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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1378 - (show annotations)
Mon Oct 14 06:55:28 2013 UTC (6 years, 5 months ago) by zherczeg
File size: 254625 byte(s)
Error occurred while calculating annotation data.
Add JIT support for the 64 bit TileGX architecture. Patch by Jiong Wang (Tilera Corporation).
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 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
127 character types such as \d or \w, too many callouts were inserted, and the
128 data that they returned was rubbish.
129
130 26. In UCP mode, \s was not matching two of the characters that Perl matches,
131 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
132 were matched by \h. The code has now been refactored so that the lists of
133 the horizontal and vertical whitespace characters used for \h and \v (which
134 are defined only in one place) are now also used for \s.
135
136 27. Add JIT support for the 64 bit TileGX architecture.
137 Patch by Jiong Wang (Tilera Corporation).
138
139
140 Version 8.33 28-May-2013
141 ------------------------
142
143 1. Added 'U' to some constants that are compared to unsigned integers, to
144 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
145 variables that are added to signed variables, to ensure the result is
146 signed and can be negated.
147
148 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
149 CMake config files.
150
151 3. Revise the creation of config.h.generic so that all boolean macros are
152 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
153 overriding via -D on the command line possible.
154
155 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
156 to unsigned int is reported to make a quite noticeable speed difference in
157 a specific Windows environment. Testing on Linux did also appear to show
158 some benefit (and it is clearly not harmful). Also fixed the definition of
159 Xop which should be unsigned.
160
161 5. Related to (4), changing the definition of the intermediate variable cc
162 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
163 improvements.
164
165 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
166 unnecessary spaces.
167
168 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
169 and later.
170
171 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
172
173 9. Optimizing fast_forward_start_bits in JIT.
174
175 10. Adding support for callouts in JIT, and fixing some issues revealed
176 during this work. Namely:
177
178 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
179
180 (b) Minimum length was not checked before the matching is started.
181
182 11. The value of capture_last that is passed to callouts was incorrect in some
183 cases when there was a capture on one path that was subsequently abandoned
184 after a backtrack. Also, the capture_last value is now reset after a
185 recursion, since all captures are also reset in this case.
186
187 12. The interpreter no longer returns the "too many substrings" error in the
188 case when an overflowing capture is in a branch that is subsequently
189 abandoned after a backtrack.
190
191 13. In the pathological case when an offset vector of size 2 is used, pcretest
192 now prints out the matched string after a yield of 0 or 1.
193
194 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
195 Only supported by the JIT compiler at the moment.
196
197 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
198
199 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
200 offset of the starting point of the matching process, provided the offsets
201 vector is large enough.
202
203 17. The \A escape now records a lookbehind value of 1, though its execution
204 does not actually inspect the previous character. This is to ensure that,
205 in partial multi-segment matching, at least one character from the old
206 segment is retained when a new segment is processed. Otherwise, if there
207 are no lookbehinds in the pattern, \A might match incorrectly at the start
208 of a new segment.
209
210 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
211
212 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
213 gives some modest performance improvement in 8-bit mode.
214
215 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
216 be expressed in certain programming languages using Universal Character
217 Names.
218
219 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
220 which points out that "non characters" are not "characters that may not
221 appear in Unicode strings" but rather "characters that are reserved for
222 internal use and have only local meaning".
223
224 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
225 there was a conditional group that depended on an assertion, if the
226 assertion was false, the callout that immediately followed the alternation
227 in the condition was skipped when pcre_exec() was used for matching.
228
229 23. Allow an explicit callout to be inserted before an assertion that is the
230 condition for a conditional group, for compatibility with automatic
231 callouts, which always insert a callout at this point.
232
233 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
234 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
235
236 25. (*PRUNE) is now supported by the JIT compiler.
237
238 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
239
240 27. Fix the case where there are two or more SKIPs with arguments that may be
241 ignored.
242
243 28. (*SKIP) is now supported by the JIT compiler.
244
245 29. (*THEN) is now supported by the JIT compiler.
246
247 30. Update RunTest with additional test selector options.
248
249 31. The way PCRE handles backtracking verbs has been changed in two ways.
250
251 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
252 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
253 backtracking. In some cases this makes it more Perl-compatible, but Perl's
254 rather obscure rules do not always do the same thing.
255
256 (2) Previously, backtracking verbs were confined within assertions. This is
257 no longer the case for positive assertions, except for (*ACCEPT). Again,
258 this sometimes improves Perl compatibility, and sometimes does not.
259
260 32. A number of tests that were in test 2 because Perl did things differently
261 have been moved to test 1, because either Perl or PCRE has changed, and
262 these tests are now compatible.
263
264 32. Backtracking control verbs are now handled in the same way in JIT and
265 interpreter.
266
267 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
268 contained a forward subroutine reference caused a compile error.
269
270 34. Auto-detect and optimize limited repetitions in JIT.
271
272 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
273 blocking (*UTF) etc.
274
275 36. In the interpreter, maximizing pattern repetitions for characters and
276 character types now use tail recursion, which reduces stack usage.
277
278 37. The value of the max lookbehind was not correctly preserved if a compiled
279 and saved regex was reloaded on a host of different endianness.
280
281 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
282 of the compiled pattern block, expand the flags field from 16 to 32 bits
283 because it was almost full.
284
285 39. Try madvise first before posix_madvise.
286
287 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
288 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
289 original change locked out all matches of empty strings. This has been
290 changed so that one match of an empty string per line is recognized.
291 Subsequent searches on the same line (for colouring or for --only-matching,
292 for example) do not recognize empty strings.
293
294 41. Applied a user patch to fix a number of spelling mistakes in comments.
295
296 42. Data lines longer than 65536 caused pcretest to crash.
297
298 43. Clarified the data type for length and startoffset arguments for pcre_exec
299 and pcre_dfa_exec in the function-specific man pages, where they were
300 explicitly stated to be in bytes, never having been updated. I also added
301 some clarification to the pcreapi man page.
302
303 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
304 a segmentation fault.
305
306
307 Version 8.32 30-November-2012
308 -----------------------------
309
310 1. Improved JIT compiler optimizations for first character search and single
311 character iterators.
312
313 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
314 Patch by Daniel Richard G.
315
316 3. Single character iterator optimizations in the JIT compiler.
317
318 4. Improved JIT compiler optimizations for character ranges.
319
320 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
321 Reported by Giuseppe D'Angelo.
322
323 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
324 of a line, was being set incorrectly in cases where .* appeared inside
325 atomic brackets at the start of a pattern, or where there was a subsequent
326 *PRUNE or *SKIP.
327
328 7. Improved instruction cache flush for POWER/PowerPC.
329 Patch by Daniel Richard G.
330
331 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
332 grep:
333
334 (a) There is now no limit to the number of patterns to be matched.
335
336 (b) An error is given if a pattern is too long.
337
338 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
339 are now supported.
340
341 (d) --exclude-from and --include-from (multiple use) have been added.
342
343 (e) Exclusions and inclusions now apply to all files and directories, not
344 just to those obtained from scanning a directory recursively.
345
346 (f) Multiple uses of -f and --file-list are now supported.
347
348 (g) In a Windows environment, the default for -d has been changed from
349 "read" (the GNU grep default) to "skip", because otherwise the presence
350 of a directory in the file list provokes an error.
351
352 (h) The documentation has been revised and clarified in places.
353
354 9. Improve the matching speed of capturing brackets.
355
356 10. Changed the meaning of \X so that it now matches a Unicode extended
357 grapheme cluster.
358
359 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
360 out POSIX threads when JIT support is configured.
361
362 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
363
364 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
365 structure could go wrong in environments where size_t is not the same size
366 as int.
367
368 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
369 set.
370
371 15. The EBCDIC support had decayed; later updates to the code had included
372 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
373 general tidy up of EBCDIC-related issues, and the documentation was also
374 not quite right. There is now a test that can be run on ASCII systems to
375 check some of the EBCDIC-related things (but is it not a full test).
376
377 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
378 in a small tidy to the code.
379
380 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
381
382 18. If the --only-matching (-o) option in pcregrep is specified multiple
383 times, each one causes appropriate output. For example, -o1 -o2 outputs the
384 substrings matched by the 1st and 2nd capturing parentheses. A separating
385 string can be specified by --om-separator (default empty).
386
387 19. Improving the first n character searches.
388
389 20. Turn case lists for horizontal and vertical white space into macros so that
390 they are defined only once.
391
392 21. This set of changes together give more compatible Unicode case-folding
393 behaviour for characters that have more than one other case when UCP
394 support is available.
395
396 (a) The Unicode property table now has offsets into a new table of sets of
397 three or more characters that are case-equivalent. The MultiStage2.py
398 script that generates these tables (the pcre_ucd.c file) now scans
399 CaseFolding.txt instead of UnicodeData.txt for character case
400 information.
401
402 (b) The code for adding characters or ranges of characters to a character
403 class has been abstracted into a generalized function that also handles
404 case-independence. In UTF-mode with UCP support, this uses the new data
405 to handle characters with more than one other case.
406
407 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
408 whose other case is greater than 256 are now correctly matched
409 caselessly. Previously, the high codepoint matched the low one, but not
410 vice versa.
411
412 (d) The processing of \h, \H, \v, and \ in character classes now makes use
413 of the new class addition function, using character lists defined as
414 macros alongside the case definitions of 20 above.
415
416 (e) Caseless back references now work with characters that have more than
417 one other case.
418
419 (f) General caseless matching of characters with more than one other case
420 is supported.
421
422 22. Unicode character properties were updated from Unicode 6.2.0
423
424 23. Improved CMake support under Windows. Patch by Daniel Richard G.
425
426 24. Add support for 32-bit character strings, and UTF-32
427
428 25. Major JIT compiler update (code refactoring and bugfixing).
429 Experimental Sparc 32 support is added.
430
431 26. Applied a modified version of Daniel Richard G's patch to create
432 pcre.h.generic and config.h.generic by "make" instead of in the
433 PrepareRelease script.
434
435 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
436 pcre_compile.c when checking for a zero character.
437
438 28. Introducing a native interface for JIT. Through this interface, the compiled
439 machine code can be directly executed. The purpose of this interface is to
440 provide fast pattern matching, so several sanity checks are not performed.
441 However, feature tests are still performed. The new interface provides
442 1.4x speedup compared to the old one.
443
444 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
445 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
446 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
447
448 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
449 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
450 the "old" RFC 2279). Instead, it ended up passing a negative length to
451 pcre_exec().
452
453 31. Add support for GCC's visibility feature to hide internal functions.
454
455 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
456 "unknown -C option" after outputting 0 or 1.
