
1459 stack; I have reserved it now, so that the stack frame size does not
1460 decrease.
1461
1462 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1463 item in branch that calls a recursion is a subroutine call - as in the
1464 second branch in the above example - was incorrectly given the compile-
1465 time error "recursive call could loop indefinitely" because pcre_compile()
1466 was not correctly checking the subroutine for matching a non-empty string.
1467
1468 14. The checks for overrunning compiling workspace could trigger after an
1469 overrun had occurred. This is a "should never occur" error, but it can be
1470 triggered by pathological patterns such as hundreds of nested parentheses.
1471 The checks now trigger 100 bytes before the end of the workspace.
1472
1473 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1474
1475
1476 Version 8.01 19-Jan-2010
1477 ------------------------
1478
1479 1. If a pattern contained a conditional subpattern with only one branch (in
1480 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1481 computed the wrong minimum data length (which is of course zero for such
1482 subpatterns). This could cause incorrect "no match" results.
1483
1484 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1485 the pattern is reset in the first branch, pcre_compile() failed with
1486 "internal error: code overflow at offset...". This happened only when
1487 the reset was to the original external option setting. (An optimization
1488 abstracts leading options settings into an external setting, which was the
1489 cause of this.)
1490
1491 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1492 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1493 assertion pattern did not match (meaning that the assertion was true), it
1494 was incorrectly treated as false if the SKIP had been reached during the
1495 matching. This also applied to assertions used as conditions.
1496
1497 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1498 assertion subpattern, including such a pattern used as a condition,
1499 unpredictable results occurred, instead of the error return
1500 PCRE_ERROR_DFA_UITEM.
1501
1502 5. The C++ GlobalReplace function was not working like Perl for the special
1503 situation when an empty string is matched. It now does the fancy magic
1504 stuff that is necessary.
1505
1506 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1507 removed. (These were left over from very, very early versions of PCRE.)
1508
1509 7. Some cosmetic changes to the code to make life easier when compiling it
1510 as part of something else:
1511
1512 (a) Change DEBUG to PCRE_DEBUG.
1513
1514 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1515 called "current" as "current_branch", to prevent a collision with the
1516 Linux macro when compiled as a kernel module.
1517
1518 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1519 prevent a collision with the Linux macro when compiled as a kernel
1520 module.
1521
1522 8. In pcre_compile() there are some checks for integer overflows that used to
1523 cast potentially large values to (double). This has been changed to that
1524 when building, a check for int64_t is made, and if it is found, it is used
1525 instead, thus avoiding the use of floating point arithmetic. (There is no
1526 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1527 double.
1528
1529 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1530 2005 (difference between two addresses compared to an unsigned value).
1531
1532 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1533 custom one, because of the following reported problem in Windows:
1534
1535 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1536 under Win32.
1537 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1538 therefore missing the function definition.
1539 - The compiler thus generates a "C" signature for the test function.
1540 - The linker fails to find the "C" function.
1541 - PCRE fails to configure if asked to do so against libbz2.
1542
1543 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1544 messages were output:
1545
1546 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1547 rerunning libtoolize, to keep the correct libtool macros in-tree.
1548 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1549
1550 I have done both of these things.
1551
1552 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1553 most of the time, it *can* run out if it is given a pattern that contains a
1554 runaway infinite recursion. I updated the discussion in the pcrestack man
1555 page.
1556
1557 13. Now that we have gone to the x.xx style of version numbers, the minor
1558 version may start with zero. Using 08 or 09 is a bad idea because users
1559 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1560 interpreted as invalid octal numbers. I've updated the previous comment in
1561 configure.ac, and also added a check that gives an error if 08 or 09 are
1562 used.
1563
1564 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1565 causing partial matching to fail when the end of the subject matched \W
1566 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1567
1568 15. There were some discrepancies between the declarations in pcre_internal.h
1569 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1570 their definitions. The declarations used "const uschar *" and the
1571 definitions used USPTR. Even though USPTR is normally defined as "const
1572 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1573 reported that: "This difference in casting confuses some C++ compilers, for
1574 example, SunCC recognizes above declarations as different functions and
1575 generates broken code for hbpcre." I have changed the declarations to use
1576 USPTR.
1577
1578 16. GNU libtool is named differently on some systems. The autogen.sh script now
1579 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1580 (FreeBSD).
1581
1582 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1583 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1584 comment: "Figure out how to create a longlong from a string: strtoll and
1585 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1586 instance, but it only takes 2 args instead of 3!"
1587
1588 18. A subtle bug concerned with back references has been fixed by a change of
1589 specification, with a corresponding code fix. A pattern such as
1590 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1591 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1592 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1593 same bug. Such groups have to be quantified to be useful, or contained
1594 inside another quantified group. (If there's no repetition, the reference
1595 can never match.) The problem arises because, having left the group and
1596 moved on to the rest of the pattern, a later failure that backtracks into
1597 the group uses the captured value from the final iteration of the group
1598 rather than the correct earlier one. I have fixed this in PCRE by forcing
1599 any group that contains a reference to itself to be an atomic group; that
1600 is, there cannot be any backtracking into it once it has completed. This is
1601 similar to recursive and subroutine calls.
1602
1603
1604 Version 8.00 19-Oct-09
1605 ----------------------
1606
1607 1. The table for translating pcre_compile() error codes into POSIX error codes
1608 was out-of-date, and there was no check on the pcre_compile() error code
1609 being within the table. This could lead to an OK return being given in
1610 error.
1611
1612 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1613 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1614 in a Windows environment.
1615
1616 3. The pcregrep --count option prints the count for each file even when it is
1617 zero, as does GNU grep. However, pcregrep was also printing all files when
1618 --files-with-matches was added. Now, when both options are given, it prints
1619 counts only for those files that have at least one match. (GNU grep just
1620 prints the file name in this circumstance, but including the count seems
1621 more useful - otherwise, why use --count?) Also ensured that the
1622 combination -clh just lists non-zero counts, with no names.
1623
1624 4. The long form of the pcregrep -F option was incorrectly implemented as
1625 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1626 but it seems right to fix it, and I didn't think it was worth preserving
1627 the old behaviour.
1628
1629 5. The command line items --regex=pattern and --regexp=pattern were not
1630 recognized by pcregrep, which required --regex pattern or --regexp pattern
1631 (with a space rather than an '='). The man page documented the '=' forms,
1632 which are compatible with GNU grep; these now work.
1633
1634 6. No libpcreposix.pc file was created for pkg-config; there was just
1635 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1636
1637 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1638 when UCP support is not needed, by modifying the Python script that
1639 generates it from Unicode data files. This should not matter if the module
1640 is correctly used as a library, but I received one complaint about 50K of
1641 unwanted data. My guess is that the person linked everything into his
1642 program rather than using a library. Anyway, it does no harm.
1643
1644 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1645 was a minimum greater than 1 for a wide character in a possessive
1646 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1647 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1648 character. Chaos in the form of incorrect output or a compiling loop could
1649 result.
1650
1651 9. The restrictions on what a pattern can contain when partial matching is
1652 requested for pcre_exec() have been removed. All patterns can now be
1653 partially matched by this function. In addition, if there are at least two
1654 slots in the offset vector, the offset of the earliest inspected character
1655 for the match and the offset of the end of the subject are set in them when
1656 PCRE_ERROR_PARTIAL is returned.
1657
1658 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1659 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1660 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1661 and may be more useful for multi-segment matching.
1662
1663 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1664 used to be given if ever the end of the subject was reached; now it is
1665 given only if matching could not proceed because another character was
1666 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1667 string "Z", which now yields "no match" instead of "partial match". In the
1668 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1669 final character ended with (*FAIL).
1670
1671 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1672 if the pattern had a "must contain" character that was already found in the
1673 earlier partial match, unless partial matching was again requested. For
1674 example, with the pattern /dog.(body)?/, the "must contain" character is
1675 "g". If the first part-match was for the string "dog", restarting with
1676 "sbody" failed. This bug has been fixed.
1677
1678 13. The string returned by pcre_dfa_exec() after a partial match has been
1679 changed so that it starts at the first inspected character rather than the
1680 first character of the match. This makes a difference only if the pattern
1681 starts with a lookbehind assertion or \b or \B (\K is not supported by
1682 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1683 matching functions compatible, and I think it's the right thing to do.
1684
1685 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1686 so that the demonstration program is easily available in environments where
1687 PCRE has not been installed from source.
1688
1689 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1690 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1691 library.
1692
1693 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1694 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1695 is not the first non-POSIX option to be added. Clearly some people find
1696 these options useful.
1697
1698 17. If a caller to the POSIX matching function regexec() passes a non-zero
1699 value for nmatch with a NULL value for pmatch, the value of
1700 nmatch is forced to zero.
1701
1702 18. RunGrepTest did not have a test for the availability of the -u option of
1703 the diff command, as RunTest does. It now checks in the same way as
1704 RunTest, and also checks for the -b option.
1705
1706 19. If an odd number of negated classes containing just a single character
1707 interposed, within parentheses, between a forward reference to a named
1708 subpattern and the definition of the subpattern, compilation crashed with
1709 an internal error, complaining that it could not find the referenced
1710 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1711 [The bug was that it was starting one character too far in when skipping
1712 over the character class, thus treating the ] as data rather than
1713 terminating the class. This meant it could skip too much.]
1714
1715 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1716 /g option in pcretest when the pattern contains \K, which makes it possible
1717 to have an empty string match not at the start, even when the pattern is
1718 anchored. Updated pcretest and pcredemo to use this option.
1719
1720 21. If the maximum number of capturing subpatterns in a recursion was greater
1721 than the maximum at the outer level, the higher number was returned, but
1722 with unset values at the outer level. The correct (outer level) value is
1723 now given.
1724
1725 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1726 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1727 make it do so. The string so far is captured, making this feature
1728 compatible with Perl.
1729
1730 23. The tests have been re-organized, adding tests 11 and 12, to make it
1731 possible to check the Perl 5.10 features against Perl 5.10.
1732
1733 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1734 pattern matches a fixed length string. PCRE did not allow this; now it
1735 does. Neither allows recursion.
1736
1737 25. I finally figured out how to implement a request to provide the minimum
1738 length of subject string that was needed in order to match a given pattern.
1739 (It was back references and recursion that I had previously got hung up
1740 on.) This code has now been added to pcre_study(); it finds a lower bound
1741 to the length of subject needed. It is not necessarily the greatest lower
1742 bound, but using it to avoid searching strings that are too short does give
1743 some useful speed-ups. The value is available to calling programs via
1744 pcre_fullinfo().
1745
1746 26. While implementing 25, I discovered to my embarrassment that pcretest had
1747 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1748 study optimizations had never been tested with that matching function.
1749 Oops. What is worse, even when it was passed study data, there was a bug in
1750 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1751 were also very few tests of studied patterns with pcre_dfa_exec().
1752
1753 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1754 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1755 on the other side of the coin, they are no longer allowed to have different
1756 names, because these cannot be distinguished in PCRE, and this has caused
1757 confusion. (This is a difference from Perl.)
1758
1759 28. When duplicate subpattern names are present (necessarily with different
1760 numbers, as required by 27 above), and a test is made by name in a
1761 conditional pattern, either for a subpattern having been matched, or for
1762 recursion in such a pattern, all the associated numbered subpatterns are
1763 tested, and the overall condition is true if the condition is true for any
1764 one of them. This is the way Perl works, and is also more like the way
1765 testing by number works.
1766
1767
1768 Version 7.9 11-Apr-09
1769 ---------------------
1770
1771 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1772 (pcretest), all targets were linked against these libraries. This included
1773 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1774 libraries. This caused unwanted dependencies to be created. This problem
1775 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1776 pcretest is linked with readline.
1777
1778 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1779 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1780 moved outside it again, because FALSE and TRUE are already defined in AIX,
1781 but BOOL is not.
1782
1783 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1784 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1785
1786 4. The pcregrep documentation said spaces were inserted as well as colons (or
1787 hyphens) following file names and line numbers when outputting matching
1788 lines. This is not true; no spaces are inserted. I have also clarified the
1789 wording for the --colour (or --color) option.
1790
1791 5. In pcregrep, when --colour was used with -o, the list of matching strings
1792 was not coloured; this is different to GNU grep, so I have changed it to be
1793 the same.
1794
1795 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1796 each matching line was coloured. Now it goes on to look for further matches
1797 of any of the test patterns, which is the same behaviour as GNU grep.
1798
1799 7. A pattern that could match an empty string could cause pcregrep to loop; it
1800 doesn't make sense to accept an empty string match in pcregrep, so I have
1801 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1802 seems to be how GNU grep behaves. [But see later change 40 for release
1803 8.33.]
1804
1805 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1806 start or after a newline", because the conditional assertion was not being
1807 correctly handled. The rule now is that both the assertion and what follows
1808 in the first alternative must satisfy the test.
1809
1810 9. If auto-callout was enabled in a pattern with a conditional group whose
1811 condition was an assertion, PCRE could crash during matching, both with
1812 pcre_exec() and pcre_dfa_exec().
1813
1814 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1815 used for matching.
1816
1817 11. Unicode property support in character classes was not working for
1818 characters (bytes) greater than 127 when not in UTF-8 mode.
1819
1820 12. Added the -M command line option to pcretest.
1821
1822 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1823
1824 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1825
1826 16. Added comments and documentation about mis-use of no_arg in the C++
1827 wrapper.
1828
1829 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1830 from Martin Jerabek that uses macro names for all relevant character and
1831 string constants.
1832
1833 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1834 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1835 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1836 these, but not everybody uses configure.
1837
1838 19. A conditional group that had only one branch was not being correctly
1839 recognized as an item that could match an empty string. This meant that an
1840 enclosing group might also not be so recognized, causing infinite looping
1841 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1842 with the subject "ab", where knowledge that the repeated group can match
1843 nothing is needed in order to break the loop.
1844
1845 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1846 exec(), but without supplying a callout function, matching went wrong.
1847
1848 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1849 leak if the size of the offset vector was greater than 30. When the vector
1850 is smaller, the saved offsets during recursion go onto a local stack
1851 vector, but for larger vectors malloc() is used. It was failing to free
1852 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1853 error, in fact).
1854
1855 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1856 heapframe that is used only when UTF-8 support is enabled. This caused no
1857 problem, but was untidy.
1858
1859 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1860 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1861 included within another project.
1862
1863 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1864 slightly modified by me:
1865
1866 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1867 not building pcregrep.
1868
1869 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1870 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1871
1872 25. Forward references, both numeric and by name, in patterns that made use of
1873 duplicate group numbers, could behave incorrectly or give incorrect errors,
1874 because when scanning forward to find the reference group, PCRE was not
1875 taking into account the duplicate group numbers. A pattern such as
1876 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1877
1878 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1879 the feature of a custom pointer more persuasive (as requested by a user).
1880
1881 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1882 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1883 pre-defined.
1884
1885 28. Added support for (*UTF8) at the start of a pattern.
1886
1887 29. Arrange for flags added by the "release type" setting in CMake to be shown
1888 in the configuration summary.
1889
1890
1891 Version 7.8 05-Sep-08
1892 ---------------------
1893
1894 1. Replaced UCP searching code with optimized version as implemented for Ad
1895 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1896 stage table and inline lookup instead of a function, giving speed ups of 2
1897 to 5 times on some simple patterns that I tested. Permission was given to
1898 distribute the MultiStage2.py script that generates the tables (it's not in
1899 the tarball, but is in the Subversion repository).
1900
1901 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1902 scripts.
1903
1904 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1905 a group with a zero qualifier. The result of the study could be incorrect,
1906 or the function might crash, depending on the pattern.
1907
1908 4. Caseless matching was not working for non-ASCII characters in back
1909 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1910 It now works when Unicode Property Support is available.
1911
1912 5. In pcretest, an escape such as \x{de} in the data was always generating
1913 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1914 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1915 truncation.
1916
1917 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1918
1919 7. Added two (int) casts to pcregrep when printing the difference of two
1920 pointers, in case they are 64-bit values.
1921
1922 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1923 test 2 if it fails.
1924
1925 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1926 and a #define of that name to empty if it is not externally set. This is to
1927 allow users of MSVC to set it if necessary.
1928
1929 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1930 the convenience functions in the pcre_get.c source file.