457
458 33. There is now support for generating a code coverage report for the test
459 suite in environments where gcc is the compiler and lcov is installed. This
460 is mainly for the benefit of the developers.
461
462 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
463 unaddressable using valgrind annotations, allowing valgrind to detect
464 invalid memory accesses. This is mainly for the benefit of the developers.
465
466 25. (*UTF) can now be used to start a pattern in any of the three libraries.
467
468 26. Give configure error if --enable-cpp but no C++ compiler found.
469
470
471 Version 8.31 06-July-2012
472 -------------------------
473
474 1. Fixing a wrong JIT test case and some compiler warnings.
475
476 2. Removed a bashism from the RunTest script.
477
478 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
479 to unsigned type, result still unsigned" that was given by an MS compiler
480 on encountering the code "-sizeof(xxx)".
481
482 4. Partial matching support is added to the JIT compiler.
483
484 5. Fixed several bugs concerned with partial matching of items that consist
485 of more than one character:
486
487 (a) /^(..)\1/ did not partially match "aba" because checking references was
488 done on an "all or nothing" basis. This also applied to repeated
489 references.
490
491 (b) \R did not give a hard partial match if \r was found at the end of the
492 subject.
493
494 (c) \X did not give a hard partial match after matching one or more
495 characters at the end of the subject.
496
497 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
498 a partial match for the string "\r".
499
500 (e) When newline was set to CRLF, the metacharacter "." did not recognize
501 a partial match for a CR character at the end of the subject string.
502
503 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
504 running pcretest, the text "(JIT)" added to the output whenever JIT is
505 actually used to run the match.
506
507 7. Individual JIT compile options can be set in pcretest by following -s+[+]
508 or /S+[+] with a digit between 1 and 7.
509
510 8. OP_NOT now supports any UTF character not just single-byte ones.
511
512 9. (*MARK) control verb is now supported by the JIT compiler.
513
514 10. The command "./RunTest list" lists the available tests without actually
515 running any of them. (Because I keep forgetting what they all are.)
516
517 11. Add PCRE_INFO_MAXLOOKBEHIND.
518
519 12. Applied a (slightly modified) user-supplied patch that improves performance
520 when the heap is used for recursion (compiled with --disable-stack-for-
521 recursion). Instead of malloc and free for each heap frame each time a
522 logical recursion happens, frames are retained on a chain and re-used where
523 possible. This sometimes gives as much as 30% improvement.
524
525 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
526 call.
527
528 14. As documented, (*COMMIT) is now confined to within a positive assertion.
529
530 15. It is now possible to link pcretest with libedit as an alternative to
531 libreadline.
532
533 16. (*COMMIT) control verb is now supported by the JIT compiler.
534
535 17. The Unicode data tables have been updated to Unicode 6.1.0.
536
537 18. Added --file-list option to pcregrep.
538
539 19. Added binary file support to pcregrep, including the -a, --binary-files,
540 -I, and --text options.
541
542 20. The madvise function is renamed for posix_madvise for QNX compatibility
543 reasons. Fixed by Giuseppe D'Angelo.
544
545 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
546
547 22. Changed the option for creating symbolic links for 16-bit man pages from
548 -s to -sf so that re-installing does not cause issues.
549
550 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
551
552 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
553 matches in certain environments (the workspace was not being correctly
554 retained). Also added to pcre_dfa_exec() a simple plausibility check on
555 some of the workspace data at the beginning of a restart.
556
557 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
558 was not doing so when it should - probably a typo introduced by SVN 528
559 (change 8.10/14).
560
561 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
562 \w+ when the character tables indicated that \x{c4} was a word character.
563 There were several related cases, all because the tests for doing a table
564 lookup were testing for characters less than 127 instead of 255.
565
566 27. If a pattern contains capturing parentheses that are not used in a match,
567 their slots in the ovector are set to -1. For those that are higher than
568 any matched groups, this happens at the end of processing. In the case when
569 there were back references that the ovector was too small to contain
570 (causing temporary malloc'd memory to be used during matching), and the
571 highest capturing number was not used, memory off the end of the ovector
572 was incorrectly being set to -1. (It was using the size of the temporary
573 memory instead of the true size.)
574
575 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
576 ovector size, it uses memory at the end of the block that it has got.
577
578 29. Check for an overlong MARK name and give an error at compile time. The
579 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
580
581 30. JIT compiler update.
582
583 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
584 Rill for the patch.
585
586 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
587
588 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
589
590 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
591 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
592 was enabled.
593
594 35. Improve JIT code generation for greedy plus quantifier.
595
596 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
597 1 to "aa" instead of to an empty string. The bug affected repeated groups
598 that could potentially match an empty string.
599
600 37. Optimizing single character iterators in JIT.
601
602 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
603 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
604 codepoints that are too big for the mode are faulted, and in a UTF mode,
605 disallowed codepoints are also faulted.
606
607 39. If PCRE was compiled with UTF support, in three places in the DFA
608 matcher there was code that should only have been obeyed in UTF mode, but
609 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
610 processing when bytes with values greater than 127 were present. In 16-bit
611 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
612 both cases the values are those that cannot be the first data item in a UTF
613 character. The three items that might have provoked this were recursions,
614 possessively repeated groups, and atomic groups.
615
616 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
617 and pcregrep, because some OS require shared objects to be explicitly
618 passed to ld, causing the link step to fail if they are not.
619
620 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
621 mode, patterns that started with \h* or \R* might be incorrectly matched.
622
623
624 Version 8.30 04-February-2012
625 -----------------------------
626
627 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
628 name is defined in ctype.h.
629
630 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
631 only in quite long subpatterns.
632
633 3. Removed the function pcre_info(), which has been obsolete and deprecated
634 since it was replaced by pcre_fullinfo() in February 2000.
635
636 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
637 match a (*MARK), and the match failed at the start of the subject, a
638 reference to memory before the start of the subject could occur. This bug
639 was introduced by fix 17 of release 8.21.
640
641 5. A reference to an unset group with zero minimum repetition was giving
642 totally wrong answers (in non-JavaScript-compatibility mode). For example,
643 /(another)?(\1?)test/ matched against "hello world test". This bug was
644 introduced in release 8.13.
645
646 6. Add support for 16-bit character strings (a large amount of work involving
647 many changes and refactorings).
648
649 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
650 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
651 from a file.
652
653 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
654 rounding is not applied in this particular case).
655
656 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
657 if they appear, or are escaped, in patterns.
658
659 10. Get rid of a number of -Wunused-but-set-variable warnings.
660
661 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
662 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
663 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
664 also returns the mark "x". This bug applied to capturing parentheses,
665 non-capturing parentheses, and atomic parentheses. It also applied to some
666 assertions.
667
668 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
669 information out of configure.ac instead of relying on pcre.h.generic, which
670 is not stored in the repository.
671
672 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
673 -lreadline.
674
675 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
676
677 15. Applied Graycode's patch to put the top-level frame on the stack rather
678 than the heap when not using the stack for recursion. This gives a
679 performance improvement in many cases when recursion is not deep.
680
681 16. Experimental code added to "pcretest -C" to output the stack frame size.
682
683
684 Version 8.21 12-Dec-2011
685 ------------------------
686
687 1. Updating the JIT compiler.
688
689 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
690 are added as well.
691
692 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
693 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
694 calling _pcre_jit_exec. Some extra comments are added.
695
696 4. (*MARK) settings inside atomic groups that do not contain any capturing
697 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
698 was introduced by change 18 for 8.20.
699
700 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
701 ECMA-262 standard.
702
703 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
704 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
705 This bug was probably introduced by change 9 of 8.13.
706
707 7. While fixing 6 above, I noticed that a number of other items were being
708 incorrectly rejected as "not fixed length". This arose partly because newer
709 opcodes had not been added to the fixed-length checking code. I have (a)
710 corrected the bug and added tests for these items, and (b) arranged for an
711 error to occur if an unknown opcode is encountered while checking for fixed
712 length instead of just assuming "not fixed length". The items that were
713 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
714 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
715 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
716
717 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
718 being incorrectly compiled and would have given unpredicatble results.
719
720 9. A possessively repeated subpattern with minimum repeat count greater than
721 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
722 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
723 the first (A) could occur when it should not.
724
725 10. Add a cast and remove a redundant test from the code.
726
727 11. JIT should use pcre_malloc/pcre_free for allocation.
728
729 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
730 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
731 is anything other than /usr, -L is still shown).
732
733 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
734
735 14. Perl does not support \N without a following name in a [] class; PCRE now
736 also gives an error.
737
738 15. If a forward reference was repeated with an upper limit of around 2000,
739 it caused the error "internal error: overran compiling workspace". The
740 maximum number of forward references (including repeats) was limited by the
741 internal workspace, and dependent on the LINK_SIZE. The code has been
742 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
743 the default depends on LINK_SIZE. There is a new upper limit (for safety)
744 of around 200,000 forward references. While doing this, I also speeded up
745 the filling in of repeated forward references.
746
747 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
748 incorrectly expecting the subject to contain another "a" after the start.
749
750 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
751 in the match, the SKIP should be ignored. This was not happening; instead
752 the SKIP was being treated as NOMATCH. For patterns such as
753 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
754 tested.
755
756 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
757 now much more compatible with Perl, in particular in cases where the result
758 is a non-match for a non-anchored pattern. For example, if
759 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
760 "m", where previously it did not return a name. A side effect of this
761 change is that for partial matches, the last encountered mark name is
762 returned, as for non matches. A number of tests that were previously not
763 Perl-compatible have been moved into the Perl-compatible test files. The
764 refactoring has had the pleasing side effect of removing one argument from
765 the match() function, thus reducing its stack requirements.
766
767 19. If the /S+ option was used in pcretest to study a pattern using JIT,
768 subsequent uses of /S (without +) incorrectly behaved like /S+.
769
770 21. Retrieve executable code size support for the JIT compiler and fixing
771 some warnings.
772
773 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
774 not work when the shorter character appeared right at the end of the
775 subject string.
776
777 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
778 systems.
779
780 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
781 output it when the /M option is used in pcretest.
782
783 25. The CheckMan script was not being included in the distribution. Also, added
784 an explicit "perl" to run Perl scripts from the PrepareRelease script
785 because this is reportedly needed in Windows.
786
787 26. If study data was being save in a file and studying had not found a set of
788 "starts with" bytes for the pattern, the data written to the file (though
789 never used) was taken from uninitialized memory and so caused valgrind to
790 complain.
791
792 27. Updated RunTest.bat as provided by Sheri Pierce.
793
794 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
795
796 29. Computation of memory usage for the table of capturing group names was
797 giving an unnecessarily large value.