1931
1932 11. An option change at the start of a pattern that had top-level alternatives
1933 could cause overwriting and/or a crash. This command provoked a crash in
1934 some environments:
1935
1936 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1937
1938 This potential security problem was recorded as CVE-2008-2371.
1939
1940 12. For a pattern where the match had to start at the beginning or immediately
1941 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1942 pcre_dfa_exec() could read past the end of the passed subject if there was
1943 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1944 pcretest so that it places the subject at the end of its malloc-ed buffer.
1945
1946 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1947 exec() might read past the end of the data buffer in UTF-8 mode.
1948
1949 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1950 the data contained the byte 0x85 as part of a UTF-8 character within its
1951 first line. This applied both to normal and DFA matching.
1952
1953 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1954 /^[^d]*?$/8 failed to match "abc".
1955
1956 16. Added a missing copyright notice to pcrecpp_internal.h.
1957
1958 17. Make it more clear in the documentation that values returned from
1959 pcre_exec() in ovector are byte offsets, not character counts.
1960
1961 18. Tidied a few places to stop certain compilers from issuing warnings.
1962
1963 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1964 supplied by Stefan Weber. I made a further small update for 7.8 because
1965 there is a change of source arrangements: the pcre_searchfuncs.c module is
1966 replaced by pcre_ucd.c.
1967
1968
1969 Version 7.7 07-May-08
1970 ---------------------
1971
1972 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1973 a string to a long long, pretend we don't even have a long long." This is
1974 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1975
1976 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1977 pre-7.6 versions, which defined a global no_arg variable instead of putting
1978 it in the RE class. (See also #8 below.)
1979
1980 3. Remove a line of dead code, identified by coverity and reported by Nuno
1981 Lopes.
1982
1983 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1984
1985 (1) The include/exclude patterns were being applied to the whole pathnames
1986 of files, instead of just to the final components.
1987
1988 (2) If there was more than one level of directory, the subdirectories were
1989 skipped unless they satisfied the include/exclude conditions. This is
1990 inconsistent with GNU grep (and could even be seen as contrary to the
1991 pcregrep specification - which I improved to make it absolutely clear).
1992 The action now is always to scan all levels of directory, and just
1993 apply the include/exclude patterns to regular files.
1994
1995 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1996 --exclude_dir in the tests to avoid scanning .svn directories.
1997
1998 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1999 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2000 doesn't support NULs in patterns.
2001
2002 7. Added some missing "const"s to declarations of static tables in
2003 pcre_compile.c and pcre_dfa_exec.c.
2004
2005 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2006 caused by fix #2 above. (Subsequently also a second patch to fix the
2007 first patch. And a third patch - this was a messy problem.)
2008
2009 9. Applied Craig's patch to remove the use of push_back().
2010
2011 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2012 matching function regexec().
2013
2014 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2015 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2016 references. PCRE supports relative numbers with this syntax (I don't think
2017 Oniguruma does).
2018
2019 12. Previously, a group with a zero repeat such as (...){0} was completely
2020 omitted from the compiled regex. However, this means that if the group
2021 was called as a subroutine from elsewhere in the pattern, things went wrong
2022 (an internal error was given). Such groups are now left in the compiled
2023 pattern, with a new opcode that causes them to be skipped at execution
2024 time.
2025
2026 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2027 to the way PCRE behaves:
2028
2029 (a) A lone ] character is dis-allowed (Perl treats it as data).
2030
2031 (b) A back reference to an unmatched subpattern matches an empty string
2032 (Perl fails the current match path).
2033
2034 (c) A data ] in a character class must be notated as \] because if the
2035 first data character in a class is ], it defines an empty class. (In
2036 Perl it is not possible to have an empty class.) The empty class []
2037 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2038 The negative empty class [^] matches any one character, independently
2039 of the DOTALL setting.
2040
2041 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2042 non-existent subpattern following a character class starting with ']' and
2043 containing () gave an internal compiling error instead of "reference to
2044 non-existent subpattern". Fortunately, when the pattern did exist, the
2045 compiled code was correct. (When scanning forwards to check for the
2046 existence of the subpattern, it was treating the data ']' as terminating
2047 the class, so got the count wrong. When actually compiling, the reference
2048 was subsequently set up correctly.)
2049
2050 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2051 it was being rejected as not supported by pcre_dfa_exec(), even though
2052 other assertions are supported. I have made pcre_dfa_exec() support
2053 (*FAIL).
2054
2055 16. The implementation of 13c above involved the invention of a new opcode,
2056 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2057 cannot be changed at match time, I realized I could make a small
2058 improvement to matching performance by compiling OP_ALLANY instead of
2059 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2060 on the OP_ANY path.
2061
2062 17. Compiling pcretest on Windows with readline support failed without the
2063 following two fixes: (1) Make the unistd.h include conditional on
2064 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2065
2066 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2067 ncurses library to be included for pcretest when ReadLine support is
2068 requested, but also to allow for it to be overridden. This patch came from
2069 Daniel Bergström.
2070
2071 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2072 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2073 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2074 spotting this.
2075
2076
2077 Version 7.6 28-Jan-08
2078 ---------------------
2079
2080 1. A character class containing a very large number of characters with
2081 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2082 overflow.
2083
2084 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2085 HAVE_LONG_LONG is not defined.
2086
2087 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2088 bring them up to date and include new features. This patch includes:
2089
2090 - Fixed PH's badly added libz and libbz2 support.
2091 - Fixed a problem with static linking.
2092 - Added pcredemo. [But later removed - see 7 below.]
2093 - Fixed dftables problem and added an option.
2094 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2095 HAVE_LONG_LONG.
2096 - Added readline support for pcretest.
2097 - Added an listing of the option settings after cmake has run.
2098
2099 4. A user submitted a patch to Makefile that makes it easy to create
2100 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2101 Makefile.am that cause it to include this special target, without
2102 affecting anything else. Note that the same mingw target plus all
2103 the other distribution libraries and programs are now supported
2104 when configuring with CMake (see 6 below) instead of with
2105 Configure/Make.
2106
2107 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2108 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2109 exported in the Windows port". It has not yet been confirmed that the patch
2110 solves the problem, but it does no harm.
2111
2112 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2113 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2114 with CMake, and also correct the comment about stack recursion.
2115
2116 7. Remove the automatic building of pcredemo from the ./configure system and
2117 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2118 of a program that users should build themselves after PCRE is installed, so
2119 building it automatically is not really right. What is more, it gave
2120 trouble in some build environments.
2121
2122 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2123
2124
2125 Version 7.5 10-Jan-08
2126 ---------------------
2127
2128 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2129 values in parens when parsing an RE using the C++ wrapper."
2130
2131 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2132 Characters greater than 255 were excluded from the class instead of being
2133 included.
2134
2135 3. The same bug as (2) above applied to negated POSIX classes such as
2136 [:^space:].
2137
2138 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2139 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2140 I have changed it.
2141
2142 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2143 first named subpattern) and a construct such as (?&a) would reference the
2144 first named subpattern whose name started with "a" (in other words, the
2145 length check was missing). Both these problems are fixed. "Subpattern name
2146 expected" is now given for (?&) (a zero-length name), and this patch also
2147 makes it give the same error for \k'' (previously it complained that that
2148 was a reference to a non-existent subpattern).
2149
2150 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2151 this is right because (?- can be followed by option settings as well as by
2152 digits. I have, however, made the messages clearer.
2153
2154 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2155 than the number used in the conditional) now cause a compile-time error.
2156 This is actually not compatible with Perl, which accepts such patterns, but
2157 treats the conditional as always being FALSE (as PCRE used to), but it
2158 seems to me that giving a diagnostic is better.
2159
2160 8. Change "alphameric" to the more common word "alphanumeric" in comments
2161 and messages.
2162
2163 9. Fix two occurrences of "backslash" in comments that should have been
2164 "backspace".
2165
2166 10. Remove two redundant lines of code that can never be obeyed (their function
2167 was moved elsewhere).
2168
2169 11. The program that makes PCRE's Unicode character property table had a bug
2170 which caused it to generate incorrect table entries for sequences of
2171 characters that have the same character type, but are in different scripts.
2172 It amalgamated them into a single range, with the script of the first of
2173 them. In other words, some characters were in the wrong script. There were
2174 thirteen such cases, affecting characters in the following ranges:
2175
2176 U+002b0 - U+002c1
2177 U+0060c - U+0060d
2178 U+0061e - U+00612
2179 U+0064b - U+0065e
2180 U+0074d - U+0076d
2181 U+01800 - U+01805
2182 U+01d00 - U+01d77
2183 U+01d9b - U+01dbf
2184 U+0200b - U+0200f
2185 U+030fc - U+030fe
2186 U+03260 - U+0327f
2187 U+0fb46 - U+0fbb1
2188 U+10450 - U+1049d
2189
2190 12. The -o option (show only the matching part of a line) for pcregrep was not
2191 compatible with GNU grep in that, if there was more than one match in a
2192 line, it showed only the first of them. It now behaves in the same way as
2193 GNU grep.
2194
2195 13. If the -o and -v options were combined for pcregrep, it printed a blank
2196 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2197 does the same. The return code can be used to tell if there were any
2198 non-matching lines.
2199
2200 14. Added --file-offsets and --line-offsets to pcregrep.
2201
2202 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2203 infinitely looping recursion. The bug was that positive lookaheads were not
2204 being skipped when checking for a possible empty match (negative lookaheads
2205 and both kinds of lookbehind were skipped).
2206
2207 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2208 inclusion of <windows.h> to before rather than after the definition of
2209 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2210
2211 17. Specifying a possessive quantifier with a specific limit for a Unicode
2212 character property caused pcre_compile() to compile bad code, which led at
2213 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2214 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2215 caused the error; without that there was no problem.
2216
2217 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2218
2219 19. Added --enable-pcretest-libreadline.
2220
2221 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2222 RE::GlobalReplace(). As a result, the number of replacements returned was
2223 double what it should be. I removed one of the increments, but Craig sent a
2224 later patch that removed the other one (the right fix) and added unit tests
2225 that check the return values (which was not done before).
2226
2227 21. Several CMake things:
2228
2229 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2230 the names libpcre and libpcreposix, not just pcre and pcreposix.
2231
2232 (2) The above change means that pcretest and pcregrep are now correctly
2233 linked with the newly-built libraries, not previously installed ones.
2234
2235 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2236
2237 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2238 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2239 UTF-8 newline character). The key issue is that the pattern starts .*;
2240 this means that the match must be either at the beginning, or after a
2241 newline. The bug was in the code for advancing after a failed match and
2242 checking that the new position followed a newline. It was not taking
2243 account of UTF-8 characters correctly.
2244
2245 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2246 character classes. PCRE was not treating the sequence [:...:] as a
2247 character class unless the ... were all letters. Perl, however, seems to
2248 allow any characters between [: and :], though of course it rejects as
2249 unknown any "names" that contain non-letters, because all the known class
2250 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2251 for example, whereas PCRE did not - it did not recognize a POSIX character
2252 class. This seemed a bit dangerous, so the code has been changed to be
2253 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2254 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2255 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2256 Perl does, and where it didn't before.
2257
2258 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2259 Windows environments %n is disabled by default.
2260
2261
2262 Version 7.4 21-Sep-07
2263 ---------------------
2264
2265 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2266 means that a class such as [\s] counted as "explicit reference to CR or
2267 LF". That isn't really right - the whole point of the change was to try to
2268 help when there was an actual mention of one of the two characters. So now
2269 the change happens only if \r or \n (or a literal CR or LF) character is
2270 encountered.
2271
2272 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2273 of both had grown to the point where there were only 3 bits left.
2274 Fortunately, there was spare space in the data structure, and so I have
2275 moved the internal flags into a new 16-bit field to free up more option
2276 bits.
2277
2278 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2279 but did not set the internal JCHANGED flag - either of these is enough to
2280 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2281 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2282 start sets both bits.
2283
2284 4. Added options (at build time, compile time, exec time) to change \R from
2285 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2286
2287 5. doc/pcresyntax.html was missing from the distribution.
2288
2289 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2290 compatibility, even though it is no longer used.
2291
2292 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2293 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2294 windows.h file is present (where different names are used). [This was
2295 reversed later after testing - see 16 below.]
2296
2297 8. Changed all #include <config.h> to #include "config.h". There were also
2298 some further <pcre.h> cases that I changed to "pcre.h".
2299
2300 9. When pcregrep was used with the --colour option, it missed the line ending
2301 sequence off the lines that it output.
2302
2303 10. It was pointed out to me that arrays of string pointers cause lots of
2304 relocations when a shared library is dynamically loaded. A technique of
2305 using a single long string with a table of offsets can drastically reduce
2306 these. I have refactored PCRE in four places to do this. The result is
2307 dramatic:
2308
2309 Originally: 290
2310 After changing UCP table: 187
2311 After changing error message table: 43
2312 After changing table of "verbs" 36
2313 After changing table of Posix names 22
2314
2315 Thanks to the folks working on Gregex for glib for this insight.
2316
2317 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2318 unicode-properties was also set.
2319
2320 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2321
2322 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2323 checked only for CRLF.
2324
2325 14. Added casts to pcretest.c to avoid compiler warnings.
2326
2327 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2328
2329 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2330 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2331 entirely. This removes changes made in 7 above.
2332
2333 17. The CMake files have been updated, and there is now more information about
2334 building with CMake in the NON-UNIX-USE document.
2335
2336
2337 Version 7.3 28-Aug-07
2338 ---------------------
2339
2340 1. In the rejigging of the build system that eventually resulted in 7.1, the
2341 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2342 brackets there is not right, since it causes compilers to look for an
2343 installed pcre.h, not the version that is in the source that is being
2344 compiled (which of course may be different). I have changed it back to:
2345
2346 #include "pcre.h"
2347
2348 I have a vague recollection that the change was concerned with compiling in
2349 different directories, but in the new build system, that is taken care of
2350 by the VPATH setting the Makefile.
2351
2352 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2353 when the subject happened to end in the byte 0x85 (e.g. if the last
2354 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2355 characters but of course it shouldn't be taken as a newline when it is part
2356 of another character. The bug was that, for an unlimited repeat of . in
2357 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2358 characters when looking for a newline.
2359
2360 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2361
2362 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2363 in debug output.
2364
2365 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2366 long printing in the pcrecpp unittest when running under MinGW.
2367
2368 6. ESC_K was left out of the EBCDIC table.
2369
2370 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2371 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2372 limit also applies to "virtual nesting" when a pattern is recursive, and in
2373 this case 1000 isn't so big. I have been able to remove this limit at the
2374 expense of backing off one optimization in certain circumstances. Normally,
2375 when pcre_exec() would call its internal match() function recursively and
2376 immediately return the result unconditionally, it uses a "tail recursion"
2377 feature to save stack. However, when a subpattern that can match an empty
2378 string has an unlimited repetition quantifier, it no longer makes this
2379 optimization. That gives it a stack frame in which to save the data for
2380 checking that an empty string has been matched. Previously this was taken
2381 from the 1000-entry workspace that had been reserved. So now there is no
2382 explicit limit, but more stack is used.
2383
2384 8. Applied Daniel's patches to solve problems with the import/export magic
2385 syntax that is required for Windows, and which was going wrong for the
2386 pcreposix and pcrecpp parts of the library. These were overlooked when this
2387 problem was solved for the main library.
2388
2389 9. There were some crude static tests to avoid integer overflow when computing
2390 the size of patterns that contain repeated groups with explicit upper
2391 limits. As the maximum quantifier is 65535, the maximum group length was
2392 set at 30,000 so that the product of these two numbers did not overflow a
2393 32-bit integer. However, it turns out that people want to use groups that
2394 are longer than 30,000 bytes (though not repeat them that many times).
2395 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2396 made it possible to implement the integer overflow checks in a much more
2397 dynamic way, which I have now done. The artificial limitation on group
2398 length has been removed - we now have only the limit on the total length of
2399 the compiled pattern, which depends on the LINK_SIZE setting.
2400
2401 10. Fixed a bug in the documentation for get/copy named substring when
2402 duplicate names are permitted. If none of the named substrings are set, the
2403 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2404 empty string.
2405
2406 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2407 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2408 because the ] is interpreted as the first data character and the
2409 terminating ] is not found. PCRE has been made compatible with Perl in this
2410 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2411 cause memory overwriting.