798
799
800 Version 8.20 21-Oct-2011
801 ------------------------
802
803 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
804 a POSIX class. After further experiments with Perl, which convinced me that
805 Perl has bugs and confusions, a closing square bracket is no longer allowed
806 in a POSIX name. This bug also affected patterns with classes that started
807 with full stops.
808
809 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
810 captured substring, but while checking the failing first alternative,
811 substring 1 is temporarily captured. If the output vector supplied to
812 pcre_exec() was not big enough for this capture, the yield of the function
813 was still zero ("insufficient space for captured substrings"). This cannot
814 be totally fixed without adding another stack variable, which seems a lot
815 of expense for a edge case. However, I have improved the situation in cases
816 such as /(a)(b)x|abc/ matched against "abc", where the return code
817 indicates that fewer than the maximum number of slots in the ovector have
818 been set.
819
820 3. Related to (2) above: when there are more back references in a pattern than
821 slots in the output vector, pcre_exec() uses temporary memory during
822 matching, and copies in the captures as far as possible afterwards. It was
823 using the entire output vector, but this conflicts with the specification
824 that only 2/3 is used for passing back captured substrings. Now it uses
825 only the first 2/3, for compatibility. This is, of course, another edge
826 case.
827
828 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
829 main code base, and can be used by building with --enable-jit. When this is
830 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
831 runtime --no-jit option is given.
832
833 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
834 ovector, the return from the function was zero, implying that there were
835 other matches that did not fit. The correct "exactly full" value is now
836 returned.
837
838 6. If a subpattern that was called recursively or as a subroutine contained
839 (*PRUNE) or any other control that caused it to give a non-standard return,
840 invalid errors such as "Error -26 (nested recursion at the same subject
841 position)" or even infinite loops could occur.
842
843 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
844 computing the minimum length on reaching *ACCEPT, and so ended up with the
845 wrong value of 1 rather than 0. Further investigation indicates that
846 computing a minimum subject length in the presence of *ACCEPT is difficult
847 (think back references, subroutine calls), and so I have changed the code
848 so that no minimum is registered for a pattern that contains *ACCEPT.
849
850 8. If (*THEN) was present in the first (true) branch of a conditional group,
851 it was not handled as intended. [But see 16 below.]
852
853 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
854 Sheri Pierce.
855
856 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
857 the first byte in a match must be "a".
858
859 11. Change 17 for 8.13 increased the recursion depth for patterns like
860 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
861 pattern contains any instances of (*THEN). If it does not, the old
862 optimizations are restored. It would be nice to do this on a per-group
863 basis, but at the moment that is not feasible.
864
865 12. In some environments, the output of pcretest -C is CRLF terminated. This
866 broke RunTest's code that checks for the link size. A single white space
867 character after the value is now allowed for.
868
869 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
870 For "fr", it uses the Windows-specific input and output files.
871
872 14. If (*THEN) appeared in a group that was called recursively or as a
873 subroutine, it did not work as intended. [But see next item.]
874
875 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
876 pattern fragments (but not containing any | characters). If A and B are
877 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
878 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
879 D. In other words, Perl considers parentheses that do not contain any |
880 characters to be part of a surrounding alternative, whereas PCRE was
881 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
882 differently. PCRE now behaves in the same way as Perl, except in the case
883 of subroutine/recursion calls such as (?1) which have in any case always
884 been different (but PCRE had them first :-).
885
886 16. Related to 15 above: Perl does not treat the | in a conditional group as
887 creating alternatives. Such a group is treated in the same way as an
888 ordinary group without any | characters when processing (*THEN). PCRE has
889 been changed to match Perl's behaviour.
890
891 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
892 RunGrepTest script failed.
893
894 18. Change 22 for version 13 caused atomic groups to use more stack. This is
895 inevitable for groups that contain captures, but it can lead to a lot of
896 stack use in large patterns. The old behaviour has been restored for atomic
897 groups that do not contain any capturing parentheses.
898
899 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
900 suppress the check for a minimum subject length at run time. (If it was
901 given to pcre_exec() or pcre_dfa_exec() it did work.)
902
903 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
904 fail to work when decoding hex characters in data strings in EBCDIC
905 environments.
906
907 21. It appears that in at least one Mac OS environment, the isxdigit() function
908 is implemented as a macro that evaluates to its argument more than once,
909 contravening the C 90 Standard (I haven't checked a later standard). There
910 was an instance in pcretest which caused it to go wrong when processing
911 \x{...} escapes in subject strings. The has been rewritten to avoid using
912 things like p++ in the argument of isxdigit().
913
914
915 Version 8.13 16-Aug-2011
916 ------------------------
917
918 1. The Unicode data tables have been updated to Unicode 6.0.0.
919
920 2. Two minor typos in pcre_internal.h have been fixed.
921
922 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
923 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
924 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
925
926 4. There were a number of related bugs in the code for matching backrefences
927 caselessly in UTF-8 mode when codes for the characters concerned were
928 different numbers of bytes. For example, U+023A and U+2C65 are an upper
929 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
930 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
931 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
932 2-byte code at the end of the subject (it thought there wasn't enough data
933 left).
934
935 5. Comprehensive information about what went wrong is now returned by
936 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
937 as the output vector has at least 2 elements. The offset of the start of
938 the failing character and a reason code are placed in the vector.
939
940 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
941 now returned is for the first byte of the failing character, instead of the
942 last byte inspected. This is an incompatible change, but I hope it is small
943 enough not to be a problem. It makes the returned offset consistent with
944 pcre_exec() and pcre_dfa_exec().
945
946 7. pcretest now gives a text phrase as well as the error number when
947 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
948 failure, the offset and reason code are output.
949
950 8. When \R was used with a maximizing quantifier it failed to skip backwards
951 over a \r\n pair if the subsequent match failed. Instead, it just skipped
952 back over a single character (\n). This seems wrong (because it treated the
953 two characters as a single entity when going forwards), conflicts with the
954 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
955 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
956 has been changed.
957
958 9. Some internal refactoring has changed the processing so that the handling
959 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
960 time (the PCRE_DOTALL option was changed this way some time ago: version
961 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
962 which was always a bit of a fudge. It also means that there is one less
963 argument for the match() function, which reduces its stack requirements
964 slightly. This change also fixes an incompatibility with Perl: the pattern
965 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
966
967 10. More internal refactoring has drastically reduced the number of recursive
968 calls to match() for possessively repeated groups such as (abc)++ when
969 using pcre_exec().
970
971 11. While implementing 10, a number of bugs in the handling of groups were
972 discovered and fixed:
973
974 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
975 (a|)*(?1) gave a compile-time internal error.
976 ((a|)+)+ did not notice that the outer group could match an empty string.
977 (^a|^)+ was not marked as anchored.
978 (.*a|.*)+ was not marked as matching at start or after a newline.
979
980 12. Yet more internal refactoring has removed another argument from the match()
981 function. Special calls to this function are now indicated by setting a
982 value in a variable in the "match data" data block.
983
984 13. Be more explicit in pcre_study() instead of relying on "default" for
985 opcodes that mean there is no starting character; this means that when new
986 ones are added and accidentally left out of pcre_study(), testing should
987 pick them up.
988
989 14. The -s option of pcretest has been documented for ages as being an old
990 synonym of -m (show memory usage). I have changed it to mean "force study
991 for every regex", that is, assume /S for every regex. This is similar to -i
992 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
993 using it. It makes it easier to run collections of tests with and without
994 study enabled, and thereby test pcre_study() more easily. All the standard
995 tests are now run with and without -s (but some patterns can be marked as
996 "never study" - see 20 below).
997
998 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
999 restoration of the capturing data to the outer values was not happening
1000 correctly.
1001
1002 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1003 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1004 pattern had matched an empty string, and so incorrectly returned a no
1005 match.
1006
1007 17. There was optimizing code for the last branch of non-capturing parentheses,
1008 and also for the obeyed branch of a conditional subexpression, which used
1009 tail recursion to cut down on stack usage. Unfortunately, now that there is
1010 the possibility of (*THEN) occurring in these branches, tail recursion is
1011 no longer possible because the return has to be checked for (*THEN). These
1012 two optimizations have therefore been removed. [But see 8.20/11 above.]
1013
1014 18. If a pattern containing \R was studied, it was assumed that \R always
1015 matched two bytes, thus causing the minimum subject length to be
1016 incorrectly computed because \R can also match just one byte.
1017
1018 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1019 was incorrectly computed.
1020
1021 20. If /S is present twice on a test pattern in pcretest input, it now
1022 *disables* studying, thereby overriding the use of -s on the command line
1023 (see 14 above). This is necessary for one or two tests to keep the output
1024 identical in both cases.
1025
1026 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1027 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1028
1029 22. When an atomic group that contained a capturing parenthesis was
1030 successfully matched, but the branch in which it appeared failed, the
1031 capturing was not being forgotten if a higher numbered group was later
1032 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1033 group 1 to "a", when in fact it should be unset. This applied to multi-
1034 branched capturing and non-capturing groups, repeated or not, and also to
1035 positive assertions (capturing in negative assertions does not happen
1036 in PCRE) and also to nested atomic groups.
1037
1038 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1039 subject after a captured substring, to make it easier to tell which of a
1040 number of identical substrings has been captured.
1041
1042 24. The way atomic groups are processed by pcre_exec() has been changed so that
1043 if they are repeated, backtracking one repetition now resets captured
1044 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1045 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1046 "aaa". Previously, it would have been "a". As part of this code
1047 refactoring, the way recursive calls are handled has also been changed.
1048
1049 25. If an assertion condition captured any substrings, they were not passed
1050 back unless some other capturing happened later. For example, if
1051 (?(?=(a))a) was matched against "a", no capturing was returned.
1052
1053 26. When studying a pattern that contained subroutine calls or assertions,
1054 the code for finding the minimum length of a possible match was handling
1055 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1056 group 1 called group 2 while simultaneously a separate group 2 called group
1057 1). A stack overflow occurred in this case. I have fixed this by limiting
1058 the recursion depth to 10.
1059
1060 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1061 Fortmann. This supports explicit test numbers on the command line, and has
1062 argument validation and error reporting.
1063
1064 28. An instance of \X with an unlimited repeat could fail if at any point the
1065 first character it looked at was a mark character.
1066
1067 29. Some minor code refactoring concerning Unicode properties and scripts
1068 should reduce the stack requirement of match() slightly.
1069
1070 30. Added the '=' option to pcretest to check the setting of unused capturing
1071 slots at the end of the pattern, which are documented as being -1, but are
1072 not included in the return count.