2412
2413 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2414 string has been matched (to stop an infinite loop). It was not recognizing
2415 a conditional subpattern that could match an empty string if that
2416 subpattern was within another subpattern. For example, it looped when
2417 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2418 condition was not nested. This bug has been fixed.
2419
2420 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2421 past the start of the subject in the presence of bytes with the top bit
2422 set, for example "\x8aBCD".
2423
2424 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2425 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2426
2427 14. Optimized (?!) to (*FAIL).
2428
2429 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2430 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2431 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2432 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2433 does: it's just the validity check that is more restrictive.
2434
2435 16. Inserted checks for integer overflows during escape sequence (backslash)
2436 processing, and also fixed erroneous offset values for syntax errors during
2437 backslash processing.
2438
2439 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2440 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2441
2442 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2443 caused an overrun.
2444
2445 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2446 something other than just ASCII characters) inside a group that had an
2447 unlimited repeat caused a loop at compile time (while checking to see
2448 whether the group could match an empty string).
2449
2450 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2451 [\P{Any}] did so. (Error in the code for printing property names.)
2452
2453 21. An orphan \E inside a character class could cause a crash.
2454
2455 22. A repeated capturing bracket such as (A)? could cause a wild memory
2456 reference during compilation.
2457
2458 23. There are several functions in pcre_compile() that scan along a compiled
2459 expression for various reasons (e.g. to see if it's fixed length for look
2460 behind). There were bugs in these functions when a repeated \p or \P was
2461 present in the pattern. These operators have additional parameters compared
2462 with \d, etc, and these were not being taken into account when moving along
2463 the compiled data. Specifically:
2464
2465 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2466 length.
2467
2468 (b) An item such as \pL+ within a repeated group could cause crashes or
2469 loops.
2470
2471 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2472 "reference to non-existent subpattern" error.
2473
2474 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2475
2476 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2477 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2478
2479 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2480
2481 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2482 character were causing crashes (broken optimization).
2483
2484 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2485 \p or \P) caused a compile-time loop.
2486
2487 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2488 break. For example, the unstudied pattern [\r\n]A does not match the string
2489 "\r\nA" because change 7.0/46 below moves the current point on by two
2490 characters after failing to match at the start. However, the pattern \nA
2491 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2492 the same is true. There doesn't seem any very clean way out of this, but
2493 what I have chosen to do makes the common cases work: PCRE now takes note
2494 of whether there can be an explicit match for \r or \n anywhere in the
2495 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2496 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2497 pattern has explicit CR or LF references.
2498
2499 29. Added (*CR) etc for changing newline setting at start of pattern.
2500
2501
2502 Version 7.2 19-Jun-07
2503 ---------------------
2504
2505 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2506 which is apparently normally available under Windows.
2507
2508 2. Re-jig the pcregrep tests with different newline settings in an attempt
2509 to make them independent of the local environment's newline setting.
2510
2511 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2512
2513 4. Some of the "internals" tests were previously cut out when the link size
2514 was not 2, because the output contained actual offsets. The recent new
2515 "Z" feature of pcretest means that these can be cut out, making the tests
2516 usable with all link sizes.
2517
2518 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2519 stack recursion. This gives a massive performance boost under BSD, but just
2520 a small improvement under Linux. However, it saves one field in the frame
2521 in all cases.
2522
2523 6. Added more features from the forthcoming Perl 5.10:
2524
2525 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2526 recursion call. It refers to the nth most recently opened parentheses.
2527
2528 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2529 to be opened parentheses.
2530
2531 (c) Conditions that refer to capturing parentheses can be specified
2532 relatively, for example, (?(-2)... or (?(+3)...
2533
2534 (d) \K resets the start of the current match so that everything before
2535 is not part of it.
2536
2537 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2538
2539 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2540 reference syntax.
2541
2542 (g) (?| introduces a group in which the numbering of parentheses in each
2543 alternative starts with the same number.
2544
2545 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2546
2547 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2548 PCRE_INFO_JCHANGED.
2549
2550 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2551 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2552 for detecting groups that can match an empty string.
2553
2554 9. A pattern with a very large number of alternatives (more than several
2555 hundred) was running out of internal workspace during the pre-compile
2556 phase, where pcre_compile() figures out how much memory will be needed. A
2557 bit of new cunning has reduced the workspace needed for groups with
2558 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2559 workspace instead of running out of the 4096 that are available.
2560
2561 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2562
2563 11. Applied patch from Google to remove an optimization that didn't quite work.
2564 The report of the bug said:
2565
2566 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2567 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2568 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2569
2570 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2571 it matched the wrong number of bytes.
2572
2573
2574 Version 7.1 24-Apr-07
2575 ---------------------
2576
2577 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2578 that is more "standard", making use of automake and other Autotools. There
2579 is some re-arrangement of the files and adjustment of comments consequent
2580 on this.
2581
2582 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2583 for recursive directory scanning broke on some systems because the files
2584 are not scanned in any specific order and on different systems the order
2585 was different. A call to "sort" has been inserted into RunGrepTest for the
2586 approprate test as a short-term fix. In the longer term there may be an
2587 alternative.
2588
2589 3. I had an email from Eric Raymond about problems translating some of PCRE's
2590 man pages to HTML (despite the fact that I distribute HTML pages, some
2591 people do their own conversions for various reasons). The problems
2592 concerned the use of low-level troff macros .br and .in. I have therefore
2593 removed all such uses from the man pages (some were redundant, some could
2594 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2595 HTML has been updated to handle .nf/.fi and to complain if it encounters
2596 .br or .in.
2597
2598 4. Updated comments in configure.ac that get placed in config.h.in and also
2599 arranged for config.h to be included in the distribution, with the name
2600 config.h.generic, for the benefit of those who have to compile without
2601 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2602
2603 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2604 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2605 makevp.bat for the current PCRE, using the additional files
2606 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2607
2608 6. A Windows user reported a minor discrepancy with test 2, which turned out
2609 to be caused by a trailing space on an input line that had got lost in his
2610 copy. The trailing space was an accident, so I've just removed it.
2611
2612 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2613 that is needed.
2614
2615 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2616 as "const" (a) because they are and (b) because it helps the PHP
2617 maintainers who have recently made a script to detect big data structures
2618 in the php code that should be moved to the .rodata section. I remembered
2619 to update Builducptable as well, so it won't revert if ucptable.h is ever
2620 re-created.
2621
2622 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2623 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2624 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2625 support is not required. This saves 1.5-2K of code, which is important in
2626 some applications.
2627
2628 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2629 so as not to refer to the tables, even though these functions will never be
2630 called when UTF-8 support is disabled. Otherwise there are problems with a
2631 shared library.
2632
2633 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2634
2635 (a) It was defining its arguments as char * instead of void *.
2636
2637 (b) It was assuming that all moves were upwards in memory; this was true
2638 a long time ago when I wrote it, but is no longer the case.
2639
2640 The emulated memove() is provided for those environments that have neither
2641 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2642 is clearly not the case, as these two bugs were recently reported.
2643
2644 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2645 and Detrail to create the HTML documentation, the .txt form of the man
2646 pages, and it removes trailing spaces from listed files. It also creates
2647 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2648 case, it wraps all the #defines with #ifndefs. This script should be run
2649 before "make dist".
2650
2651 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2652 with Unicode property support.
2653
2654 (a) For a maximizing quantifier, if the two different cases of the
2655 character were of different lengths in their UTF-8 codings (there are
2656 some cases like this - I found 11), and the matching function had to
2657 back up over a mixture of the two cases, it incorrectly assumed they
2658 were both the same length.
2659
2660 (b) When PCRE was configured to use the heap rather than the stack for
2661 recursion during matching, it was not correctly preserving the data for
2662 the other case of a UTF-8 character when checking ahead for a match
2663 while processing a minimizing repeat. If the check also involved
2664 matching a wide character, but failed, corruption could cause an
2665 erroneous result when trying to check for a repeat of the original
2666 character.
2667
2668 13. Some tidying changes to the testing mechanism:
2669
2670 (a) The RunTest script now detects the internal link size and whether there
2671 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2672 values substituted by "configure". (The RunGrepTest script already did
2673 this for UTF-8.) The configure.ac script no longer substitutes the
2674 relevant variables.
2675
2676 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2677 with length and offset values. This means that the output is different
2678 for different internal link sizes. Test 2 is skipped for link sizes
2679 other than 2 because of this, bypassing the problem. Unfortunately,
2680 there was also a test in test 3 (the locale tests) that used /B and
2681 failed for link sizes other than 2. Rather than cut the whole test out,
2682 I have added a new /Z option to pcretest that replaces the length and
2683 offset values with spaces. This is now used to make test 3 independent
2684 of link size. (Test 2 will be tidied up later.)
2685
2686 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2687 segmentation fault instead of returning the appropriate error message.
2688
2689 15. In multiline mode when the newline sequence was set to "any", the pattern
2690 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2691 This doesn't seem right; it now treats the CRLF combination as the line
2692 ending, and so does not match in that case. It's only a pattern such as ^$
2693 that would hit this one: something like ^ABC$ would have failed after \r
2694 and then tried again after \r\n.
2695
2696 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2697 in an attempt to make files that differ only in their line terminators
2698 compare equal. This works on Linux.
2699
2700 17. Under certain error circumstances pcregrep might try to free random memory
2701 as it exited. This is now fixed, thanks to valgrind.
2702
2703 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2704 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2705 was because its rules for how to advance for /g after matching an empty
2706 string at the end of a line did not allow for this case. They now check for
2707 it specially.
2708
2709 20. pcretest is supposed to handle patterns and data of any length, by
2710 extending its buffers when necessary. It was getting this wrong when the
2711 buffer for a data line had to be extended.
2712
2713 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2714 CRLF as a newline sequence.
2715
2716 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2717 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2718 I have nevertheless tidied it up.
2719
2720 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2721
2722 24. Added a man page for pcre-config.
2723
2724
2725 Version 7.0 19-Dec-06
2726 ---------------------
2727
2728 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2729 moving to gcc 4.1.1.
2730
2731 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2732 sys/time.h, which is documented as needed for this function. It doesn't
2733 seem to matter on Linux, but it showed up on some releases of OS X.
2734
2735 3. It seems that there are systems where bytes whose values are greater than
2736 127 match isprint() in the "C" locale. The "C" locale should be the
2737 default when a C program starts up. In most systems, only ASCII printing
2738 characters match isprint(). This difference caused the output from pcretest
2739 to vary, making some of the tests fail. I have changed pcretest so that:
2740
2741 (a) When it is outputting text in the compiled version of a pattern, bytes
2742 other than 32-126 are always shown as hex escapes.
2743
2744 (b) When it is outputting text that is a matched part of a subject string,
2745 it does the same, unless a different locale has been set for the match
2746 (using the /L modifier). In this case, it uses isprint() to decide.
2747
2748 4. Fixed a major bug that caused incorrect computation of the amount of memory
2749 required for a compiled pattern when options that changed within the
2750 pattern affected the logic of the preliminary scan that determines the
2751 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2752 that the computed length was too small. The symptoms of this bug were
2753 either the PCRE error "internal error: code overflow" from pcre_compile(),
2754 or a glibc crash with a message such as "pcretest: free(): invalid next
2755 size (fast)". Examples of patterns that provoked this bug (shown in
2756 pcretest format) are:
2757
2758 /(?-x: )/x
2759 /(?x)(?-x: \s*#\s*)/
2760 /((?i)[\x{c0}])/8
2761 /(?i:[\x{c0}])/8
2762
2763 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2764 is now done differently.
2765
2766 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2767 wrapper classes; (b) implement a new function in the C++ scanner that is
2768 more efficient than the old way of doing things because it avoids levels of
2769 recursion in the regex matching; (c) add a paragraph to the documentation
2770 for the FullMatch() function.
2771
2772 6. The escape sequence \n was being treated as whatever was defined as
2773 "newline". Not only was this contrary to the documentation, which states
2774 that \n is character 10 (hex 0A), but it also went horribly wrong when
2775 "newline" was defined as CRLF. This has been fixed.
2776
2777 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2778 was being set to -1 for the "end of line" case (supposedly a value that no
2779 character can have). Though this value is never used (the check for end of
2780 line is "zero bytes in current character"), it caused compiler complaints.
2781 I've changed it to 0xffffffff.
2782
2783 8. In pcre_version.c, the version string was being built by a sequence of
2784 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2785 string (as it is for production releases) called a macro with an empty
2786 argument. The C standard says the result of this is undefined. The gcc
2787 compiler treats it as an empty string (which was what was wanted) but it is
2788 reported that Visual C gives an error. The source has been hacked around to
2789 avoid this problem.
2790
2791 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2792 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2793 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2794 of them did).
2795
2796 10. Originally, pcretest opened its input and output without "b"; then I was
2797 told that "b" was needed in some environments, so it was added for release
2798 5.0 to both the input and output. (It makes no difference on Unix-like
2799 systems.) Later I was told that it is wrong for the input on Windows. I've
2800 now abstracted the modes into two macros, to make it easier to fiddle with
2801 them, and removed "b" from the input mode under Windows.
2802
2803 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2804
2805 12. Added -help and --help to pcretest as an official way of being reminded
2806 of the options.
2807
2808 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2809 and pcrecpp.cc because they annoy compilers at high warning levels.
2810
2811 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2812
2813 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2814 scripts are not C programs :-) and which was not noticed because it works
2815 on Linux.
2816
2817 16. pcretest is supposed to handle any length of pattern and data line (as one
2818 line or as a continued sequence of lines) by extending its input buffer if
2819 necessary. This feature was broken for very long pattern lines, leading to
2820 a string of junk being passed to pcre_compile() if the pattern was longer
2821 than about 50K.
2822
2823 17. I have done a major re-factoring of the way pcre_compile() computes the
2824 amount of memory needed for a compiled pattern. Previously, there was code
2825 that made a preliminary scan of the pattern in order to do this. That was
2826 OK when PCRE was new, but as the facilities have expanded, it has become
2827 harder and harder to keep it in step with the real compile phase, and there
2828 have been a number of bugs (see for example, 4 above). I have now found a
2829 cunning way of running the real compile function in a "fake" mode that
2830 enables it to compute how much memory it would need, while actually only
2831 ever using a few hundred bytes of working memory and without too many
2832 tests of the mode. This should make future maintenance and development
2833 easier. A side effect of this work is that the limit of 200 on the nesting
2834 depth of parentheses has been removed (though this was never a serious
2835 limitation, I suspect). However, there is a downside: pcre_compile() now
2836 runs more slowly than before (30% or more, depending on the pattern). I
2837 hope this isn't a big issue. There is no effect on runtime performance.
2838
2839 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2840 newline (only possible for the last line of a file) and it was a
2841 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2842
2843 19. Added additional timing features to pcretest. (1) The -tm option now times
2844 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2845 separate command line item, by a number that specifies the number of
2846 repeats to use when timing. The default is 50000; this gives better
2847 precision, but takes uncomfortably long for very large patterns.
2848
2849 20. Extended pcre_study() to be more clever in cases where a branch of a
2850 subpattern has no definite first character. For example, (a*|b*)[cd] would
2851 previously give no result from pcre_study(). Now it recognizes that the
2852 first character must be a, b, c, or d.
2853
2854 21. There was an incorrect error "recursive call could loop indefinitely" if
2855 a subpattern (or the entire pattern) that was being tested for matching an
2856 empty string contained only one non-empty item after a nested subpattern.
2857 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2858 incorrectly, because the \d was being skipped in the check.
2859
2860 22. The pcretest program now has a new pattern option /B and a command line
2861 option -b, which is equivalent to adding /B to every pattern. This causes
2862 it to show the compiled bytecode, without the additional information that
2863 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2864 is the same as /B/I).
2865
2866 23. A new optimization is now able automatically to treat some sequences such
2867 as a*b as a*+b. More specifically, if something simple (such as a character
2868 or a simple class like \d) has an unlimited quantifier, and is followed by
2869 something that cannot possibly match the quantified thing, the quantifier
2870 is automatically "possessified".
2871
2872 24. A recursive reference to a subpattern whose number was greater than 39
2873 went wrong under certain circumstances in UTF-8 mode. This bug could also
2874 have affected the operation of pcre_study().
2875
2876 25. Realized that a little bit of performance could be had by replacing
2877 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2878
2879 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2880
2881 27. Possessive quantifiers such as a++ were previously implemented by turning
2882 them into atomic groups such as ($>a+). Now they have their own opcodes,
2883 which improves performance. This includes the automatically created ones
2884 from 23 above.
2885
2886 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2887 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2888 the first matched character to be a colon. This applied both to named and
2889 numbered groups.
2890
2891 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2892
2893 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2894 building PCRE on Windows easier, so I have included it in the distribution.
2895
2896 31. There is now a check in pcretest against a ridiculously large number being
2897 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2898 loop, the loop is abandoned.
2899
2900 32. Forward references to subpatterns in conditions such as (?(2)...) where
2901 subpattern 2 is defined later cause pcre_compile() to search forwards in
2902 the pattern for the relevant set of parentheses. This search went wrong
2903 when there were unescaped parentheses in a character class, parentheses
2904 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2905
2906 33. "Subroutine" calls and backreferences were previously restricted to
2907 referencing subpatterns earlier in the regex. This restriction has now
2908 been removed.
2909
2910 34. Added a number of extra features that are going to be in Perl 5.10. On the
2911 whole, these are just syntactic alternatives for features that PCRE had
2912 previously implemented using the Python syntax or my own invention. The
2913 other formats are all retained for compatibility.
2914
2915 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2916 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2917 also .NET compatible.
2918
2919 (b) A recursion or subroutine call to a named group can now be defined as
2920 (?&name) as well as (?P>name).
2921
2922 (c) A backreference to a named group can now be defined as \k<name> or
2923 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2924 5.10, are also .NET compatible.
2925
2926 (d) A conditional reference to a named group can now use the syntax
2927 (?(<name>) or (?('name') as well as (?(name).
2928
2929 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2930 groups (named and numbered) that are never evaluated inline, but can be
2931 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2932 is always false. There may be only one alternative in such a group.
2933
2934 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2935 as the simple (?(R). The condition is true only if the most recent
2936 recursion is that of the given number or name. It does not search out
2937 through the entire recursion stack.
2938
2939 (g) The escape \gN or \g{N} has been added, where N is a positive or
2940 negative number, specifying an absolute or relative reference.
2941
2942 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2943 some "unreachable code" warnings.
2944
2945 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2946 things, this adds five new scripts.
2947
2948 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2949 There were also incompatibilities regarding the handling of \Q..\E inside
2950 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2951 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2952
2953 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2954 matches an empty string, and forcibly breaks the loop. There were bugs in
2955 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2956 against aaaa the result was just "a" rather than "aaaa", for example. Two
2957 separate and independent bugs (that affected different cases) have been
2958 fixed.
2959
2960 39. Refactored the code to abolish the use of different opcodes for small
2961 capturing bracket numbers. This is a tidy that I avoided doing when I
2962 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2963 The new approach is not only tidier, it makes it possible to reduce the
2964 memory needed to fix the previous bug (38).
2965
2966 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2967 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2968 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2969 mode.
2970
2971 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2972 report.
2973
2974 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2975 copy construction and assignment in the C++ wrapper.
2976
2977 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2978 couple of bugs that could have given wrong results in the "--newline=crlf"
2979 case.
2980
2981 44. Added a number of casts and did some reorganization of signed/unsigned int
2982 variables following suggestions from Dair Grant. Also renamed the variable
2983 "this" as "item" because it is a C++ keyword.
2984
2985 45. Arranged for dftables to add
2986
2987 #include "pcre_internal.h"
2988
2989 to pcre_chartables.c because without it, gcc 4.x may remove the array
2990 definition from the final binary if PCRE is built into a static library and
2991 dead code stripping is activated.
2992
2993 46. For an unanchored pattern, if a match attempt fails at the start of a
2994 newline sequence, and the newline setting is CRLF or ANY, and the next two
2995 characters are CRLF, advance by two characters instead of one.
2996
2997
2998 Version 6.7 04-Jul-06
2999 ---------------------
3000
3001 1. In order to handle tests when input lines are enormously long, pcretest has
3002 been re-factored so that it automatically extends its buffers when
3003 necessary. The code is crude, but this _is_ just a test program. The
3004 default size has been increased from 32K to 50K.
3005
3006 2. The code in pcre_study() was using the value of the re argument before
3007 testing it for NULL. (Of course, in any sensible call of the function, it
3008 won't be NULL.)
3009
3010 3. The memmove() emulation function in pcre_internal.h, which is used on
3011 systems that lack both memmove() and bcopy() - that is, hardly ever -
3012 was missing a "static" storage class specifier.
3013
3014 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3015 containing an extended class (one that cannot be represented by a bitmap
3016 because it contains high-valued characters or Unicode property items, e.g.
3017 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3018 pattern, but PCRE should not loop if you do not (it no longer does).
3019 [Detail: two cases were found: (a) a repeated subpattern containing an
3020 extended class; (b) a recursive reference to a subpattern that followed a
3021 previous extended class. It wasn't skipping over the extended class
3022 correctly when UTF-8 mode was not set.]
3023
3024 5. A negated single-character class was not being recognized as fixed-length
3025 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3026 compile error "lookbehind assertion is not fixed length".
3027
3028 6. The RunPerlTest auxiliary script was showing an unexpected difference
3029 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3030 write a Perl script that can interpret lines of an input file either as
3031 byte characters or as UTF-8, which is what "perltest" was being required to
3032 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3033 can't do is switch easily at run time between having the "use utf8;" pragma
3034 or not. In the end, I fudged it by using the RunPerlTest script to insert
3035 "use utf8;" explicitly for the UTF-8 tests.
3036
3037 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3038 the end of the subject string, contrary to the documentation and to what
3039 Perl does. This was true of both matching functions. Now it matches only at
3040 the start of the subject and immediately after *internal* newlines.
3041
3042 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3043 a pointer to an int instead of a pointer to an unsigned long int. This
3044 caused problems on 64-bit systems.
3045
3046 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3047 instance of the 'standard' template library not being so standard".
3048
3049 10. There was no check on the number of named subpatterns nor the maximum
3050 length of a subpattern name. The product of these values is used to compute
3051 the size of the memory block for a compiled pattern. By supplying a very
3052 long subpattern name and a large number of named subpatterns, the size
3053 computation could be caused to overflow. This is now prevented by limiting
3054 the length of names to 32 characters, and the number of named subpatterns
3055 to 10,000.
3056
3057 11. Subpatterns that are repeated with specific counts have to be replicated in
3058 the compiled pattern. The size of memory for this was computed from the
3059 length of the subpattern and the repeat count. The latter is limited to
3060 65535, but there was no limit on the former, meaning that integer overflow
3061 could in principle occur. The compiled length of a repeated subpattern is
3062 now limited to 30,000 bytes in order to prevent this.
3063
3064 12. Added the optional facility to have named substrings with the same name.
3065
3066 13. Added the ability to use a named substring as a condition, using the
3067 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3068 are numbers (not recommended). Forward references are permitted.
3069
3070 14. Added forward references in named backreferences (if you see what I mean).
3071
3072 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3073 pattern could run off the end of the subject. For example, the pattern
3074 "(?s)(.{1,5})"8 did this with the subject "ab".
3075
3076 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3077 PCRE_CASELESS was set when matching characters that were quantified with ?
3078 or *.
3079
3080 17. A character class other than a single negated character that had a minimum
3081 but no maximum quantifier - for example [ab]{6,} - was not handled
3082 correctly by pce_dfa_exec(). It would match only one character.
3083
3084 18. A valid (though odd) pattern that looked like a POSIX character
3085 class but used an invalid character after [ (for example [[,abc,]]) caused
3086 pcre_compile() to give the error "Failed: internal error: code overflow" or
3087 in some cases to crash with a glibc free() error. This could even happen if
3088 the pattern terminated after [[ but there just happened to be a sequence of
3089 letters, a binary zero, and a closing ] in the memory that followed.
3090
3091 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3092 over the years. Originally (before any Unicode support), just the bottom 8
3093 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3094 output from "man perlunicode" includes this:
3095
3096 The regular expression compiler produces polymorphic opcodes. That
3097 is, the pattern adapts to the data and automatically switches to
3098 the Unicode character scheme when presented with Unicode data--or
3099 instead uses a traditional byte scheme when presented with byte
3100 data.
3101
3102 Sadly, a wide octal escape does not cause a switch, and in a string with
3103 no other multibyte characters, these octal escapes are treated as before.
3104 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3105 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3106 Unicode string.
3107
3108 I have not perpetrated such confusion in PCRE. Up till now, it took just
3109 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3110 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3111 translate to the appropriate multibyte character.
3112
3113 29. Applied some refactoring to reduce the number of warnings from Microsoft
3114 and Borland compilers. This has included removing the fudge introduced
3115 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3116 a warning about an unused variable.
3117
3118 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3119 characters since release 4.0, because Perl (from release 5.004) does not.
3120 [Or at least, is documented not to: some releases seem to be in conflict
3121 with the documentation.] However, when a pattern was studied with
3122 pcre_study() and all its branches started with \s, PCRE still included VT
3123 as a possible starting character. Of course, this did no harm; it just
3124 caused an unnecessary match attempt.
3125
3126 22. Removed a now-redundant internal flag bit that recorded the fact that case
3127 dependency changed within the pattern. This was once needed for "required
3128 byte" processing, but is no longer used. This recovers a now-scarce options
3129 bit. Also moved the least significant internal flag bit to the most-
3130 significant bit of the word, which was not previously used (hangover from
3131 the days when it was an int rather than a uint) to free up another bit for
3132 the future.
3133
3134 23. Added support for CRLF line endings as well as CR and LF. As well as the
3135 default being selectable at build time, it can now be changed at runtime
3136 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3137 specify that it is scanning data with non-default line endings.
3138
3139 24. Changed the definition of CXXLINK to make it agree with the definition of
3140 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3141
3142 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3143 recursions. This makes a big different to stack usage for some patterns.
3144
3145 26. If a subpattern containing a named recursion or subroutine reference such
3146 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3147 the space required for the compiled pattern went wrong and gave too small a
3148 value. Depending on the environment, this could lead to "Failed: internal
3149 error: code overflow at offset 49" or "glibc detected double free or
3150 corruption" errors.
3151
3152 27. Applied patches from Google (a) to support the new newline modes and (b) to
3153 advance over multibyte UTF-8 characters in GlobalReplace.
3154
3155 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3156 difference for some implementation of PCRE in some Windows version.
3157
3158 29. Added some extra testing facilities to pcretest:
3159
3160 \q<number> in a data line sets the "match limit" value
3161 \Q<number> in a data line sets the "match recursion limt" value
3162 -S <number> sets the stack size, where <number> is in megabytes
3163
3164 The -S option isn't available for Windows.
3165
3166
3167 Version 6.6 06-Feb-06
3168 ---------------------
3169
3170 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3171 in pcreposix.h. I have copied the definition from pcre.h.
3172
3173 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3174 because pcre.h is no longer a built file.
3175
3176 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3177 not normally included in the compiled code.
3178
3179
3180 Version 6.5 01-Feb-06
3181 ---------------------
3182
3183 1. When using the partial match feature with pcre_dfa_exec(), it was not
3184 anchoring the second and subsequent partial matches at the new starting
3185 point. This could lead to incorrect results. For example, with the pattern
3186 /1234/, partially matching against "123" and then "a4" gave a match.
3187
3188 2. Changes to pcregrep:
3189
3190 (a) All non-match returns from pcre_exec() were being treated as failures
3191 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3192 error message is output. Some extra information is given for the
3193 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3194 probably the only errors that are likely to be caused by users (by
3195 specifying a regex that has nested indefinite repeats, for instance).
3196 If there are more than 20 of these errors, pcregrep is abandoned.
3197
3198 (b) A binary zero was treated as data while matching, but terminated the
3199 output line if it was written out. This has been fixed: binary zeroes
3200 are now no different to any other data bytes.
3201
3202 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3203 used to set a locale for matching. The --locale=xxxx long option has
3204 been added (no short equivalent) to specify a locale explicitly on the
3205 pcregrep command, overriding the environment variables.
3206
3207 (d) When -B was used with -n, some line numbers in the output were one less
3208 than they should have been.
3209
3210 (e) Added the -o (--only-matching) option.
3211
3212 (f) If -A or -C was used with -c (count only), some lines of context were
3213 accidentally printed for the final match.
3214
3215 (g) Added the -H (--with-filename) option.
3216
3217 (h) The combination of options -rh failed to suppress file names for files
3218 that were found from directory arguments.
3219
3220 (i) Added the -D (--devices) and -d (--directories) options.
3221
3222 (j) Added the -F (--fixed-strings) option.
3223
3224 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3225
3226 (l) Added the --colo(u)r option.
3227
3228 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3229 is not present by default.
3230
3231 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3232 items such as (?R) or (?1), when the recursion could match a number of
3233 alternatives. If it matched one of the alternatives, but subsequently,
3234 outside the recursion, there was a failure, the code tried to back up into
3235 the recursion. However, because of the way PCRE is implemented, this is not
3236 possible, and the result was an incorrect result from the match.
3237
3238 In order to prevent this happening, the specification of recursion has
3239 been changed so that all such subpatterns are automatically treated as
3240 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3241
3242 4. I had overlooked the fact that, in some locales, there are characters for
3243 which isalpha() is true but neither isupper() nor islower() are true. In
3244 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3245 and ordfeminine) are like this. This affected the treatment of \w and \W
3246 when they appeared in character classes, but not when they appeared outside
3247 a character class. The bit map for "word" characters is now created
3248 separately from the results of isalnum() instead of just taking it from the
3249 upper, lower, and digit maps. (Plus the underscore character, of course.)
3250
3251 5. The above bug also affected the handling of POSIX character classes such as
3252 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3253 permanent tables. Instead, the bit maps for such a class were previously
3254 created as the appropriate unions of the upper, lower, and digit bitmaps.
3255 Now they are created by subtraction from the [[:word:]] class, which has
3256 its own bitmap.
3257
3258 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3259 It is created by subtracting the vertical space characters (\x09, \x0a,
3260 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3261 subtraction was done in the overall bitmap for a character class, meaning
3262 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3263 be recognized. This bug has been fixed.
3264
3265 7. Patches from the folks at Google:
3266
3267 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3268 real life, but is still worth protecting against".
3269
3270 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3271 regular expressions".
3272
3273 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3274 have it.
3275
3276 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3277 "configure" and the latter not, in order to fix a problem somebody had
3278 with compiling the Arg class on HP-UX.
3279
3280 (e) Improve the error-handling of the C++ wrapper a little bit.
3281
3282 (f) New tests for checking recursion limiting.
3283
3284 8. The pcre_memmove() function, which is used only if the environment does not
3285 have a standard memmove() function (and is therefore rarely compiled),
3286 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3287 returning a result (though PCRE never actually uses the result).
3288
3289 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3290 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3291 returned instead of calling malloc() with an overflowing number that would
3292 most likely cause subsequent chaos.
3293
3294 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3295
3296 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3297 with this option is matched, the nmatch and pmatch options of regexec() are
3298 ignored.
3299
3300 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3301 provided in case anyone wants to the the POSIX interface with UTF-8
3302 strings.
3303
3304 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3305 C++ linking (needed for some HP-UX environments).
3306
3307 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3308 (unused parameter) and in the pcre_printint() function (omitted "default"
3309 switch label when the default is to do nothing).
3310
3311 15. Added some code to make it possible, when PCRE is compiled as a C++
3312 library, to replace subject pointers for pcre_exec() with a smart pointer
3313 class, thus making it possible to process discontinuous strings.
3314
3315 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3316 much the same function. They were added by different people who were trying
3317 to make PCRE easy to compile on non-Unix systems. It has been suggested
3318 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3319 for compiling on Windows systems. I have therefore replaced it with
3320 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3321 defaults to "extern" for C or "extern C" for C++, which works fine on
3322 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3323 SCOPE with something explicit in config.h. In addition:
3324
3325 (a) pcreposix.h still had just "extern" instead of either of these macros;
3326 I have replaced it with PCRE_DATA_SCOPE.
3327
3328 (b) Functions such as _pcre_xclass(), which are internal to the library,
3329 but external in the C sense, all had PCRE_EXPORT in their definitions.