1073
1074 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1075 compiled something random. Now it gives a compile-time error (as does
1076 Perl).
1077
1078 32. A *MARK encountered during the processing of a positive assertion is now
1079 recorded and passed back (compatible with Perl).
1080
1081 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1082 had alternative anchored branches, the search for a second match in a line
1083 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1084 matched the line "0102" twice. The same bug affected patterns that started
1085 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1086 twice.
1087
1088 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1089 does, and because of capturing effects, quantifying parenthesized
1090 assertions may at times be useful. Quantifiers are now allowed for
1091 parenthesized assertions.
1092
1093 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1094
1095 36. \g was being checked for fancy things in a character class, when it should
1096 just be a literal "g".
1097
1098 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1099 appearance of a nested POSIX class supersedes an apparent external class.
1100 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1101 unescaped square brackets may also appear as part of class names. For
1102 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1103 more like Perl. (But see 8.20/1 above.)
1104
1105 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1106 was because it thought it was \N{name}, which is not supported).
1107
1108 39. Add minix to OS list not supporting the -S option in pcretest.
1109
1110 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1111 cannot analyze patterns in sufficient detail to catch mutual recursions
1112 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1113 subgroup is called recursively as a subpattern for a second time at the
1114 same position in the subject string. In previous releases this might have
1115 been caught by the recursion limit, or it might have run out of stack.
1116
1117 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1118 happen only once. PCRE was, however incorrectly giving a compile time error
1119 "recursive call could loop indefinitely" because it cannot analyze the
1120 pattern in sufficient detail. The compile time test no longer happens when
1121 PCRE is compiling a conditional subpattern, but actual runaway loops are
1122 now caught at runtime (see 40 above).
1123
1124 42. It seems that Perl allows any characters other than a closing parenthesis
1125 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1126 has been changed to be the same.
1127
1128 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1129 as not to get warnings when autogen.sh is called. Also changed
1130 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1131
1132 44. To help people who use pcregrep to scan files containing exceedingly long
1133 lines, the following changes have been made:
1134
1135 (a) The default value of the buffer size parameter has been increased from
1136 8K to 20K. (The actual buffer used is three times this size.)
1137
1138 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1139 PCRE is built.
1140
1141 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1142 to be set at run time.
1143
1144 (d) Numerical values in pcregrep options can be followed by K or M, for
1145 example --buffer-size=50K.
1146
1147 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1148 given and the return code is set to 2.
1149
1150 45. Add a pointer to the latest mark to the callout data block.
1151
1152 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1153 partial match of an empty string instead of no match. This was specific to
1154 the use of ".".
1155
1156 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1157 complete match instead of a partial match. This bug was dependent on both
1158 the PCRE_UTF8 and PCRE_DOTALL options being set.
1159
1160 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1161 starting byte set, because \b was not being ignored.
1162
1163
1164 Version 8.12 15-Jan-2011
1165 ------------------------
1166
1167 1. Fixed some typos in the markup of the man pages, and wrote a script that
1168 checks for such things as part of the documentation building process.
1169
1170 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1171 --match-limit and --recursion-limit options (added for 8.11). In
1172 particular, this made one of the standard tests fail. (The integer value
1173 went into the wrong half of a long int.)
1174
1175 3. If the --colour option was given to pcregrep with -v (invert match), it
1176 did strange things, either producing crazy output, or crashing. It should,
1177 of course, ignore a request for colour when reporting lines that do not
1178 match.
1179
1180 4. Another pcregrep bug caused similar problems if --colour was specified with
1181 -M (multiline) and the pattern match finished with a line ending.
1182
1183 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1184 matched in multiline mode, the following line was shown as part of the
1185 match. This seems wrong, so I have changed it.
1186
1187 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1188 the check for further matches in the same line (so they could be coloured)
1189 to overrun the end of the current line. If another match was found, it was
1190 incorrectly shown (and then shown again when found in the next line).
1191
1192 7. If pcregrep was compiled under Windows, there was a reference to the
1193 function pcregrep_exit() before it was defined. I am assuming this was
1194 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1195 reported by a user. I've moved the definition above the reference.
1196
1197
1198 Version 8.11 10-Dec-2010
1199 ------------------------
1200
1201 1. (*THEN) was not working properly if there were untried alternatives prior
1202 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1203 backtracked to try for "b" instead of moving to the next alternative branch
1204 at the same level (in this case, to look for "c"). The Perl documentation
1205 is clear that when (*THEN) is backtracked onto, it goes to the "next
1206 alternative in the innermost enclosing group".
1207
1208 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1209 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1210 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1211 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1212 (*THEN).
1213
1214 3. If \s appeared in a character class, it removed the VT character from
1215 the class, even if it had been included by some previous item, for example
1216 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1217 of \s, but is part of the POSIX "space" class.)
1218
1219 4. A partial match never returns an empty string (because you can always
1220 match an empty string at the end of the subject); however the checking for
1221 an empty string was starting at the "start of match" point. This has been
1222 changed to the "earliest inspected character" point, because the returned
1223 data for a partial match starts at this character. This means that, for
1224 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1225 (previously it gave "no match").
1226
1227 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1228 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1229 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1230 has an implication that the given string is incomplete (because a partial
1231 match is preferred over a full match). For this reason, these items now
1232 give a partial match in this situation. [Aside: previously, the one case
1233 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1234 match rather than a full match, which was wrong by the old rules, but is
1235 now correct.]
1236
1237 6. There was a bug in the handling of #-introduced comments, recognized when
1238 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1239 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1240 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1241 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1242 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1243 places in pcre_compile().
1244
1245 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1246 comments when looking ahead for named forward references to subpatterns,
1247 the only newline sequence it recognized was NL. It now handles newlines
1248 according to the set newline convention.
1249
1250 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1251 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1252 cater for a lack of strerror(). These oversights have been fixed.
1253
1254 9. Added --match-limit and --recursion-limit to pcregrep.
1255
1256 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1257
1258 11. When the -o option was used, pcregrep was setting a return code of 1, even
1259 when matches were found, and --line-buffered was not being honoured.
1260
1261 12. Added an optional parentheses number to the -o and --only-matching options
1262 of pcregrep.
1263
1264 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1265 can match an empty string. The code to do it in pcretest and pcredemo
1266 needed fixing:
1267
1268 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1269 only one byte after an empty string match just before CRLF (this case
1270 just got forgotten; "any" and "anycrlf" were OK).
1271
1272 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1273 mode when an empty string match preceded an ASCII character followed by
1274 a non-ASCII character. (The code for advancing by one character rather
1275 than one byte was nonsense.)
1276
1277 (c) The pcredemo.c sample program did not have any code at all to handle
1278 the cases when CRLF is a valid newline sequence.
1279
1280 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1281 as a starting offset was within the subject string. There is now a new
1282 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1283 negative or greater than the length of the string. In order to test this,
1284 pcretest is extended to allow the setting of negative starting offsets.
1285
1286 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1287 starting offset points to the beginning of a UTF-8 character was
1288 unnecessarily clumsy. I tidied it up.
1289
1290 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1291 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1292
1293 17. Nobody had reported that the --include_dir option, which was added in
1294 release 7.7 should have been called --include-dir (hyphen, not underscore)
1295 for compatibility with GNU grep. I have changed it to --include-dir, but
1296 left --include_dir as an undocumented synonym, and the same for
1297 --exclude-dir, though that is not available in GNU grep, at least as of
1298 release 2.5.4.
1299
1300 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1301 characters from a string of bytes) have been redefined so as not to use
1302 loops, in order to improve performance in some environments. At the same
1303 time, I abstracted some of the common code into auxiliary macros to save
1304 repetition (this should not affect the compiled code).
1305
1306 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1307 compile-time error is now given if \c is not followed by an ASCII
1308 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1309 different, and any byte value is allowed.)
1310
1311 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1312 START_OPTIMIZE option, which is now allowed at compile time - but just
1313 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1314 to pcregrep and other applications that have no direct access to PCRE
1315 options. The new /Y option in pcretest sets this option when calling
1316 pcre_compile().
1317
1318 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1319 back references. Groups containing recursive back references were forced to
1320 be atomic by that change, but in the case of named groups, the amount of
1321 memory required was incorrectly computed, leading to "Failed: internal
1322 error: code overflow". This has been fixed.
1323
1324 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1325 pcretest.c, to avoid build problems in some Borland environments.
1326
1327
1328 Version 8.10 25-Jun-2010
1329 ------------------------
1330
1331 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1332 THEN.
1333
1334 2. (*ACCEPT) was not working when inside an atomic group.
1335
1336 3. Inside a character class, \B is treated as a literal by default, but
1337 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1338 causes the error). The code is unchanged, but I tidied the documentation.
1339
1340 4. Inside a character class, PCRE always treated \R and \X as literals,
1341 whereas Perl faults them if its -w option is set. I have changed PCRE so
1342 that it faults them when PCRE_EXTRA is set.
1343
1344 5. Added support for \N, which always matches any character other than
1345 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1346
1347 6. When compiling pcregrep with newer versions of gcc which may have
1348 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1349 declared with attribute warn_unused_result" were given. Just casting the
1350 result to (void) does not stop the warnings; a more elaborate fudge is
1351 needed. I've used a macro to implement this.
1352
1353 7. Minor change to pcretest.c to avoid a compiler warning.
1354
1355 8. Added four artifical Unicode properties to help with an option to make
1356 \s etc use properties (see next item). The new properties are: Xan
1357 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1358
1359 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1360 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1361 this option. Modified pcretest to add /W to test this facility. Added
1362 REG_UCP to make it available via the POSIX interface.
1363
1364 10. Added --line-buffered to pcregrep.
1365
1366 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1367 studied, and the match started with a letter with a code point greater than
1368 127 whose first byte was different to the first byte of the other case of
1369 the letter, the other case of this starting letter was not recognized
1370 (#976).
1371
1372 12. If a pattern that was studied started with a repeated Unicode property
1373 test, for example, \p{Nd}+, there was the theoretical possibility of
1374 setting up an incorrect bitmap of starting bytes, but fortunately it could
1375 not have actually happened in practice until change 8 above was made (it
1376 added property types that matched character-matching opcodes).
1377
1378 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1379 possible starting bytes for non-anchored patterns.
1380
1381 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1382 \R, and also a number of cases that involve Unicode properties, both
1383 explicit and implicit when PCRE_UCP is set.
1384
1385 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1386 input, it could crash or give wrong results if characters with values
1387 greater than 0xc0 were present in the subject string. (Detail: it assumed
1388 UTF-8 input when processing these items.)
1389
1390 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1391 size_t is 64-bit (#991).