3330 This is apparently wrong for the Windows case, so I have removed it.
3331 (It makes no difference on Unix-like systems.)
3332
3333 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3334 of recursive calls to match(). This is different to MATCH_LIMIT because
3335 that limits the total number of calls to match(), not all of which increase
3336 the depth of recursion. Limiting the recursion depth limits the amount of
3337 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3338 when PCRE is compiled, and changed at run time. A patch from Google adds
3339 this functionality to the C++ interface.
3340
3341 18. Changes to the handling of Unicode character properties:
3342
3343 (a) Updated the table to Unicode 4.1.0.
3344
3345 (b) Recognize characters that are not in the table as "Cn" (undefined).
3346
3347 (c) I revised the way the table is implemented to a much improved format
3348 which includes recognition of ranges. It now supports the ranges that
3349 are defined in UnicodeData.txt, and it also amalgamates other
3350 characters into ranges. This has reduced the number of entries in the
3351 table from around 16,000 to around 3,000, thus reducing its size
3352 considerably. I realized I did not need to use a tree structure after
3353 all - a binary chop search is just as efficient. Having reduced the
3354 number of entries, I extended their size from 6 bytes to 8 bytes to
3355 allow for more data.
3356
3357 (d) Added support for Unicode script names via properties such as \p{Han}.
3358
3359 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3360 matching that character.
3361
3362 20. When matching a repeated Unicode property with a minimum greater than zero,
3363 (for example \pL{2,}), PCRE could look past the end of the subject if it
3364 reached it while seeking the minimum number of characters. This could
3365 happen only if some of the characters were more than one byte long, because
3366 there is a check for at least the minimum number of bytes.
3367
3368 21. Refactored the implementation of \p and \P so as to be more general, to
3369 allow for more different types of property in future. This has changed the
3370 compiled form incompatibly. Anybody with saved compiled patterns that use
3371 \p or \P will have to recompile them.
3372
3373 22. Added "Any" and "L&" to the supported property types.
3374
3375 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3376 but give a compile time error if the value is greater than 0xff.
3377
3378 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3379 accidentally not being installed or uninstalled.
3380
3381 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3382 made were to insert the current release number. This seemed silly, because
3383 it made things harder for people building PCRE on systems that don't run
3384 "configure". I have turned pcre.h into a distributed file, no longer built
3385 by "configure", with the version identification directly included. There is
3386 no longer a pcre.h.in file.
3387
3388 However, this change necessitated a change to the pcre-config script as
3389 well. It is built from pcre-config.in, and one of the substitutions was the
3390 release number. I have updated configure.ac so that ./configure now finds
3391 the release number by grepping pcre.h.
3392
3393 26. Added the ability to run the tests under valgrind.
3394
3395
3396 Version 6.4 05-Sep-05
3397 ---------------------
3398
3399 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3400 "--" to be printed when multiple files were scanned, even when none of the
3401 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3402 consider it to be a bug, and have restored the previous behaviour.
3403
3404 2. A couple of code tidies to get rid of compiler warnings.
3405
3406 3. The pcretest program used to cheat by referring to symbols in the library
3407 whose names begin with _pcre_. These are internal symbols that are not
3408 really supposed to be visible externally, and in some environments it is
3409 possible to suppress them. The cheating is now confined to including
3410 certain files from the library's source, which is a bit cleaner.
3411
3412 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3413 file's purpose clearer.
3414
3415 5. Reorganized pcre_ucp_findchar().
3416
3417
3418 Version 6.3 15-Aug-05
3419 ---------------------
3420
3421 1. The file libpcre.pc.in did not have general read permission in the tarball.
3422
3423 2. There were some problems when building without C++ support:
3424
3425 (a) If C++ support was not built, "make install" and "make test" still
3426 tried to test it.
3427
3428 (b) There were problems when the value of CXX was explicitly set. Some
3429 changes have been made to try to fix these, and ...
3430
3431 (c) --disable-cpp can now be used to explicitly disable C++ support.
3432
3433 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3434 backslash in a target when C++ was disabled. This confuses some
3435 versions of "make", apparently. Using an intermediate variable solves
3436 this. (Same for CPP_LOBJ.)
3437
3438 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3439 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3440 necessary on certain architectures.
3441
3442 4. Added a setting of -export-symbols-regex to the link command to remove
3443 those symbols that are exported in the C sense, but actually are local
3444 within the library, and not documented. Their names all begin with
3445 "_pcre_". This is not a perfect job, because (a) we have to except some
3446 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3447 available (and never for static libraries). I have made a note to try to
3448 find a way round (a) in the future.
3449
3450
3451 Version 6.2 01-Aug-05
3452 ---------------------
3453
3454 1. There was no test for integer overflow of quantifier values. A construction
3455 such as {1111111111111111} would give undefined results. What is worse, if
3456 a minimum quantifier for a parenthesized subpattern overflowed and became
3457 negative, the calculation of the memory size went wrong. This could have
3458 led to memory overwriting.
3459
3460 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3461
3462 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3463 operating environments where this matters.
3464
3465 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3466 PCRE options from within the C++ wrapper.
3467
3468 5. Named capturing subpatterns were not being correctly counted when a pattern
3469 was compiled. This caused two problems: (a) If there were more than 100
3470 such subpatterns, the calculation of the memory needed for the whole
3471 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3472 back references of the form \12, where the number was greater than 9, were
3473 not recognized as back references, even though there were sufficient
3474 previous subpatterns.
3475
3476 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3477 versions of gcc, e.g. 2.95.4.
3478
3479
3480 Version 6.1 21-Jun-05
3481 ---------------------
3482
3483 1. There was one reference to the variable "posix" in pcretest.c that was not
3484 surrounded by "#if !defined NOPOSIX".
3485
3486 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3487 the cross-check on the old pcre_info() function, for the benefit of the
3488 cut-down version of PCRE that is currently imported into Exim.
3489
3490 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3491 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3492 patterns that start (?i) but I don't think that matters. The use of (?i) is
3493 just an example; this all applies to the other options as well.
3494
3495 4. Since libtool seems to echo the compile commands it is issuing, the output
3496 from "make" can be reduced a bit by putting "@" in front of each libtool
3497 compile command.
3498
3499 5. Patch from the folks at Google for configure.in to be a bit more thorough
3500 in checking for a suitable C++ installation before trying to compile the
3501 C++ stuff. This should fix a reported problem when a compiler was present,
3502 but no suitable headers.
3503
3504 6. The man pages all had just "PCRE" as their title. I have changed them to
3505 be the relevant file name. I have also arranged that these names are
3506 retained in the file doc/pcre.txt, which is a concatenation in text format
3507 of all the man pages except the little individual ones for each function.
3508
3509 7. The NON-UNIX-USE file had not been updated for the different set of source
3510 files that come with release 6. I also added a few comments about the C++
3511 wrapper.
3512
3513
3514 Version 6.0 07-Jun-05
3515 ---------------------
3516
3517 1. Some minor internal re-organization to help with my DFA experiments.
3518
3519 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3520 didn't matter for the library itself when fully configured, but did matter
3521 when compiling without UCP support, or within Exim, where the ucp files are
3522 not imported.
3523
3524 3. Refactoring of the library code to split up the various functions into
3525 different source modules. The addition of the new DFA matching code (see
3526 below) to a single monolithic source would have made it really too
3527 unwieldy, quite apart from causing all the code to be include in a
3528 statically linked application, when only some functions are used. This is
3529 relevant even without the DFA addition now that patterns can be compiled in
3530 one application and matched in another.
3531
3532 The downside of splitting up is that there have to be some external
3533 functions and data tables that are used internally in different modules of
3534 the library but which are not part of the API. These have all had their
3535 names changed to start with "_pcre_" so that they are unlikely to clash
3536 with other external names.
3537
3538 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3539 a different (DFA) algorithm. Although it is slower than the original
3540 function, it does have some advantages for certain types of matching
3541 problem.
3542
3543 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3544 including restarting after a partial match.
3545
3546 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3547 defined when compiling for Windows was sent to me. I have put it into the
3548 code, though I have no means of testing or verifying it.
3549
3550 7. Added the pcre_refcount() auxiliary function.
3551
3552 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3553 match before or at the first newline in the subject string. In pcretest,
3554 the /f option on a pattern can be used to set this.
3555
3556 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3557 would behave wrongly. This has been present in PCRE since release 4.0.
3558
3559 10. A number of changes to the pcregrep command:
3560
3561 (a) Refactored how -x works; insert ^(...)$ instead of setting
3562 PCRE_ANCHORED and checking the length, in preparation for adding
3563 something similar for -w.
3564
3565 (b) Added the -w (match as a word) option.
3566
3567 (c) Refactored the way lines are read and buffered so as to have more
3568 than one at a time available.
3569
3570 (d) Implemented a pcregrep test script.
3571
3572 (e) Added the -M (multiline match) option. This allows patterns to match
3573 over several lines of the subject. The buffering ensures that at least
3574 8K, or the rest of the document (whichever is the shorter) is available
3575 for matching (and similarly the previous 8K for lookbehind assertions).
3576
3577 (f) Changed the --help output so that it now says
3578
3579 -w, --word-regex(p)
3580
3581 instead of two lines, one with "regex" and the other with "regexp"
3582 because that confused at least one person since the short forms are the
3583 same. (This required a bit of code, as the output is generated
3584 automatically from a table. It wasn't just a text change.)
3585
3586 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3587 option but starts with a hyphen. Could be a pattern or a path name
3588 starting with a hyphen, for instance.
3589
3590 (h) "-" can be given as a file name to represent stdin.
3591
3592 (i) When file names are being printed, "(standard input)" is used for
3593 the standard input, for compatibility with GNU grep. Previously
3594 "<stdin>" was used.
3595
3596 (j) The option --label=xxx can be used to supply a name to be used for
3597 stdin when file names are being printed. There is no short form.
3598
3599 (k) Re-factored the options decoding logic because we are going to add
3600 two more options that take data. Such options can now be given in four
3601 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3602
3603 (l) Added the -A, -B, and -C options for requesting that lines of context
3604 around matches be printed.
3605
3606 (m) Added the -L option to print the names of files that do not contain
3607 any matching lines, that is, the complement of -l.
3608
3609 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3610 continue to scan other files.
3611
3612 (o) The -s option was incorrectly implemented. For compatibility with other
3613 greps, it now suppresses the error message for a non-existent or non-
3614 accessible file (but not the return code). There is a new option called
3615 -q that suppresses the output of matching lines, which was what -s was
3616 previously doing.
3617
3618 (p) Added --include and --exclude options to specify files for inclusion
3619 and exclusion when recursing.
3620
3621 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3622 Hopefully, it now does.
3623
3624 12. Missing cast in pcre_study().
3625
3626 13. Added an "uninstall" target to the makefile.
3627
3628 14. Replaced "extern" in the function prototypes in Makefile.in with
3629 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3630 world, but is set differently for Windows.
3631
3632 15. Added a second compiling function called pcre_compile2(). The only
3633 difference is that it has an extra argument, which is a pointer to an
3634 integer error code. When there is a compile-time failure, this is set
3635 non-zero, in addition to the error test pointer being set to point to an
3636 error message. The new argument may be NULL if no error number is required
3637 (but then you may as well call pcre_compile(), which is now just a
3638 wrapper). This facility is provided because some applications need a
3639 numeric error indication, but it has also enabled me to tidy up the way
3640 compile-time errors are handled in the POSIX wrapper.
3641
3642 16. Added VPATH=.libs to the makefile; this should help when building with one
3643 prefix path and installing with another. (Or so I'm told by someone who
3644 knows more about this stuff than I do.)
3645
3646 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3647 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3648 match everything, including newlines. This is not POSIX-compatible, but
3649 somebody wanted the feature. From pcretest it can be activated by using
3650 both the P and the s flags.
3651
3652 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3653
3654 19. libpcre.pc was being incorrectly installed as executable.
3655
3656 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3657 it now also looks for '\r' so that it will work unmodified on Windows.
3658
3659 21. Added Google's contributed C++ wrapper to the distribution.
3660
3661 22. Added some untidy missing memory free() calls in pcretest, to keep
3662 Electric Fence happy when testing.
3663
3664
3665
3666 Version 5.0 13-Sep-04
3667 ---------------------
3668
3669 1. Internal change: literal characters are no longer packed up into items
3670 containing multiple characters in a single byte-string. Each character
3671 is now matched using a separate opcode. However, there may be more than one
3672 byte in the character in UTF-8 mode.
3673
3674 2. The pcre_callout_block structure has two new fields: pattern_position and
3675 next_item_length. These contain the offset in the pattern to the next match
3676 item, and its length, respectively.
3677
3678 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3679 insertion of callouts before each pattern item. Added the /C option to
3680 pcretest to make use of this.
3681
3682 4. On the advice of a Windows user, the lines
3683
3684 #if defined(_WIN32) || defined(WIN32)
3685 _setmode( _fileno( stdout ), 0x8000 );
3686 #endif /* defined(_WIN32) || defined(WIN32) */
3687
3688 have been added to the source of pcretest. This apparently does useful
3689 magic in relation to line terminators.
3690
3691 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3692 for the benefit of those environments where the "b" makes a difference.
3693
3694 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3695 to know about it. I have put a hack into configure.in that adds in code
3696 to set GCC=yes if CC=icc. This seems to end up at a point in the
3697 generated configure script that is early enough to affect the setting of
3698 compiler options, which is what is needed, but I have no means of testing
3699 whether it really works. (The user who reported this had patched the
3700 generated configure script, which of course I cannot do.)
3701
3702 LATER: After change 22 below (new libtool files), the configure script
3703 seems to know about icc (and also ecc). Therefore, I have commented out
3704 this hack in configure.in.
3705
3706 7. Added support for pkg-config (2 patches were sent in).
3707
3708 8. Negated POSIX character classes that used a combination of internal tables
3709 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3710 [[:^ascii]]. Typically, they would match almost any characters. The other
3711 POSIX classes were not broken in this way.
3712
3713 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3714 to find the match, as PCRE was deluded into thinking that the match had to
3715 start at the start point or following a newline. The same bug applied to
3716 patterns with negative forward assertions or any backward assertions
3717 preceding ".*" at the start, unless the pattern required a fixed first
3718 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3719
3720 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3721 starting at the last subject character, bytes beyond the end of the subject
3722 string were read.
3723
3724 11. Renamed the variable "class" as "classbits" to make life easier for C++
3725 users. (Previously there was a macro definition, but it apparently wasn't
3726 enough.)
3727
3728 12. Added the new field "tables" to the extra data so that tables can be passed
3729 in at exec time, or the internal tables can be re-selected. This allows
3730 a compiled regex to be saved and re-used at a later time by a different
3731 program that might have everything at different addresses.
3732
3733 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3734 -R library as well as a -L library.
3735
3736 14. The debugging options of pcretest (-d on the command line or D on a
3737 pattern) showed incorrect output for anything following an extended class
3738 that contained multibyte characters and which was followed by a quantifier.
3739
3740 15. Added optional support for general category Unicode character properties
3741 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3742 support. It adds about 90K to the size of the library. The meanings of the
3743 inbuilt class escapes such as \d and \s have NOT been changed.
3744
3745 16. Updated pcredemo.c to include calls to free() to release the memory for the
3746 compiled pattern.
3747
3748 17. The generated file chartables.c was being created in the source directory
3749 instead of in the building directory. This caused the build to fail if the
3750 source directory was different from the building directory, and was
3751 read-only.
3752
3753 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3754 file. No doubt somebody will tell me if they don't make sense... Also added
3755 Dan Mooney's comments about building on OpenVMS.
3756
3757 19. Added support for partial matching via the PCRE_PARTIAL option for
3758 pcre_exec() and the \P data escape in pcretest.
3759
3760 20. Extended pcretest with 3 new pattern features:
3761
3762 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3763 write the compiled pattern to the file whose name is "rest-of-line".
3764 This is a straight binary dump of the data, with the saved pointer to
3765 the character tables forced to be NULL. The study data, if any, is
3766 written too. After writing, pcretest reads a new pattern.