1392
1393 17. Added a check for running out of memory when PCRE is compiled with
1394 --disable-stack-for-recursion (#990).
1395
1396 18. If the last data line in a file for pcretest does not have a newline on
1397 the end, a newline was missing in the output.
1398
1399 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1400 less than 128) in its various bitmaps. However, there is a facility for
1401 generating tables according to the current locale when PCRE is compiled. It
1402 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1403 characters, are recognized by isspace() and therefore were getting set in
1404 these tables, and indeed these tables seem to approximate to ISO 8859. This
1405 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1406 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1407 which of course cannot start UTF-8 characters. I have changed the code so
1408 that only real ASCII characters (less than 128) and the correct starting
1409 bytes for UTF-8 encodings are set for characters greater than 127 when in
1410 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1411 altogether.)
1412
1413 20. Added the /T option to pcretest so as to be able to run tests with non-
1414 standard character tables, thus making it possible to include the tests
1415 used for 19 above in the standard set of tests.
1416
1417 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1418 reference to a subpattern the other side of a comment that contains an
1419 opening parenthesis caused either an internal compiling error, or a
1420 reference to the wrong subpattern.
1421
1422
1423 Version 8.02 19-Mar-2010
1424 ------------------------
1425
1426 1. The Unicode data tables have been updated to Unicode 5.2.0.
1427
1428 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1429 configured.
1430
1431 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1432 original author of that file, following a query about its status.
1433
1434 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1435 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1436
1437 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1438 quantifier applied to a forward-referencing subroutine call, could compile
1439 incorrect code or give the error "internal error: previously-checked
1440 referenced subpattern not found".
1441
1442 6. Both MS Visual Studio and Symbian OS have problems with initializing
1443 variables to point to external functions. For these systems, therefore,
1444 pcre_malloc etc. are now initialized to local functions that call the
1445 relevant global functions.
1446
1447 7. There were two entries missing in the vectors called coptable and poptable
1448 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1449 I've fixed the data, and added a kludgy way of testing at compile time that
1450 the lengths are correct (equal to the number of opcodes).
1451
1452 8. Following on from 7, I added a similar kludge to check the length of the
1453 eint vector in pcreposix.c.
1454
1455 9. Error texts for pcre_compile() are held as one long string to avoid too
1456 much relocation at load time. To find a text, the string is searched,
1457 counting zeros. There was no check for running off the end of the string,
1458 which could happen if a new error number was added without updating the
1459 string.
1460
1461 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1462
1463 11. \K was not working if it appeared in an atomic group or in a group that
1464 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1465 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1466 the assertion is positive, but not if it is negative.
1467
1468 12. Change 11 fortuitously reduced the size of the stack frame used in the
1469 "match()" function of pcre_exec.c by one pointer. Forthcoming
1470 implementation of support for (*MARK) will need an extra pointer on the
1471 stack; I have reserved it now, so that the stack frame size does not
1472 decrease.
1473
1474 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1475 item in branch that calls a recursion is a subroutine call - as in the
1476 second branch in the above example - was incorrectly given the compile-
1477 time error "recursive call could loop indefinitely" because pcre_compile()
1478 was not correctly checking the subroutine for matching a non-empty string.
1479
1480 14. The checks for overrunning compiling workspace could trigger after an
1481 overrun had occurred. This is a "should never occur" error, but it can be
1482 triggered by pathological patterns such as hundreds of nested parentheses.
1483 The checks now trigger 100 bytes before the end of the workspace.
1484
1485 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1486
1487
1488 Version 8.01 19-Jan-2010
1489 ------------------------
1490
1491 1. If a pattern contained a conditional subpattern with only one branch (in
1492 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1493 computed the wrong minimum data length (which is of course zero for such
1494 subpatterns). This could cause incorrect "no match" results.
1495
1496 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1497 the pattern is reset in the first branch, pcre_compile() failed with
1498 "internal error: code overflow at offset...". This happened only when
1499 the reset was to the original external option setting. (An optimization
1500 abstracts leading options settings into an external setting, which was the
1501 cause of this.)
1502
1503 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1504 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1505 assertion pattern did not match (meaning that the assertion was true), it
1506 was incorrectly treated as false if the SKIP had been reached during the
1507 matching. This also applied to assertions used as conditions.
1508
1509 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1510 assertion subpattern, including such a pattern used as a condition,
1511 unpredictable results occurred, instead of the error return
1512 PCRE_ERROR_DFA_UITEM.
1513
1514 5. The C++ GlobalReplace function was not working like Perl for the special
1515 situation when an empty string is matched. It now does the fancy magic
1516 stuff that is necessary.
1517
1518 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1519 removed. (These were left over from very, very early versions of PCRE.)
1520
1521 7. Some cosmetic changes to the code to make life easier when compiling it
1522 as part of something else:
1523
1524 (a) Change DEBUG to PCRE_DEBUG.
1525
1526 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1527 called "current" as "current_branch", to prevent a collision with the
1528 Linux macro when compiled as a kernel module.
1529
1530 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1531 prevent a collision with the Linux macro when compiled as a kernel
1532 module.
1533
1534 8. In pcre_compile() there are some checks for integer overflows that used to
1535 cast potentially large values to (double). This has been changed to that
1536 when building, a check for int64_t is made, and if it is found, it is used
1537 instead, thus avoiding the use of floating point arithmetic. (There is no
1538 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1539 double.
1540
1541 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1542 2005 (difference between two addresses compared to an unsigned value).
1543
1544 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1545 custom one, because of the following reported problem in Windows:
1546
1547 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1548 under Win32.
1549 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1550 therefore missing the function definition.
1551 - The compiler thus generates a "C" signature for the test function.
1552 - The linker fails to find the "C" function.
1553 - PCRE fails to configure if asked to do so against libbz2.
1554
1555 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1556 messages were output:
1557
1558 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1559 rerunning libtoolize, to keep the correct libtool macros in-tree.
1560 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1561
1562 I have done both of these things.
1563
1564 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1565 most of the time, it *can* run out if it is given a pattern that contains a
1566 runaway infinite recursion. I updated the discussion in the pcrestack man
1567 page.
1568
1569 13. Now that we have gone to the x.xx style of version numbers, the minor
1570 version may start with zero. Using 08 or 09 is a bad idea because users
1571 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1572 interpreted as invalid octal numbers. I've updated the previous comment in
1573 configure.ac, and also added a check that gives an error if 08 or 09 are
1574 used.
1575
1576 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1577 causing partial matching to fail when the end of the subject matched \W
1578 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1579
1580 15. There were some discrepancies between the declarations in pcre_internal.h
1581 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1582 their definitions. The declarations used "const uschar *" and the
1583 definitions used USPTR. Even though USPTR is normally defined as "const
1584 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1585 reported that: "This difference in casting confuses some C++ compilers, for
1586 example, SunCC recognizes above declarations as different functions and
1587 generates broken code for hbpcre." I have changed the declarations to use
1588 USPTR.
1589
1590 16. GNU libtool is named differently on some systems. The autogen.sh script now
1591 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1592 (FreeBSD).
1593
1594 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1595 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1596 comment: "Figure out how to create a longlong from a string: strtoll and
1597 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1598 instance, but it only takes 2 args instead of 3!"
1599
1600 18. A subtle bug concerned with back references has been fixed by a change of
1601 specification, with a corresponding code fix. A pattern such as
1602 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1603 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1604 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1605 same bug. Such groups have to be quantified to be useful, or contained
1606 inside another quantified group. (If there's no repetition, the reference
1607 can never match.) The problem arises because, having left the group and
1608 moved on to the rest of the pattern, a later failure that backtracks into
1609 the group uses the captured value from the final iteration of the group
1610 rather than the correct earlier one. I have fixed this in PCRE by forcing
1611 any group that contains a reference to itself to be an atomic group; that
1612 is, there cannot be any backtracking into it once it has completed. This is
1613 similar to recursive and subroutine calls.
1614
1615
1616 Version 8.00 19-Oct-09
1617 ----------------------
1618
1619 1. The table for translating pcre_compile() error codes into POSIX error codes
1620 was out-of-date, and there was no check on the pcre_compile() error code
1621 being within the table. This could lead to an OK return being given in
1622 error.
1623
1624 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1625 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1626 in a Windows environment.
1627
1628 3. The pcregrep --count option prints the count for each file even when it is
1629 zero, as does GNU grep. However, pcregrep was also printing all files when
1630 --files-with-matches was added. Now, when both options are given, it prints
1631 counts only for those files that have at least one match. (GNU grep just
1632 prints the file name in this circumstance, but including the count seems
1633 more useful - otherwise, why use --count?) Also ensured that the
1634 combination -clh just lists non-zero counts, with no names.
1635
1636 4. The long form of the pcregrep -F option was incorrectly implemented as
1637 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1638 but it seems right to fix it, and I didn't think it was worth preserving
1639 the old behaviour.
1640
1641 5. The command line items --regex=pattern and --regexp=pattern were not
1642 recognized by pcregrep, which required --regex pattern or --regexp pattern
1643 (with a space rather than an '='). The man page documented the '=' forms,
1644 which are compatible with GNU grep; these now work.
1645
1646 6. No libpcreposix.pc file was created for pkg-config; there was just
1647 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1648
1649 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1650 when UCP support is not needed, by modifying the Python script that
1651 generates it from Unicode data files. This should not matter if the module
1652 is correctly used as a library, but I received one complaint about 50K of
1653 unwanted data. My guess is that the person linked everything into his
1654 program rather than using a library. Anyway, it does no harm.
1655
1656 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1657 was a minimum greater than 1 for a wide character in a possessive
1658 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1659 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1660 character. Chaos in the form of incorrect output or a compiling loop could
1661 result.
1662
1663 9. The restrictions on what a pattern can contain when partial matching is
1664 requested for pcre_exec() have been removed. All patterns can now be
1665 partially matched by this function. In addition, if there are at least two
1666 slots in the offset vector, the offset of the earliest inspected character
1667 for the match and the offset of the end of the subject are set in them when
1668 PCRE_ERROR_PARTIAL is returned.
1669
1670 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1671 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1672 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1673 and may be more useful for multi-segment matching.
1674
1675 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1676 used to be given if ever the end of the subject was reached; now it is
1677 given only if matching could not proceed because another character was
1678 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1679 string "Z", which now yields "no match" instead of "partial match". In the
1680 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1681 final character ended with (*FAIL).
1682
1683 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1684 if the pattern had a "must contain" character that was already found in the
1685 earlier partial match, unless partial matching was again requested. For
1686 example, with the pattern /dog.(body)?/, the "must contain" character is
1687 "g". If the first part-match was for the string "dog", restarting with
1688 "sbody" failed. This bug has been fixed.
1689
1690 13. The string returned by pcre_dfa_exec() after a partial match has been
1691 changed so that it starts at the first inspected character rather than the
1692 first character of the match. This makes a difference only if the pattern
1693 starts with a lookbehind assertion or \b or \B (\K is not supported by
1694 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1695 matching functions compatible, and I think it's the right thing to do.
1696
1697 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1698 so that the demonstration program is easily available in environments where
1699 PCRE has not been installed from source.
1700
1701 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1702 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1703 library.
1704
1705 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1706 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1707 is not the first non-POSIX option to be added. Clearly some people find
1708 these options useful.
1709
1710 17. If a caller to the POSIX matching function regexec() passes a non-zero
1711 value for nmatch with a NULL value for pmatch, the value of
1712 nmatch is forced to zero.
1713
1714 18. RunGrepTest did not have a test for the availability of the -u option of
1715 the diff command, as RunTest does. It now checks in the same way as
1716 RunTest, and also checks for the -b option.
1717
1718 19. If an odd number of negated classes containing just a single character
1719 interposed, within parentheses, between a forward reference to a named
1720 subpattern and the definition of the subpattern, compilation crashed with
1721 an internal error, complaining that it could not find the referenced
1722 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1723 [The bug was that it was starting one character too far in when skipping
1724 over the character class, thus treating the ] as data rather than
1725 terminating the class. This meant it could skip too much.]
1726
1727 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1728 /g option in pcretest when the pattern contains \K, which makes it possible
1729 to have an empty string match not at the start, even when the pattern is
1730 anchored. Updated pcretest and pcredemo to use this option.
1731
1732 21. If the maximum number of capturing subpatterns in a recursion was greater
1733 than the maximum at the outer level, the higher number was returned, but
1734 with unset values at the outer level. The correct (outer level) value is
1735 now given.
1736
1737 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1738 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1739 make it do so. The string so far is captured, making this feature
1740 compatible with Perl.
1741
1742 23. The tests have been re-organized, adding tests 11 and 12, to make it
1743 possible to check the Perl 5.10 features against Perl 5.10.
1744
1745 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1746 pattern matches a fixed length string. PCRE did not allow this; now it
1747 does. Neither allows recursion.
1748
1749 25. I finally figured out how to implement a request to provide the minimum
1750 length of subject string that was needed in order to match a given pattern.
1751 (It was back references and recursion that I had previously got hung up
1752 on.) This code has now been added to pcre_study(); it finds a lower bound
1753 to the length of subject needed. It is not necessarily the greatest lower
1754 bound, but using it to avoid searching strings that are too short does give
1755 some useful speed-ups. The value is available to calling programs via
1756 pcre_fullinfo().
1757
1758 26. While implementing 25, I discovered to my embarrassment that pcretest had
1759 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1760 study optimizations had never been tested with that matching function.
1761 Oops. What is worse, even when it was passed study data, there was a bug in
1762 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1763 were also very few tests of studied patterns with pcre_dfa_exec().
1764
1765 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1766 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1767 on the other side of the coin, they are no longer allowed to have different
1768 names, because these cannot be distinguished in PCRE, and this has caused
1769 confusion. (This is a difference from Perl.)
1770
1771 28. When duplicate subpattern names are present (necessarily with different
1772 numbers, as required by 27 above), and a test is made by name in a
1773 conditional pattern, either for a subpattern having been matched, or for
1774 recursion in such a pattern, all the associated numbered subpatterns are
1775 tested, and the overall condition is true if the condition is true for any
1776 one of them. This is the way Perl works, and is also more like the way
1777 testing by number works.
1778
1779
1780 Version 7.9 11-Apr-09
1781 ---------------------
1782
1783 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1784 (pcretest), all targets were linked against these libraries. This included
1785 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1786 libraries. This caused unwanted dependencies to be created. This problem
1787 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1788 pcretest is linked with readline.
1789
1790 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1791 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1792 moved outside it again, because FALSE and TRUE are already defined in AIX,
1793 but BOOL is not.
1794
1795 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1796 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1797
1798 4. The pcregrep documentation said spaces were inserted as well as colons (or
1799 hyphens) following file names and line numbers when outputting matching
1800 lines. This is not true; no spaces are inserted. I have also clarified the
1801 wording for the --colour (or --color) option.
1802
1803 5. In pcregrep, when --colour was used with -o, the list of matching strings
1804 was not coloured; this is different to GNU grep, so I have changed it to be
1805 the same.
1806
1807 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1808 each matching line was coloured. Now it goes on to look for further matches
1809 of any of the test patterns, which is the same behaviour as GNU grep.
1810
1811 7. A pattern that could match an empty string could cause pcregrep to loop; it
1812 doesn't make sense to accept an empty string match in pcregrep, so I have
1813 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1814 seems to be how GNU grep behaves. [But see later change 40 for release
1815 8.33.]
1816
1817 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1818 start or after a newline", because the conditional assertion was not being
1819 correctly handled. The rule now is that both the assertion and what follows
1820 in the first alternative must satisfy the test.
1821
1822 9. If auto-callout was enabled in a pattern with a conditional group whose
1823 condition was an assertion, PCRE could crash during matching, both with
1824 pcre_exec() and pcre_dfa_exec().
1825
1826 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1827 used for matching.
1828
1829 11. Unicode property support in character classes was not working for
1830 characters (bytes) greater than 127 when not in UTF-8 mode.
1831
1832 12. Added the -M command line option to pcretest.
1833
1834 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1835
1836 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1837
1838 16. Added comments and documentation about mis-use of no_arg in the C++
1839 wrapper.
1840
1841 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1842 from Martin Jerabek that uses macro names for all relevant character and
1843 string constants.
1844
1845 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1846 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1847 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1848 these, but not everybody uses configure.
1849
1850 19. A conditional group that had only one branch was not being correctly
1851 recognized as an item that could match an empty string. This meant that an
1852 enclosing group might also not be so recognized, causing infinite looping
1853 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1854 with the subject "ab", where knowledge that the repeated group can match
1855 nothing is needed in order to break the loop.
1856
1857 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1858 exec(), but without supplying a callout function, matching went wrong.
1859
1860 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1861 leak if the size of the offset vector was greater than 30. When the vector
1862 is smaller, the saved offsets during recursion go onto a local stack
1863 vector, but for larger vectors malloc() is used. It was failing to free
1864 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1865 error, in fact).
1866
1867 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1868 heapframe that is used only when UTF-8 support is enabled. This caused no
1869 problem, but was untidy.
1870
1871 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1872 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1873 included within another project.
1874
1875 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1876 slightly modified by me:
1877
1878 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1879 not building pcregrep.
1880
1881 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1882 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1883
1884 25. Forward references, both numeric and by name, in patterns that made use of
1885 duplicate group numbers, could behave incorrectly or give incorrect errors,
1886 because when scanning forward to find the reference group, PCRE was not
1887 taking into account the duplicate group numbers. A pattern such as
1888 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1889
1890 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1891 the feature of a custom pointer more persuasive (as requested by a user).
1892
1893 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1894 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1895 pre-defined.
1896
1897 28. Added support for (*UTF8) at the start of a pattern.
1898
1899 29. Arrange for flags added by the "release type" setting in CMake to be shown
1900 in the configuration summary.
1901
1902
1903 Version 7.8 05-Sep-08
1904 ---------------------
1905
1906 1. Replaced UCP searching code with optimized version as implemented for Ad
1907 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1908 stage table and inline lookup instead of a function, giving speed ups of 2
1909 to 5 times on some simple patterns that I tested. Permission was given to
1910 distribute the MultiStage2.py script that generates the tables (it's not in
1911 the tarball, but is in the Subversion repository).
1912
1913 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1914 scripts.
1915
1916 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1917 a group with a zero qualifier. The result of the study could be incorrect,
1918 or the function might crash, depending on the pattern.
1919
1920 4. Caseless matching was not working for non-ASCII characters in back
1921 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1922 It now works when Unicode Property Support is available.
1923
1924 5. In pcretest, an escape such as \x{de} in the data was always generating
1925 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1926 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1927 truncation.
1928
1929 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1930
1931 7. Added two (int) casts to pcregrep when printing the difference of two
1932 pointers, in case they are 64-bit values.
1933
1934 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1935 test 2 if it fails.
1936
1937 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1938 and a #define of that name to empty if it is not externally set. This is to
1939 allow users of MSVC to set it if necessary.
1940
1941 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1942 the convenience functions in the pcre_get.c source file.
1943
1944 11. An option change at the start of a pattern that had top-level alternatives
1945 could cause overwriting and/or a crash. This command provoked a crash in
1946 some environments:
1947
1948 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1949
1950 This potential security problem was recorded as CVE-2008-2371.
1951
1952 12. For a pattern where the match had to start at the beginning or immediately
1953 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1954 pcre_dfa_exec() could read past the end of the passed subject if there was
1955 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1956 pcretest so that it places the subject at the end of its malloc-ed buffer.
1957
1958 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1959 exec() might read past the end of the data buffer in UTF-8 mode.
1960
1961 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1962 the data contained the byte 0x85 as part of a UTF-8 character within its
1963 first line. This applied both to normal and DFA matching.
1964
1965 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1966 /^[^d]*?$/8 failed to match "abc".
1967
1968 16. Added a missing copyright notice to pcrecpp_internal.h.
1969
1970 17. Make it more clear in the documentation that values returned from
1971 pcre_exec() in ovector are byte offsets, not character counts.
1972
1973 18. Tidied a few places to stop certain compilers from issuing warnings.
1974
1975 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1976 supplied by Stefan Weber. I made a further small update for 7.8 because
1977 there is a change of source arrangements: the pcre_searchfuncs.c module is
1978 replaced by pcre_ucd.c.
1979
1980
1981 Version 7.7 07-May-08
1982 ---------------------
1983
1984 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1985 a string to a long long, pretend we don't even have a long long." This is
1986 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1987
1988 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1989 pre-7.6 versions, which defined a global no_arg variable instead of putting
1990 it in the RE class. (See also #8 below.)
1991
1992 3. Remove a line of dead code, identified by coverity and reported by Nuno
1993 Lopes.
1994
1995 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1996
1997 (1) The include/exclude patterns were being applied to the whole pathnames
1998 of files, instead of just to the final components.
1999
2000 (2) If there was more than one level of directory, the subdirectories were
2001 skipped unless they satisfied the include/exclude conditions. This is
2002 inconsistent with GNU grep (and could even be seen as contrary to the
2003 pcregrep specification - which I improved to make it absolutely clear).