3767
3768 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3769 compiled pattern from the given file. There must not be any
3770 occurrences of "<" in the file name (pretty unlikely); if there are,
3771 pcretest will instead treat the initial "<" as a pattern delimiter.
3772 After reading in the pattern, pcretest goes on to read data lines as
3773 usual.
3774
3775 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3776 and 16-bit fields in a compiled pattern, to simulate a pattern that
3777 was compiled on a host of opposite endianness.
3778
3779 21. The pcre-exec() function can now cope with patterns that were compiled on
3780 hosts of opposite endianness, with this restriction:
3781
3782 As for any compiled expression that is saved and used later, the tables
3783 pointer field cannot be preserved; the extra_data field in the arguments
3784 to pcre_exec() should be used to pass in a tables address if a value
3785 other than the default internal tables were used at compile time.
3786
3787 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3788 now diagnosed as an error. Previously, most of the time, a negative number
3789 would have been treated as zero, but if in addition "ovector" was passed as
3790 NULL, a crash could occur.
3791
3792 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3793 new versions from the libtool 1.5 distribution (the last one is a copy of
3794 a file called libtool.m4). This seems to have fixed the need to patch
3795 "configure" to support Darwin 1.3 (which I used to do). However, I still
3796 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3797 workstation).
3798
3799 24. Changed the PCRE licence to be the more standard "BSD" licence.
3800
3801
3802 Version 4.5 01-Dec-03
3803 ---------------------
3804
3805 1. There has been some re-arrangement of the code for the match() function so
3806 that it can be compiled in a version that does not call itself recursively.
3807 Instead, it keeps those local variables that need separate instances for
3808 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3809 needs to "recurse". Keeping track of where control must go is done by means
3810 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3811 hide most of the details from the main code, and operates only if
3812 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3813 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3814 operating.
3815
3816 To make it easier for callers to provide specially tailored get/free
3817 functions for this usage, two new functions, pcre_stack_malloc, and
3818 pcre_stack_free, are used. They are always called in strict stacking order,
3819 and the size of block requested is always the same.
3820
3821 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3822 PCRE has been compiled to use the stack or the heap for recursion. The
3823 -C option of pcretest uses this to show which version is compiled.
3824
3825 A new data escape \S, is added to pcretest; it causes the amounts of store
3826 obtained and freed by both kinds of malloc/free at match time to be added
3827 to the output.
3828
3829 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3830 what's available on my current Linux desktop machine.
3831
3832 3. When matching a UTF-8 string, the test for a valid string at the start has
3833 been extended. If start_offset is not zero, PCRE now checks that it points
3834 to a byte that is the start of a UTF-8 character. If not, it returns
3835 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3836 this is necessary because there may be backward assertions in the pattern.
3837 When matching the same subject several times, it may save resources to use
3838 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3839
3840 4. The code for checking the validity of UTF-8 strings has been tightened so
3841 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3842 containing "overlong sequences".
3843
3844 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3845 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3846 should have been "&", but it just so happened that all the cases this let
3847 through by mistake were picked up later in the function.
3848
3849 6. I had used a variable called "isblank" - this is a C99 function, causing
3850 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3851
3852 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3853 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3854 so the version is shown; (c) stop "make test" echoing "./RunTest".
3855
3856 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3857
3858 9. The prototype for memmove() for systems that don't have it was using
3859 size_t, but the inclusion of the header that defines size_t was later. I've
3860 moved the #includes for the C headers earlier to avoid this.
3861
3862 10. Added some adjustments to the code to make it easier to compiler on certain
3863 special systems:
3864
3865 (a) Some "const" qualifiers were missing.
3866 (b) Added the macro EXPORT before all exported functions; by default this
3867 is defined to be empty.
3868 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3869 that it reads its output file name as an argument instead of writing
3870 to the standard output and assuming this can be redirected.
3871
3872 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3873 class containing characters with values greater than 255, PCRE compilation
3874 went into a loop.
3875
3876 12. A recursive reference to a subpattern that was within another subpattern
3877 that had a minimum quantifier of zero caused PCRE to crash. For example,
3878 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3879 recursion. If the recursively-called subpattern itself had a zero repeat,
3880 that was OK.
3881
3882 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3883 buffer into which it was copied (for escape processing) was still set at
3884 1024, so long lines caused crashes.
3885
3886 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3887 "internal error: code overflow...". This applied to any character class
3888 that was followed by a possessive quantifier.
3889
3890 15. Modified the Makefile to add libpcre.la as a prerequisite for
3891 libpcreposix.la because I was told this is needed for a parallel build to
3892 work.
3893
3894 16. If a pattern that contained .* following optional items at the start was
3895 studied, the wrong optimizing data was generated, leading to matching
3896 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3897 matching string must start with a or b or c. The correct conclusion for
3898 this pattern is that a match can start with any character.
3899
3900
3901 Version 4.4 13-Aug-03
3902 ---------------------
3903
3904 1. In UTF-8 mode, a character class containing characters with values between
3905 127 and 255 was not handled correctly if the compiled pattern was studied.
3906 In fixing this, I have also improved the studying algorithm for such
3907 classes (slightly).
3908
3909 2. Three internal functions had redundant arguments passed to them. Removal
3910 might give a very teeny performance improvement.
3911
3912 3. Documentation bug: the value of the capture_top field in a callout is *one
3913 more than* the number of the hightest numbered captured substring.
3914
3915 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3916 in incorrectly linking with a previously installed version. They now link
3917 explicitly with libpcre.la.
3918
3919 5. configure.in no longer needs to recognize Cygwin specially.
3920
3921 6. A problem in pcre.in for Windows platforms is fixed.
3922
3923 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3924 pcretest, it used to include the size of the study block as part of its
3925 output. Unfortunately, the structure contains a field that has a different
3926 size on different hardware architectures. This meant that the tests that
3927 showed this size failed. As the block is currently always of a fixed size,
3928 this information isn't actually particularly useful in pcretest output, so
3929 I have just removed it.
3930
3931 8. Three pre-processor statements accidentally did not start in column 1.
3932 Sadly, there are *still* compilers around that complain, even though
3933 standard C has not required this for well over a decade. Sigh.
3934
3935 9. In pcretest, the code for checking callouts passed small integers in the
3936 callout_data field, which is a void * field. However, some picky compilers
3937 complained about the casts involved for this on 64-bit systems. Now
3938 pcretest passes the address of the small integer instead, which should get
3939 rid of the warnings.
3940
3941 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3942 both compile and run time, and gives an error if an invalid UTF-8 sequence
3943 is found. There is a option for disabling this check in cases where the
3944 string is known to be correct and/or the maximum performance is wanted.
3945
3946 11. In response to a bug report, I changed one line in Makefile.in from
3947
3948 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3949 to
3950 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3951
3952 to look similar to other lines, but I have no way of telling whether this
3953 is the right thing to do, as I do not use Windows. No doubt I'll get told
3954 if it's wrong...
3955
3956
3957 Version 4.3 21-May-03
3958 ---------------------
3959
3960 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3961 Makefile.
3962
3963 2. Some refactoring to improve the quality of the code:
3964
3965 (i) The utf8_table... variables are now declared "const".
3966
3967 (ii) The code for \cx, which used the "case flipping" table to upper case
3968 lower case letters, now just substracts 32. This is ASCII-specific,
3969 but the whole concept of \cx is ASCII-specific, so it seems
3970 reasonable.
3971
3972 (iii) PCRE was using its character types table to recognize decimal and
3973 hexadecimal digits in the pattern. This is silly, because it handles
3974 only 0-9, a-f, and A-F, but the character types table is locale-
3975 specific, which means strange things might happen. A private
3976 table is now used for this - though it costs 256 bytes, a table is
3977 much faster than multiple explicit tests. Of course, the standard
3978 character types table is still used for matching digits in subject
3979 strings against \d.
3980
3981 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3982 ending in _t are). So I've renamed it as ESC_tee.
3983
3984 3. The first argument for regexec() in the POSIX wrapper should have been
3985 defined as "const".
3986
3987 4. Changed pcretest to use malloc() for its buffers so that they can be
3988 Electric Fenced for debugging.
3989
3990 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3991 to read one or more bytes before the start of the subject string. Often this
3992 had no effect on PCRE's behaviour, but in some circumstances it could
3993 provoke a segmentation fault.
3994
3995 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3996 to try to read one or more bytes before the start of the subject string.
3997
3998 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3999 UTF-8 support could misbehave in various ways if the subject string
4000 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4001 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4002 back over UTF-8 characters.)
4003
4004
4005 Version 4.2 14-Apr-03
4006 ---------------------
4007
4008 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4009
4010 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4011 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4012 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4013 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4014 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4015 and BUILD_EXEEXT
4016 Note: automatic setting of the BUILD variables is not yet working
4017 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4018 compile-time but not at link-time
4019 [LINK]: use for linking executables only
4020 make different versions for Windows and non-Windows
4021 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4022 libraries
4023 [LINK_FOR_BUILD]: new variable
4024 [OBJEXT]: use throughout
4025 [EXEEXT]: use throughout
4026 <winshared>: new target
4027 <wininstall>: new target
4028 <dftables.o>: use native compiler
4029 <dftables>: use native linker
4030 <install>: handle Windows platform correctly
4031 <clean>: ditto
4032 <check>: ditto
4033 copy DLL to top builddir before testing
4034
4035 As part of these changes, -no-undefined was removed again. This was reported
4036 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4037 in any case.
4038
4039 3. Some tidies to get rid of compiler warnings:
4040
4041 . In the match_data structure, match_limit was an unsigned long int, whereas
4042 match_call_count was an int. I've made them both unsigned long ints.
4043
4044 . In pcretest the fact that a const uschar * doesn't automatically cast to
4045 a void * provoked a warning.
4046
4047 . Turning on some more compiler warnings threw up some "shadow" variables
4048 and a few more missing casts.
4049
4050 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4051 option, a class that contained a single character with a value between 128
4052 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4053
4054 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4055 option, a class that contained several characters, but with at least one
4056 whose value was between 128 and 255 caused PCRE to crash.
4057
4058
4059 Version 4.1 12-Mar-03
4060 ---------------------
4061
4062 1. Compiling with gcc -pedantic found a couple of places where casts were
4063 needed, and a string in dftables.c that was longer than standard compilers are
4064 required to support.
4065
4066 2. Compiling with Sun's compiler found a few more places where the code could
4067 be tidied up in order to avoid warnings.
4068
4069 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4070 first of these names is deprecated in the latest Autoconf in favour of the name
4071 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4072 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4073 analogy I have changed it to CFLAGS_FOR_BUILD.
4074
4075 4. Added -no-undefined to the linking command in the Makefile, because this is
4076 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4077 linking step for the pcreposix library.
4078
4079 5. PCRE was failing to diagnose the case of two named groups with the same
4080 name.
4081
4082 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4083 literal character that is needed in the subject for a match, and scans along to
4084 ensure that it is present before embarking on the full matching process. This
4085 saves time in cases of nested unlimited repeats that are never going to match.
4086 Problem: the scan can take a lot of time if the subject is very long (e.g.
4087 megabytes), thus penalizing straightforward matches. It is now done only if the
4088 amount of subject to be scanned is less than 1000 bytes.
4089
4090 7. A lesser problem with the same optimization is that it was recording the
4091 first character of an anchored pattern as "needed", thus provoking a search
4092 right along the subject, even when the first match of the pattern was going to
4093 fail. The "needed" character is now not set for anchored patterns, unless it
4094 follows something in the pattern that is of non-fixed length. Thus, it still
4095 fulfils its original purpose of finding quick non-matches in cases of nested
4096 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4097
4098
4099 Version 4.0 17-Feb-03
4100 ---------------------
4101
4102 1. If a comment in an extended regex that started immediately after a meta-item
4103 extended to the end of string, PCRE compiled incorrect data. This could lead to
4104 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4105
4106 2. Moved to autoconf 2.53 and libtool 1.4.2.
4107
4108 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4109 the special perltest8 script is no longer needed - all the tests can be run
4110 from a single perltest script.
4111
4112 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4113 by \s. It has now been removed in PCRE. This means it isn't recognized as
4114 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4115 class [:space:] *does* include VT, thereby creating a mess.
4116
4117 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4118 space and tab.
4119
4120 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4121 its new features into the main test script, reducing the number of scripts.
4122
4123 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4124 were backward compatible, and made the (?i) apply to the whole pattern, as if
4125 /i were given. Now it behaves more logically, and applies the option setting
4126 only to what follows. PCRE has been changed to follow suit. However, if it
4127 finds options settings right at the start of the pattern, it extracts them into
4128 the global options, as before. Thus, they show up in the info data.
4129
4130 8. Added support for the \Q...\E escape sequence. Characters in between are
4131 treated as literals. This is slightly different from Perl in that $ and @ are
4132 also handled as literals inside the quotes. In Perl, they will cause variable
4133 interpolation. Note the following examples:
4134
4135 Pattern PCRE matches Perl matches
4136
4137 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4138 \Qabc\$xyz\E abc\$xyz abc\$xyz
4139 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4140
4141 For compatibility with Perl, \Q...\E sequences are recognized inside character
4142 classes as well as outside them.
4143
4144 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4145 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4146 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4147 signed/unsigned warnings.
4148
4149 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4150 option for pcretest, so I've replaced it by a simple function that does just
4151 that job.
4152
4153 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4154 "pcregrep -".
4155
4156 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4157 Java package. This provides some syntactic sugar for simple cases of what my
4158 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4159 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4160 item, you can use this simplified notation. Note that only makes sense with
4161 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4162 greediness, whatever the setting of the PCRE_UNGREEDY option.
4163
4164 13. A change of greediness default within a pattern was not taking effect at
4165 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4166 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4167 was abstracted outside.
4168
4169 14. PCRE now supports the \G assertion. It is true when the current matching
4170 position is at the start point of the match. This differs from \A when the
4171 starting offset is non-zero. Used with the /g option of pcretest (or similar
4172 code), it works in the same way as it does for Perl's /g option. If all
4173 alternatives of a regex begin with \G, the expression is anchored to the start
4174 match position, and the "anchored" flag is set in the compiled expression.
4175
4176 15. Some bugs concerning the handling of certain option changes within patterns
4177 have been fixed. These applied to options other than (?ims). For example,
4178 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4179 been the other way round. Some of this was related to change 7 above.
4180
4181 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4182 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4183 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4184 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4185
4186 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4187 mode. Unlike ".", it always matches newline, whatever the setting of
4188 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4189 assertions. Perl allows it, but it doesn't (in general) work because it can't
4190 calculate the length of the lookbehind. At least, that's the case for Perl
4191 5.8.0 - I've been told they are going to document that it doesn't work in
4192 future.
4193
4194 18. Added an error diagnosis for escapes that PCRE does not support: these are
4195 \L, \l, \N, \P, \p, \U, \u, and \X.
4196
4197 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4198 reading past the end of the pattern in cases such as /[abcd/.
4199
4200 20. PCRE was getting more memory than necessary for patterns with classes that
4201 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4202
4203 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4204 compiling PCRE for use with Virtual Pascal.
4205
4206 22. Small fix to the Makefile to make it work properly if the build is done
4207 outside the source tree.
4208
4209 23. Added a new extension: a condition to go with recursion. If a conditional
4210 subpattern starts with (?(R) the "true" branch is used if recursion has
4211 happened, whereas the "false" branch is used only at the top level.
4212
4213 24. When there was a very long string of literal characters (over 255 bytes
4214 without UTF support, over 250 bytes with UTF support), the computation of how
4215 much memory was required could be incorrect, leading to segfaults or other
4216 strange effects.
4217
4218 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4219 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4220 there was a subsequent back reference to those brackets. This meant that, for
4221 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4222 possible to check for precisely this case. All we can do is abandon the
4223 optimization if .* occurs inside capturing brackets when there are any back
4224 references whatsoever. (See below for a better fix that came later.)
4225
4226 26. The handling of the optimization for finding the first character of a
4227 non-anchored pattern, and for finding a character that is required later in the
4228 match were failing in some cases. This didn't break the matching; it just
4229 failed to optimize when it could. The way this is done has been re-implemented.