2004 The action now is always to scan all levels of directory, and just
2005 apply the include/exclude patterns to regular files.
2006
2007 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2008 --exclude_dir in the tests to avoid scanning .svn directories.
2009
2010 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2011 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2012 doesn't support NULs in patterns.
2013
2014 7. Added some missing "const"s to declarations of static tables in
2015 pcre_compile.c and pcre_dfa_exec.c.
2016
2017 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2018 caused by fix #2 above. (Subsequently also a second patch to fix the
2019 first patch. And a third patch - this was a messy problem.)
2020
2021 9. Applied Craig's patch to remove the use of push_back().
2022
2023 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2024 matching function regexec().
2025
2026 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2027 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2028 references. PCRE supports relative numbers with this syntax (I don't think
2029 Oniguruma does).
2030
2031 12. Previously, a group with a zero repeat such as (...){0} was completely
2032 omitted from the compiled regex. However, this means that if the group
2033 was called as a subroutine from elsewhere in the pattern, things went wrong
2034 (an internal error was given). Such groups are now left in the compiled
2035 pattern, with a new opcode that causes them to be skipped at execution
2036 time.
2037
2038 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2039 to the way PCRE behaves:
2040
2041 (a) A lone ] character is dis-allowed (Perl treats it as data).
2042
2043 (b) A back reference to an unmatched subpattern matches an empty string
2044 (Perl fails the current match path).
2045
2046 (c) A data ] in a character class must be notated as \] because if the
2047 first data character in a class is ], it defines an empty class. (In
2048 Perl it is not possible to have an empty class.) The empty class []
2049 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2050 The negative empty class [^] matches any one character, independently
2051 of the DOTALL setting.
2052
2053 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2054 non-existent subpattern following a character class starting with ']' and
2055 containing () gave an internal compiling error instead of "reference to
2056 non-existent subpattern". Fortunately, when the pattern did exist, the
2057 compiled code was correct. (When scanning forwards to check for the
2058 existence of the subpattern, it was treating the data ']' as terminating
2059 the class, so got the count wrong. When actually compiling, the reference
2060 was subsequently set up correctly.)
2061
2062 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2063 it was being rejected as not supported by pcre_dfa_exec(), even though
2064 other assertions are supported. I have made pcre_dfa_exec() support
2065 (*FAIL).
2066
2067 16. The implementation of 13c above involved the invention of a new opcode,
2068 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2069 cannot be changed at match time, I realized I could make a small
2070 improvement to matching performance by compiling OP_ALLANY instead of
2071 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2072 on the OP_ANY path.
2073
2074 17. Compiling pcretest on Windows with readline support failed without the
2075 following two fixes: (1) Make the unistd.h include conditional on
2076 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2077
2078 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2079 ncurses library to be included for pcretest when ReadLine support is
2080 requested, but also to allow for it to be overridden. This patch came from
2081 Daniel Bergström.
2082
2083 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2084 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2085 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2086 spotting this.
2087
2088
2089 Version 7.6 28-Jan-08
2090 ---------------------
2091
2092 1. A character class containing a very large number of characters with
2093 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2094 overflow.
2095
2096 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2097 HAVE_LONG_LONG is not defined.
2098
2099 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2100 bring them up to date and include new features. This patch includes:
2101
2102 - Fixed PH's badly added libz and libbz2 support.
2103 - Fixed a problem with static linking.
2104 - Added pcredemo. [But later removed - see 7 below.]
2105 - Fixed dftables problem and added an option.
2106 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2107 HAVE_LONG_LONG.
2108 - Added readline support for pcretest.
2109 - Added an listing of the option settings after cmake has run.
2110
2111 4. A user submitted a patch to Makefile that makes it easy to create
2112 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2113 Makefile.am that cause it to include this special target, without
2114 affecting anything else. Note that the same mingw target plus all
2115 the other distribution libraries and programs are now supported
2116 when configuring with CMake (see 6 below) instead of with
2117 Configure/Make.
2118
2119 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2120 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2121 exported in the Windows port". It has not yet been confirmed that the patch
2122 solves the problem, but it does no harm.
2123
2124 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2125 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2126 with CMake, and also correct the comment about stack recursion.
2127
2128 7. Remove the automatic building of pcredemo from the ./configure system and
2129 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2130 of a program that users should build themselves after PCRE is installed, so
2131 building it automatically is not really right. What is more, it gave
2132 trouble in some build environments.
2133
2134 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2135
2136
2137 Version 7.5 10-Jan-08
2138 ---------------------
2139
2140 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2141 values in parens when parsing an RE using the C++ wrapper."
2142
2143 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2144 Characters greater than 255 were excluded from the class instead of being
2145 included.
2146
2147 3. The same bug as (2) above applied to negated POSIX classes such as
2148 [:^space:].
2149
2150 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2151 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2152 I have changed it.
2153
2154 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2155 first named subpattern) and a construct such as (?&a) would reference the
2156 first named subpattern whose name started with "a" (in other words, the
2157 length check was missing). Both these problems are fixed. "Subpattern name
2158 expected" is now given for (?&) (a zero-length name), and this patch also
2159 makes it give the same error for \k'' (previously it complained that that
2160 was a reference to a non-existent subpattern).
2161
2162 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2163 this is right because (?- can be followed by option settings as well as by
2164 digits. I have, however, made the messages clearer.
2165
2166 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2167 than the number used in the conditional) now cause a compile-time error.
2168 This is actually not compatible with Perl, which accepts such patterns, but
2169 treats the conditional as always being FALSE (as PCRE used to), but it
2170 seems to me that giving a diagnostic is better.
2171
2172 8. Change "alphameric" to the more common word "alphanumeric" in comments
2173 and messages.
2174
2175 9. Fix two occurrences of "backslash" in comments that should have been
2176 "backspace".
2177
2178 10. Remove two redundant lines of code that can never be obeyed (their function
2179 was moved elsewhere).
2180
2181 11. The program that makes PCRE's Unicode character property table had a bug
2182 which caused it to generate incorrect table entries for sequences of
2183 characters that have the same character type, but are in different scripts.
2184 It amalgamated them into a single range, with the script of the first of
2185 them. In other words, some characters were in the wrong script. There were
2186 thirteen such cases, affecting characters in the following ranges:
2187
2188 U+002b0 - U+002c1
2189 U+0060c - U+0060d
2190 U+0061e - U+00612
2191 U+0064b - U+0065e
2192 U+0074d - U+0076d
2193 U+01800 - U+01805
2194 U+01d00 - U+01d77
2195 U+01d9b - U+01dbf
2196 U+0200b - U+0200f
2197 U+030fc - U+030fe
2198 U+03260 - U+0327f
2199 U+0fb46 - U+0fbb1
2200 U+10450 - U+1049d
2201
2202 12. The -o option (show only the matching part of a line) for pcregrep was not
2203 compatible with GNU grep in that, if there was more than one match in a
2204 line, it showed only the first of them. It now behaves in the same way as
2205 GNU grep.
2206
2207 13. If the -o and -v options were combined for pcregrep, it printed a blank
2208 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2209 does the same. The return code can be used to tell if there were any
2210 non-matching lines.
2211
2212 14. Added --file-offsets and --line-offsets to pcregrep.
2213
2214 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2215 infinitely looping recursion. The bug was that positive lookaheads were not
2216 being skipped when checking for a possible empty match (negative lookaheads
2217 and both kinds of lookbehind were skipped).
2218
2219 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2220 inclusion of <windows.h> to before rather than after the definition of
2221 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2222
2223 17. Specifying a possessive quantifier with a specific limit for a Unicode
2224 character property caused pcre_compile() to compile bad code, which led at
2225 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2226 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2227 caused the error; without that there was no problem.
2228
2229 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2230
2231 19. Added --enable-pcretest-libreadline.
2232
2233 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2234 RE::GlobalReplace(). As a result, the number of replacements returned was
2235 double what it should be. I removed one of the increments, but Craig sent a
2236 later patch that removed the other one (the right fix) and added unit tests
2237 that check the return values (which was not done before).
2238
2239 21. Several CMake things:
2240
2241 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2242 the names libpcre and libpcreposix, not just pcre and pcreposix.
2243
2244 (2) The above change means that pcretest and pcregrep are now correctly
2245 linked with the newly-built libraries, not previously installed ones.
2246
2247 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2248
2249 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2250 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2251 UTF-8 newline character). The key issue is that the pattern starts .*;
2252 this means that the match must be either at the beginning, or after a
2253 newline. The bug was in the code for advancing after a failed match and
2254 checking that the new position followed a newline. It was not taking
2255 account of UTF-8 characters correctly.
2256
2257 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2258 character classes. PCRE was not treating the sequence [:...:] as a
2259 character class unless the ... were all letters. Perl, however, seems to
2260 allow any characters between [: and :], though of course it rejects as
2261 unknown any "names" that contain non-letters, because all the known class
2262 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2263 for example, whereas PCRE did not - it did not recognize a POSIX character
2264 class. This seemed a bit dangerous, so the code has been changed to be
2265 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2266 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2267 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2268 Perl does, and where it didn't before.
2269
2270 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2271 Windows environments %n is disabled by default.
2272
2273
2274 Version 7.4 21-Sep-07
2275 ---------------------
2276
2277 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2278 means that a class such as [\s] counted as "explicit reference to CR or
2279 LF". That isn't really right - the whole point of the change was to try to
2280 help when there was an actual mention of one of the two characters. So now
2281 the change happens only if \r or \n (or a literal CR or LF) character is
2282 encountered.
2283
2284 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2285 of both had grown to the point where there were only 3 bits left.
2286 Fortunately, there was spare space in the data structure, and so I have
2287 moved the internal flags into a new 16-bit field to free up more option
2288 bits.
2289
2290 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2291 but did not set the internal JCHANGED flag - either of these is enough to
2292 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2293 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2294 start sets both bits.
2295
2296 4. Added options (at build time, compile time, exec time) to change \R from
2297 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2298
2299 5. doc/pcresyntax.html was missing from the distribution.
2300
2301 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2302 compatibility, even though it is no longer used.
2303
2304 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2305 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2306 windows.h file is present (where different names are used). [This was
2307 reversed later after testing - see 16 below.]
2308
2309 8. Changed all #include <config.h> to #include "config.h". There were also
2310 some further <pcre.h> cases that I changed to "pcre.h".
2311
2312 9. When pcregrep was used with the --colour option, it missed the line ending
2313 sequence off the lines that it output.