4230
4231 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4232
4233 28. Added a new feature that provides some of the functionality that Perl
4234 provides with (?{...}). The facility is termed a "callout". The way it is done
4235 in PCRE is for the caller to provide an optional function, by setting
4236 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4237 global variable. By default it is unset, which disables all calling out. To get
4238 the function called, the regex must include (?C) at appropriate points. This
4239 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4240 This provides a means of identifying different callout points. When PCRE
4241 reaches such a point in the regex, if pcre_callout has been set, the external
4242 function is called. It is provided with data in a structure called
4243 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4244 matching continues; if it returns a non-zero value, the match at the current
4245 point fails. However, backtracking will occur if possible. [This was changed
4246 later and other features added - see item 49 below.]
4247
4248 29. pcretest is upgraded to test the callout functionality. It provides a
4249 callout function that displays information. By default, it shows the start of
4250 the match and the current position in the text. There are some new data escapes
4251 to vary what happens:
4252
4253 \C+ in addition, show current contents of captured substrings
4254 \C- do not supply a callout function
4255 \C!n return 1 when callout number n is reached
4256 \C!n!m return 1 when callout number n is reached for the mth time
4257
4258 30. If pcregrep was called with the -l option and just a single file name, it
4259 output "<stdin>" if a match was found, instead of the file name.
4260
4261 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4262 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4263 pcre_exec(). This saves a malloc/free per call. The default value of
4264 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4265 when configuring.
4266
4267 32. The default maximum size of a compiled pattern is 64K. There have been a
4268 few cases of people hitting this limit. The code now uses macros to handle the
4269 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4270 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4271 configuring. Tests 2 and 5 work only with 2-byte links because they output
4272 debugging information about compiled patterns.
4273
4274 33. Internal code re-arrangements:
4275
4276 (a) Moved the debugging function for printing out a compiled regex into
4277 its own source file (printint.c) and used #include to pull it into
4278 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4279 separate copies.
4280
4281 (b) Defined the list of op-code names for debugging as a macro in
4282 internal.h so that it is next to the definition of the opcodes.
4283
4284 (c) Defined a table of op-code lengths for simpler skipping along compiled
4285 code. This is again a macro in internal.h so that it is next to the
4286 definition of the opcodes.
4287
4288 34. Added support for recursive calls to individual subpatterns, along the
4289 lines of Robin Houston's patch (but implemented somewhat differently).
4290
4291 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4292 allow it to read and process whole directories in Win32. This code was
4293 contributed by Lionel Fourquaux; it has not been tested by me.
4294
4295 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4296 used to name a group. Names consist of alphanumerics and underscores, and must
4297 be unique. Back references use the syntax (?P=name) and recursive calls use
4298 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4299 numbers. The function pcre_fullinfo() can be used after compilation to extract
4300 a name/number map. There are three relevant calls:
4301
4302 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4303 PCRE_INFO_NAMECOUNT yields the number of entries
4304 PCRE_INFO_NAMETABLE yields a pointer to the map.
4305
4306 The map is a vector of fixed-size entries. The size of each entry depends on
4307 the length of the longest name used. The first two bytes of each entry are the
4308 group number, most significant byte first. There follows the corresponding
4309 name, zero terminated. The names are in alphabetical order.
4310
4311 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4312 case instead of 255. Making it the same both with and without UTF-8 support
4313 means that the same test output works with both.
4314
4315 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4316 calling malloc() with a zero argument.
4317
4318 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4319 optimization. I've improved things by keeping a bitmap of backreferences with
4320 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4321 fact referenced, the optimization can be applied. It is unlikely that a
4322 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4323 the match to follow \n) will appear inside brackets with a number greater than
4324 31, but if it does, any back reference > 31 suppresses the optimization.
4325
4326 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4327 of disabling numbered capturing parentheses. Any opening parenthesis that is
4328 not followed by ? behaves as if it were followed by ?: but named parentheses
4329 can still be used for capturing (and they will acquire numbers in the usual
4330 way).
4331
4332 41. Redesigned the return codes from the match() function into yes/no/error so
4333 that errors can be passed back from deep inside the nested calls. A malloc
4334 failure while inside a recursive subpattern call now causes the
4335 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4336
4337 42. It is now possible to set a limit on the number of times the match()
4338 function is called in a call to pcre_exec(). This facility makes it possible to
4339 limit the amount of recursion and backtracking, though not in a directly
4340 obvious way, because the match() function is used in a number of different
4341 circumstances. The count starts from zero for each position in the subject
4342 string (for non-anchored patterns). The default limit is, for compatibility, a
4343 large number, namely 10 000 000. You can change this in two ways:
4344
4345 (a) When configuring PCRE before making, you can use --with-match-limit=n
4346 to set a default value for the compiled library.
4347
4348 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4349 a different value is set. See 45 below.
4350
4351 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4352
4353 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4354 of things that can be changed at compile time. The first argument specifies
4355 what is wanted and the second points to where the information is to be placed.
4356 The current list of available information is:
4357
4358 PCRE_CONFIG_UTF8
4359
4360 The output is an integer that is set to one if UTF-8 support is available;
4361 otherwise it is set to zero.
4362
4363 PCRE_CONFIG_NEWLINE
4364
4365 The output is an integer that it set to the value of the code that is used for
4366 newline. It is either LF (10) or CR (13).
4367
4368 PCRE_CONFIG_LINK_SIZE
4369
4370 The output is an integer that contains the number of bytes used for internal
4371 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4372
4373 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4374
4375 The output is an integer that contains the threshold above which the POSIX
4376 interface uses malloc() for output vectors. See item 31 above.
4377
4378 PCRE_CONFIG_MATCH_LIMIT
4379
4380 The output is an unsigned integer that contains the default limit of the number
4381 of match() calls in a pcre_exec() execution. See 42 above.
4382
4383 44. pcretest has been upgraded by the addition of the -C option. This causes it
4384 to extract all the available output from the new pcre_config() function, and to
4385 output it. The program then exits immediately.
4386
4387 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4388 order to support additional features. One way would have been to define
4389 pcre_exec2() (for example) with extra arguments, but this would not have been
4390 extensible, and would also have required all calls to the original function to
4391 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4392 is used for passing in "extra" data from pcre_study().
4393
4394 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4395 contains the following fields:
4396
4397 flags a bitmap indicating which of the following fields are set
4398 study_data opaque data from pcre_study()
4399 match_limit a way of specifying a limit on match() calls for a specific
4400 call to pcre_exec()
4401 callout_data data for callouts (see 49 below)
4402
4403 The flag bits are also defined in pcre.h, and are
4404
4405 PCRE_EXTRA_STUDY_DATA
4406 PCRE_EXTRA_MATCH_LIMIT
4407 PCRE_EXTRA_CALLOUT_DATA
4408
4409 The pcre_study() function now returns one of these new pcre_extra blocks, with
4410 the actual study data pointed to by the study_data field, and the
4411 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4412 before. That is, this change is entirely upwards-compatible and requires no
4413 change to existing code.
4414
4415 If you want to pass in additional data to pcre_exec(), you can either place it
4416 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4417 block.
4418
4419 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4420 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4421 times with different match limits, until it finds the minimum value needed for
4422 pcre_exec() to complete. The value is then output. This can be instructive; for
4423 most simple matches the number is quite small, but for pathological cases it
4424 gets very large very quickly.
4425
4426 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4427 returns the size of the data block pointed to by the study_data field in a
4428 pcre_extra block, that is, the value that was passed as the argument to
4429 pcre_malloc() when PCRE was getting memory in which to place the information
4430 created by pcre_study(). The fourth argument should point to a size_t variable.
4431 pcretest has been extended so that this information is shown after a successful
4432 pcre_study() call when information about the compiled regex is being displayed.
4433
4434 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4435 because what follows is always an absolute path. (Later: it turns out that this
4436 is more than cosmetic for MinGW, because it doesn't like empty path
4437 components.)
4438
4439 49. Some changes have been made to the callout feature (see 28 above):
4440
4441 (i) A callout function now has three choices for what it returns:
4442
4443 0 => success, carry on matching
4444 > 0 => failure at this point, but backtrack if possible
4445 < 0 => serious error, return this value from pcre_exec()
4446
4447 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4448 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4449 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4450 use by callout functions. It will never be used by PCRE itself.
4451
4452 (ii) The pcre_extra structure (see 45 above) has a void * field called
4453 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4454 pcre_callout_block structure has a field of the same name. The contents of
4455 the field passed in the pcre_extra structure are passed to the callout
4456 function in the corresponding field in the callout block. This makes it
4457 easier to use the same callout-containing regex from multiple threads. For
4458 testing, the pcretest program has a new data escape
4459
4460 \C*n pass the number n (may be negative) as callout_data
4461
4462 If the callout function in pcretest receives a non-zero value as
4463 callout_data, it returns that value.
4464
4465 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4466 there were some redundant $(CFLAGS) in commands that are now specified as
4467 $(LINK), which already includes $(CFLAGS).
4468
4469 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4470 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4471 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4472 one-byte characters throughout. Note that case-insensitive matching applies
4473 only to characters whose values are less than 256. PCRE doesn't support the
4474 notion of cases for higher-valued characters.
4475
4476 (i) A character class whose characters are all within 0-255 is handled as
4477 a bit map, and the map is inverted for negative classes. Previously, a
4478 character > 255 always failed to match such a class; however it should
4479 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4480
4481 (ii) A negated character class with a single character < 255 is coded as
4482 "not this character" (OP_NOT). This wasn't working properly when the test
4483 character was multibyte, either singly or repeated.
4484
4485 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4486 mode, for example: \x{100}{2,3}.
4487
4488 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4489 singly or repeated) now correctly test multibyte characters. However,
4490 PCRE doesn't recognize any characters with values greater than 255 as
4491 digits, spaces, or word characters. Such characters always match \D, \S,
4492 and \W, and never match \d, \s, or \w.
4493
4494 (v) Classes may now contain characters and character ranges with values
4495 greater than 255. For example: [ab\x{100}-\x{400}].
4496
4497 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4498 PCRE in UTF-8 mode.
4499
4500 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4501 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4502 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4503 value.)
4504
4505 53. The single man page has become too large. I have therefore split it up into
4506 a number of separate man pages. These also give rise to individual HTML pages;
4507 these are now put in a separate directory, and there is an index.html page that
4508 lists them all. Some hyperlinking between the pages has been installed.
4509
4510 54. Added convenience functions for handling named capturing parentheses.
4511
4512 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4513 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4514 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4515 are faulted.
4516
4517 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4518 calling configure. These values are used when compiling the dftables.c program
4519 which is run to generate the source of the default character tables. They
4520 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4521 you will need to set these values.
4522
4523 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4524
4525
4526 Version 3.9 02-Jan-02
4527 ---------------------
4528
4529 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4530
4531 2. If --disable-static was given, the building process failed when trying to
4532 build pcretest and pcregrep. (For some reason it was using libtool to compile
4533 them, which is not right, as they aren't part of the library.)
4534
4535
4536 Version 3.8 18-Dec-01
4537 ---------------------
4538
4539 1. The experimental UTF-8 code was completely screwed up. It was packing the
4540 bytes in the wrong order. How dumb can you get?
4541
4542
4543 Version 3.7 29-Oct-01
4544 ---------------------
4545
4546 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4547 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4548 this didn't happen under Solaris 8, where I normally test things.
4549
4550 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4551 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4552 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4553 this fix an hour or so after the initial 3.7 release.)
4554
4555
4556 Version 3.6 23-Oct-01
4557 ---------------------
4558
4559 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4560 offsets passed as NULL with zero offset count.
4561
4562 2. The config.guess and config.sub files had not been updated when I moved to
4563 the latest autoconf.
4564
4565
4566 Version 3.5 15-Aug-01
4567 ---------------------
4568
4569 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4570 had been forgotten.
4571
4572 2. By using declared but undefined structures, we can avoid using "void"
4573 definitions in pcre.h while keeping the internal definitions of the structures
4574 private.
4575
4576 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4577 user point of view, this means that both static and shared libraries are built
4578 by default, but this can be individually controlled. More of the work of
4579 handling this static/shared cases is now inside libtool instead of PCRE's make
4580 file.
4581
4582 4. The pcretest utility is now installed along with pcregrep because it is
4583 useful for users (to test regexs) and by doing this, it automatically gets
4584 relinked by libtool. The documentation has been turned into a man page, so
4585 there are now .1, .txt, and .html versions in /doc.
4586
4587 5. Upgrades to pcregrep:
4588 (i) Added long-form option names like gnu grep.
4589 (ii) Added --help to list all options with an explanatory phrase.
4590 (iii) Added -r, --recursive to recurse into sub-directories.
4591 (iv) Added -f, --file to read patterns from a file.
4592
4593 6. pcre_exec() was referring to its "code" argument before testing that
4594 argument for NULL (and giving an error if it was NULL).
4595
4596 7. Upgraded Makefile.in to allow for compiling in a different directory from
4597 the source directory.
4598
4599 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4600 options bits, the pointer it was passed was to an int instead of to an unsigned
4601 long int. This mattered only on 64-bit systems.
4602
4603 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4604 generated) instead of pcre.in, which it its source. Also made the same change
4605 in several of the .c files.
4606
4607 10. A new release of gcc defines printf() as a macro, which broke pcretest
4608 because it had an ifdef in the middle of a string argument for printf(). Fixed
4609 by using separate calls to printf().
4610
4611 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4612 script, to force use of CR or LF instead of \n in the source. On non-Unix
4613 systems, the value can be set in config.h.
4614
4615 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4616 absolute limit. Changed the text of the error message to make this clear, and
4617 likewise updated the man page.
4618
4619 13. The limit of 99 on the number of capturing subpatterns has been removed.
4620 The new limit is 65535, which I hope will not be a "real" limit.
4621
4622
4623 Version 3.4 22-Aug-00
4624 ---------------------
4625
4626 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4627
4628 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4629
4630
4631 Version 3.3 01-Aug-00
4632 ---------------------
4633
4634 1. If an octal character was given, but the value was greater than \377, it
4635 was not getting masked to the least significant bits, as documented. This could
4636 lead to crashes in some systems.
4637
4638 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4639 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4640
4641 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4642 These just pass their arguments on to (pcre_free)(), but they are provided
4643 because some uses of PCRE bind it to non-C systems that can call its functions,
4644 but cannot call free() or pcre_free() directly.
4645
4646 4. Add "make test" as a synonym for "make check". Corrected some comments in
4647 the Makefile.
4648
4649 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4650 Makefile.
4651
4652 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4653 command called pgrep for grepping around the active processes.
4654
4655 7. Added the beginnings of support for UTF-8 character strings.
4656
4657 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4658 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4659 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4660 out for the ar command.)
4661
4662
4663 Version 3.2 12-May-00
4664 ---------------------
4665
4666 This is purely a bug fixing release.
4667
4668 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4669 of ZA. This was just one example of several cases that could provoke this bug,
4670 which was introduced by change 9 of version 2.00. The code for breaking
4671 infinite loops after an iteration that matches an empty string was't working
4672 correctly.
4673
4674 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4675 when matched against abbab (for example). After matching an empty string, it
4676 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4677 caused it to match further down the string than it should.
4678
4679 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4680 was there because it doesn't seem to be needed, and it causes trouble on some
4681 systems, as it is not a Standard C header. It has been removed.
4682
4683 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4684 were reported on the Macintosh. The changes were from
4685
4686 while ((c = *(++ptr)) != 0 && c != '\n');
4687 to
4688 while ((c = *(++ptr)) != 0 && c != '\n') ;
4689
4690 Totally extraordinary, but if that's what it takes...
4691
4692 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4693 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4694 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4695 assumes the way PCRE uses memmove() (always moving upwards).
4696
4697 6. PCRE is being used in one environment where strchr() is not available. There
4698 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4699 faster code anyway.
4700
4701
4702 Version 3.1 09-Feb-00
4703 ---------------------
4704
4705 The only change in this release is the fixing of some bugs in Makefile.in for
4706 the "install" target:
4707
4708 (1) It was failing to install pcreposix.h.
4709
4710 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4711
4712
4713 Version 3.0 01-Feb-00
4714 ---------------------
4715
4716 1. Add support for the /+ modifier to perltest (to output $` like it does in
4717 pcretest).
4718
4719 2. Add support for the /g modifier to perltest.
4720
4721 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4722 matches null strings.
4723
4724 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4725 pattern. Perl treats empty patterns specially - it reuses the most recent
4726 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4727 effect.
4728
4729 5. The POSIX interface was broken in that it was just handing over the POSIX
4730 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4731 required a bigger vector, with some working space on the end. This means that
4732 the POSIX wrapper now has to get and free some memory, and copy the results.
4733
4734 6. Added some simple autoconf support, placing the test data and the
4735 documentation in separate directories, re-organizing some of the
4736 information files, and making it build pcre-config (a GNU standard). Also added
4737 libtool support for building PCRE as a shared library, which is now the
4738 default.
4739
4740 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4741 09 are not valid octal constants. Single digits will be used for minor values
4742 less than 10.
4743
4744 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4745 existing programs that set these in the POSIX interface can use PCRE without
4746 modification.
4747
4748 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4749 return all that pcre_info() returns, plus additional data. The pcre_info()
4750 function is retained for compatibility, but is considered to be obsolete.
4751
4752 10. Added experimental recursion feature (?R) to handle one common case that
4753 Perl 5.6 will be able to do with (?p{...}).
4754
4755 11. Added support for POSIX character classes like [:alpha:], which Perl is
4756 adopting.
4757
4758
4759 Version 2.08 31-Aug-99
4760 ----------------------
4761
4762 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4763 trying to match at the startoffset position, but instead was moving forward to
4764 the next newline as if a previous match had failed.
4765
4766 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4767 and could get into a loop if a null string was matched other than at the start
4768 of the subject.
4769
4770 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4771 be distinguished at compile time, and for completeness also added PCRE_DATE.
4772
4773 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4774 in GnuWin32 environments.
4775
4776
4777 Version 2.07 29-Jul-99
4778 ----------------------
4779
4780 1. The documentation is now supplied in plain text form and HTML as well as in
4781 the form of man page sources.
4782
4783 2. C++ compilers don't like assigning (void *) values to other pointer types.
4784 In particular this affects malloc(). Although there is no problem in Standard
4785 C, I've put in casts to keep C++ compilers happy.
4786
4787 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4788 should be (const char *).
4789
4790 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4791 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4792 However, I haven't made this a standard facility. The documentation doesn't
4793 mention it, and the Makefile doesn't support it.
4794
4795 5. The Makefile now contains an "install" target, with editable destinations at
4796 the top of the file. The pcretest program is not installed.
4797
4798 6. pgrep -V now gives the PCRE version number and date.
4799
4800 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4801 causing the entire string to be ignored, instead of just the last character.
4802
4803 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4804 non-matching string, it can take a very, very long time, even for strings of
4805 quite modest length, because of the nested recursion. PCRE now does better in
4806 some of these cases. It does this by remembering the last required literal
4807 character in the pattern, and pre-searching the subject to ensure it is present
4808 before running the real match. In other words, it applies a heuristic to detect
4809 some types of certain failure quickly, and in the above example, if presented
4810 with a string that has no trailing " it gives "no match" very quickly.
4811
4812 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4813 other alternatives are tried instead.
4814
4815
4816 Version 2.06 09-Jun-99
4817 ----------------------
4818
4819 1. Change pcretest's output for amount of store used to show just the code
4820 space, because the remainder (the data block) varies in size between 32-bit and
4821 64-bit systems.
4822
4823 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4824 start matching at. This allows lookbehinds to work when searching for multiple
4825 occurrences in a string.
4826
4827 3. Added additional options to pcretest for testing multiple occurrences:
4828
4829 /+ outputs the rest of the string that follows a match
4830 /g loops for multiple occurrences, using the new startoffset argument
4831 /G loops for multiple occurrences by passing an incremented pointer
4832
4833 4. PCRE wasn't doing the "first character" optimization for patterns starting
4834 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4835 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4836 the letter 'x'. On long subject strings, this gives a significant speed-up.
4837
4838
4839 Version 2.05 21-Apr-99
4840 ----------------------
4841
4842 1. Changed the type of magic_number from int to long int so that it works
4843 properly on 16-bit systems.
4844
4845 2. Fixed a bug which caused patterns starting with .* not to work correctly
4846 when the subject string contained newline characters. PCRE was assuming
4847 anchoring for such patterns in all cases, which is not correct because .* will
4848 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4849 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4850 must be retried after every newline in the subject.
4851
4852
4853 Version 2.04 18-Feb-99
4854 ----------------------
4855
4856 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4857 computation of the store needed to hold the pattern was incorrect (too large).
4858 If such patterns were nested a few deep, this could multiply and become a real
4859 problem.
4860
4861 2. Added /M option to pcretest to show the memory requirement of a specific
4862 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4863
4864 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4865 compiled in such a way that the backtracking after subsequent failure was
4866 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4867 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4868
4869
4870 Version 2.03 02-Feb-99
4871 ----------------------
4872
4873 1. Fixed typo and small mistake in man page.
4874
4875 2. Added 4th condition (GPL supersedes if conflict) and created separate
4876 LICENCE file containing the conditions.
4877
4878 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4879 Perl, that is the internal \ allows the delimiter to be included in the
4880 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4881 the final delimiter, add \ to the end of the pattern (to test the error).
4882
4883 4. Added the convenience functions for extracting substrings after a successful
4884 match. Updated pcretest to make it able to test these functions.
4885
4886
4887 Version 2.02 14-Jan-99
4888 ----------------------
4889
4890 1. Initialized the working variables associated with each extraction so that
4891 their saving and restoring doesn't refer to uninitialized store.
4892
4893 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4894 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4895 fix the problem.
4896
4897 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4898 calls, and wasn't printing the correct value for compiling calls. Increased the
4899 default value of LOOPREPEAT, and the number of significant figures in the
4900 times.
4901
4902 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4903
4904 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4905 a building problem on Windows NT with a FAT file system.
4906
4907
4908 Version 2.01 21-Oct-98
4909 ----------------------
4910
4911 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4912 to character tables built by pcre_maketables() in the current locale. If NULL
4913 is passed, the default tables are used.
4914
4915
4916 Version 2.00 24-Sep-98
4917 ----------------------
4918
4919 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4920 it any more.
4921
4922 2. Allow quantification of (?>) groups, and make it work correctly.
4923
4924 3. The first character computation wasn't working for (?>) groups.
4925
4926 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4927 end of the subject) and add 5.005's \z, which really does match only at the
4928 very end of the subject.
4929
4930 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4931
4932 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4933 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4934 localized options. All options to pcre_study() were also removed.
4935
4936 7. Add other new features from 5.005:
4937
4938 $(?<= positive lookbehind
4939 $(?<! negative lookbehind
4940 (?imsx-imsx) added the unsetting capability
4941 such a setting is global if at outer level; local otherwise
4942 (?imsx-imsx:) non-capturing groups with option setting
4943 (?(cond)re|re) conditional pattern matching
4944
4945 A backreference to itself in a repeated group matches the previous
4946 captured string.
4947
4948 8. General tidying up of studying (both automatic and via "study")
4949 consequential on the addition of new assertions.
4950
4951 9. As in 5.005, unlimited repeated groups that could match an empty substring
4952 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4953 runtime if any iteration does actually match an empty substring.
4954
4955 10. Include the RunTest script in the distribution.
4956
4957 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4958 discrepancies, some of which were old and were also with respect to 5.004. They
4959 have now been fixed.
4960
4961
4962 Version 1.09 28-Apr-98
4963 ----------------------
4964
4965 1. A negated single character class followed by a quantifier with a minimum
4966 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4967 program crashes, or just wrong answers. This did not apply to negated classes
4968 containing more than one character, or to minima other than one.
4969
4970
4971 Version 1.08 27-Mar-98
4972 ----------------------
4973
4974 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4975
4976 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4977 latter must appear before anything that relies on it in the pattern.
4978
4979
4980 Version 1.07 16-Feb-98
4981 ----------------------
4982
4983 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4984 repeat of a potentially empty string).
4985
4986
4987 Version 1.06 23-Jan-98
4988 ----------------------
4989
4990 1. Added Markus Oberhumer's little patches for C++.
4991
4992 2. Literal strings longer than 255 characters were broken.
4993
4994
4995 Version 1.05 23-Dec-97
4996 ----------------------
4997
4998 1. Negated character classes containing more than one character were failing if
4999 PCRE_CASELESS was set at run time.
5000
5001
5002 Version 1.04 19-Dec-97
5003 ----------------------
5004
5005 1. Corrected the man page, where some "const" qualifiers had been omitted.
5006
5007 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5008 input syntax.
5009
5010 3. Fixed memory leak which occurred when a regex with back references was
5011 matched with an offsets vector that wasn't big enough. The temporary memory
5012 that is used in this case wasn't being freed if the match failed.
5013
5014 4. Tidied pcretest to ensure it frees memory that it gets.
5015
5016 5. Temporary memory was being obtained in the case where the passed offsets
5017 vector was exactly big enough.
5018
5019 6. Corrected definition of offsetof() from change 5 below.
5020
5021 7. I had screwed up change 6 below and broken the rules for the use of
5022 setjmp(). Now fixed.
5023
5024
5025 Version 1.03 18-Dec-97
5026 ----------------------
5027
5028 1. A erroneous regex with a missing opening parenthesis was correctly
5029 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5030 on some systems.
5031
5032 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5033 it was reported that one broken compiler failed on the former because "code" is
5034 also an independent variable.
5035
5036 3. The erroneous regex a[]b caused an array overrun reference.
5037
5038 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5039 fail on data ending with that character. (It was going on too far, and checking
5040 the next character, typically a binary zero.) This was specific to the
5041 optimized code for single-character negative classes.
5042
5043 5. Added a contributed patch from the TIN world which does the following:
5044
5045 + Add an undef for memmove, in case the the system defines a macro for it.
5046
5047 + Add a definition of offsetof(), in case there isn't one. (I don't know
5048 the reason behind this - offsetof() is part of the ANSI standard - but
5049 it does no harm).
5050
5051 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5052 most of the places where whitespace preceded '#'. I have given up and
5053 allowed the remaining 2 cases to be at the margin.
5054
5055 + Rename some variables in pcre to eliminate shadowing. This seems very
5056 pedantic, but does no harm, of course.
5057
5058 6. Moved the call to setjmp() into its own function, to get rid of warnings
5059 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5060
5061 7. Constructs such as \d{8,} were compiling into the equivalent of
5062 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5063 outcome, but in this particular case used more store than had been allocated,
5064 which caused the bug to be discovered because it threw up an internal error.
5065
5066 8. The debugging code in both pcre and pcretest for outputting the compiled
5067 form of a regex was going wrong in the case of back references followed by
5068 curly-bracketed repeats.
5069
5070
5071 Version 1.02 12-Dec-97
5072 ----------------------
5073
5074 1. Typos in pcre.3 and comments in the source fixed.
5075
5076 2. Applied a contributed patch to get rid of places where it used to remove
5077 'const' from variables, and fixed some signed/unsigned and uninitialized
5078 variable warnings.
5079
5080 3. Added the "runtest" target to Makefile.
5081
5082 4. Set default compiler flag to -O2 rather than just -O.
5083
5084
5085 Version 1.01 19-Nov-97
5086 ----------------------
5087
5088 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5089 like /([ab]*)*/, that is, for classes with more than one character in them.
5090
5091 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5092 as /((?>a*))*/ (a PCRE_EXTRA facility).
5093
5094
5095 Version 1.00 18-Nov-97
5096 ----------------------
5097
5098 1. Added compile-time macros to support systems such as SunOS4 which don't have
5099 memmove() or strerror() but have other things that can be used instead.
5100
5101 2. Arranged that "make clean" removes the executables.
5102
5103
5104 Version 0.99 27-Oct-97
5105 ----------------------
5106
5107 1. Fixed bug in code for optimizing classes with only one character. It was
5108 initializing a 32-byte map regardless, which could cause it to run off the end
5109 of the memory it had got.
5110
5111 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5112
5113
5114 Version 0.98 22-Oct-97
5115 ----------------------
5116
5117 1. Fixed bug in code for handling temporary memory usage when there are more
5118 back references than supplied space in the ovector. This could cause segfaults.
5119
5120
5121 Version 0.97 21-Oct-97
5122 ----------------------
5123
5124 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5125
5126 2. Optimized negated single characters not to use a bit map.
5127
5128 3. Brought error texts together as macro definitions; clarified some of them;
5129 fixed one that was wrong - it said "range out of order" when it meant "invalid
5130 escape sequence".
5131
5132 4. Changed some char * arguments to const char *.
5133
5134 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5135
5136 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5137 pcretest.
5138
5139
5140 Version 0.96 16-Oct-97
5141 ----------------------
5142
5143 1. Added a simple "pgrep" utility to the distribution.
5144
5145 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5146 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5147 where "ddd" means "one or more decimal digits".
5148
5149 3. Fixed serious bug. If a pattern had a back reference, but the call to
5150 pcre_exec() didn't supply a large enough ovector to record the related
5151 identifying subpattern, the match always failed. PCRE now remembers the number
5152 of the largest back reference, and gets some temporary memory in which to save
5153 the offsets during matching if necessary, in order to ensure that
5154 backreferences always work.
5155
5156 4. Increased the compatibility with Perl in a number of ways:
5157
5158 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5159 to request this handling. The option can be set at compile or exec time.
5160
5161 (b) $ matches before a terminating newline by default; an option
5162 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5163 mode). The option can be set at compile or exec time.
5164
5165 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5166 the same as Perl's. If the decimal number it represents is less than 10
5167 or there aren't that many previous left capturing parentheses, an octal
5168 escape is read. Inside a character class, it's always an octal escape,
5169 even if it is a single digit.
5170
5171 (d) An escaped but undefined alphabetic character is taken as a literal,
5172 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5173 escapes.
5174
5175 (e) {0} is now permitted. (The previous item is removed from the compiled
5176 pattern).
5177
5178 5. Changed all the names of code files so that the basic parts are no longer
5179 than 10 characters, and abolished the teeny "globals.c" file.
5180
5181 6. Changed the handling of character classes; they are now done with a 32-byte
5182 bit map always.
5183
5184 7. Added the -d and /D options to pcretest to make it possible to look at the
5185 internals of compilation without having to recompile pcre.
5186
5187
5188 Version 0.95 23-Sep-97
5189 ----------------------
5190
5191 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5192 \x20 at the start of a run of normal characters. These were being treated as
5193 real characters, instead of the source characters being re-checked.
5194
5195
5196 Version 0.94 18-Sep-97
5197 ----------------------
5198
5199 1. The functions are now thread-safe, with the caveat that the global variables
5200 containing pointers to malloc() and free() or alternative functions are the
5201 same for all threads.
5202
5203 2. Get pcre_study() to generate a bitmap of initial characters for non-
5204 anchored patterns when this is possible, and use it if passed to pcre_exec().
5205
5206
5207 Version 0.93 15-Sep-97
5208 ----------------------
5209
5210 1. /(b)|(:+)/ was computing an incorrect first character.
5211
5212 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5213 but not actually doing anything yet.
5214
5215 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5216 as Perl does (e.g. [-az] or [az-]).
5217
5218 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5219 all possible positions.
5220
5221 5. Split up into different modules to avoid including unneeded functions in a
5222 compiled binary. However, compile and exec are still in one module. The "study"
5223 function is split off.
5224
5225 6. The character tables are now in a separate module whose source is generated
5226 by an auxiliary program - but can then be edited by hand if required. There are
5227 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5228 toupper() in the code.
5229
5230 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5231 make them global. Abolish the function for setting them, as the caller can now
5232 set them directly.
5233
5234
5235 Version 0.92 11-Sep-97
5236 ----------------------
5237
5238 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5239 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5240
5241 2. Caseless matching was not working in character classes if the characters in
5242 the pattern were in upper case.
5243
5244 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5245
5246 4. Make PCRE_ANCHORED public and accept as a compile option.
5247
5248 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5249 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5250 pass them.
5251
5252 6. Give an error if bad option bits passed at compile or run time.
5253
5254 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5255 pcretest to cause it to pass that flag.
5256
5257 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5258 options, and the first character, if set.
5259
5260 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5261
5262
5263 Version 0.91 10-Sep-97
5264 ----------------------
5265
5266 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5267 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5268
5269 2. PCRE was looping on encountering an indefinitely repeated back reference to
5270 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5271 Perl does - treats the match as successful.
5272
5273 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5