2314
2315 10. It was pointed out to me that arrays of string pointers cause lots of
2316 relocations when a shared library is dynamically loaded. A technique of
2317 using a single long string with a table of offsets can drastically reduce
2318 these. I have refactored PCRE in four places to do this. The result is
2319 dramatic:
2320
2321 Originally: 290
2322 After changing UCP table: 187
2323 After changing error message table: 43
2324 After changing table of "verbs" 36
2325 After changing table of Posix names 22
2326
2327 Thanks to the folks working on Gregex for glib for this insight.
2328
2329 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2330 unicode-properties was also set.
2331
2332 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2333
2334 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2335 checked only for CRLF.
2336
2337 14. Added casts to pcretest.c to avoid compiler warnings.
2338
2339 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2340
2341 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2342 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2343 entirely. This removes changes made in 7 above.
2344
2345 17. The CMake files have been updated, and there is now more information about
2346 building with CMake in the NON-UNIX-USE document.
2347
2348
2349 Version 7.3 28-Aug-07
2350 ---------------------
2351
2352 1. In the rejigging of the build system that eventually resulted in 7.1, the
2353 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2354 brackets there is not right, since it causes compilers to look for an
2355 installed pcre.h, not the version that is in the source that is being
2356 compiled (which of course may be different). I have changed it back to:
2357
2358 #include "pcre.h"
2359
2360 I have a vague recollection that the change was concerned with compiling in
2361 different directories, but in the new build system, that is taken care of
2362 by the VPATH setting the Makefile.
2363
2364 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2365 when the subject happened to end in the byte 0x85 (e.g. if the last
2366 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2367 characters but of course it shouldn't be taken as a newline when it is part
2368 of another character. The bug was that, for an unlimited repeat of . in
2369 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2370 characters when looking for a newline.
2371
2372 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2373
2374 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2375 in debug output.
2376
2377 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2378 long printing in the pcrecpp unittest when running under MinGW.
2379
2380 6. ESC_K was left out of the EBCDIC table.
2381
2382 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2383 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2384 limit also applies to "virtual nesting" when a pattern is recursive, and in
2385 this case 1000 isn't so big. I have been able to remove this limit at the
2386 expense of backing off one optimization in certain circumstances. Normally,
2387 when pcre_exec() would call its internal match() function recursively and
2388 immediately return the result unconditionally, it uses a "tail recursion"
2389 feature to save stack. However, when a subpattern that can match an empty
2390 string has an unlimited repetition quantifier, it no longer makes this
2391 optimization. That gives it a stack frame in which to save the data for
2392 checking that an empty string has been matched. Previously this was taken
2393 from the 1000-entry workspace that had been reserved. So now there is no
2394 explicit limit, but more stack is used.
2395
2396 8. Applied Daniel's patches to solve problems with the import/export magic
2397 syntax that is required for Windows, and which was going wrong for the
2398 pcreposix and pcrecpp parts of the library. These were overlooked when this
2399 problem was solved for the main library.
2400
2401 9. There were some crude static tests to avoid integer overflow when computing
2402 the size of patterns that contain repeated groups with explicit upper
2403 limits. As the maximum quantifier is 65535, the maximum group length was
2404 set at 30,000 so that the product of these two numbers did not overflow a
2405 32-bit integer. However, it turns out that people want to use groups that
2406 are longer than 30,000 bytes (though not repeat them that many times).
2407 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2408 made it possible to implement the integer overflow checks in a much more
2409 dynamic way, which I have now done. The artificial limitation on group
2410 length has been removed - we now have only the limit on the total length of
2411 the compiled pattern, which depends on the LINK_SIZE setting.
2412
2413 10. Fixed a bug in the documentation for get/copy named substring when
2414 duplicate names are permitted. If none of the named substrings are set, the
2415 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2416 empty string.
2417
2418 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2419 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2420 because the ] is interpreted as the first data character and the
2421 terminating ] is not found. PCRE has been made compatible with Perl in this
2422 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2423 cause memory overwriting.
2424
2425 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2426 string has been matched (to stop an infinite loop). It was not recognizing
2427 a conditional subpattern that could match an empty string if that
2428 subpattern was within another subpattern. For example, it looped when
2429 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2430 condition was not nested. This bug has been fixed.
2431
2432 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2433 past the start of the subject in the presence of bytes with the top bit
2434 set, for example "\x8aBCD".
2435
2436 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2437 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2438
2439 14. Optimized (?!) to (*FAIL).
2440
2441 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2442 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2443 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2444 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2445 does: it's just the validity check that is more restrictive.
2446
2447 16. Inserted checks for integer overflows during escape sequence (backslash)
2448 processing, and also fixed erroneous offset values for syntax errors during
2449 backslash processing.
2450
2451 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2452 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2453
2454 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2455 caused an overrun.
2456
2457 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2458 something other than just ASCII characters) inside a group that had an
2459 unlimited repeat caused a loop at compile time (while checking to see
2460 whether the group could match an empty string).
2461
2462 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2463 [\P{Any}] did so. (Error in the code for printing property names.)
2464
2465 21. An orphan \E inside a character class could cause a crash.
2466
2467 22. A repeated capturing bracket such as (A)? could cause a wild memory
2468 reference during compilation.
2469
2470 23. There are several functions in pcre_compile() that scan along a compiled
2471 expression for various reasons (e.g. to see if it's fixed length for look
2472 behind). There were bugs in these functions when a repeated \p or \P was
2473 present in the pattern. These operators have additional parameters compared
2474 with \d, etc, and these were not being taken into account when moving along
2475 the compiled data. Specifically:
2476
2477 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2478 length.
2479
2480 (b) An item such as \pL+ within a repeated group could cause crashes or
2481 loops.
2482
2483 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2484 "reference to non-existent subpattern" error.
2485
2486 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2487
2488 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2489 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2490
2491 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2492
2493 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2494 character were causing crashes (broken optimization).
2495
2496 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2497 \p or \P) caused a compile-time loop.
2498
2499 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2500 break. For example, the unstudied pattern [\r\n]A does not match the string
2501 "\r\nA" because change 7.0/46 below moves the current point on by two
2502 characters after failing to match at the start. However, the pattern \nA
2503 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2504 the same is true. There doesn't seem any very clean way out of this, but
2505 what I have chosen to do makes the common cases work: PCRE now takes note
2506 of whether there can be an explicit match for \r or \n anywhere in the
2507 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2508 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2509 pattern has explicit CR or LF references.
2510
2511 29. Added (*CR) etc for changing newline setting at start of pattern.
2512
2513
2514 Version 7.2 19-Jun-07
2515 ---------------------
2516
2517 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2518 which is apparently normally available under Windows.
2519
2520 2. Re-jig the pcregrep tests with different newline settings in an attempt
2521 to make them independent of the local environment's newline setting.
2522
2523 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2524
2525 4. Some of the "internals" tests were previously cut out when the link size
2526 was not 2, because the output contained actual offsets. The recent new
2527 "Z" feature of pcretest means that these can be cut out, making the tests
2528 usable with all link sizes.
2529
2530 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2531 stack recursion. This gives a massive performance boost under BSD, but just
2532 a small improvement under Linux. However, it saves one field in the frame
2533 in all cases.
2534
2535 6. Added more features from the forthcoming Perl 5.10:
2536
2537 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2538 recursion call. It refers to the nth most recently opened parentheses.
2539
2540 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2541 to be opened parentheses.
2542
2543 (c) Conditions that refer to capturing parentheses can be specified
2544 relatively, for example, (?(-2)... or (?(+3)...
2545
2546 (d) \K resets the start of the current match so that everything before
2547 is not part of it.
2548
2549 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2550
2551 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2552 reference syntax.
2553
2554 (g) (?| introduces a group in which the numbering of parentheses in each
2555 alternative starts with the same number.
2556
2557 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2558
2559 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2560 PCRE_INFO_JCHANGED.
2561
2562 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2563 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2564 for detecting groups that can match an empty string.
2565
2566 9. A pattern with a very large number of alternatives (more than several
2567 hundred) was running out of internal workspace during the pre-compile
2568 phase, where pcre_compile() figures out how much memory will be needed. A
2569 bit of new cunning has reduced the workspace needed for groups with
2570 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2571 workspace instead of running out of the 4096 that are available.
2572
2573 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2574
2575 11. Applied patch from Google to remove an optimization that didn't quite work.
2576 The report of the bug said:
2577
2578 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2579 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2580 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2581
2582 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2583 it matched the wrong number of bytes.
2584
2585
2586 Version 7.1 24-Apr-07
2587 ---------------------
2588
2589 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2590 that is more "standard", making use of automake and other Autotools. There
2591 is some re-arrangement of the files and adjustment of comments consequent
2592 on this.
2593
2594 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2595 for recursive directory scanning broke on some systems because the files
2596 are not scanned in any specific order and on different systems the order
2597 was different. A call to "sort" has been inserted into RunGrepTest for the
2598 approprate test as a short-term fix. In the longer term there may be an
2599 alternative.
2600
2601 3. I had an email from Eric Raymond about problems translating some of PCRE's
2602 man pages to HTML (despite the fact that I distribute HTML pages, some
2603 people do their own conversions for various reasons). The problems
2604 concerned the use of low-level troff macros .br and .in. I have therefore
2605 removed all such uses from the man pages (some were redundant, some could
2606 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2607 HTML has been updated to handle .nf/.fi and to complain if it encounters
2608 .br or .in.
2609
2610 4. Updated comments in configure.ac that get placed in config.h.in and also
2611 arranged for config.h to be included in the distribution, with the name
2612 config.h.generic, for the benefit of those who have to compile without
2613 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2614
2615 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2616 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2617 makevp.bat for the current PCRE, using the additional files
2618 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2619
2620 6. A Windows user reported a minor discrepancy with test 2, which turned out
2621 to be caused by a trailing space on an input line that had got lost in his
2622 copy. The trailing space was an accident, so I've just removed it.
2623
2624 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2625 that is needed.
2626
2627 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2628 as "const" (a) because they are and (b) because it helps the PHP
2629 maintainers who have recently made a script to detect big data structures
2630 in the php code that should be moved to the .rodata section. I remembered
2631 to update Builducptable as well, so it won't revert if ucptable.h is ever
2632 re-created.
2633
2634 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2635 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2636 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2637 support is not required. This saves 1.5-2K of code, which is important in
2638 some applications.
2639
2640 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2641 so as not to refer to the tables, even though these functions will never be
2642 called when UTF-8 support is disabled. Otherwise there are problems with a
2643 shared library.
2644
2645 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2646
2647 (a) It was defining its arguments as char * instead of void *.
2648
2649 (b) It was assuming that all moves were upwards in memory; this was true
2650 a long time ago when I wrote it, but is no longer the case.
2651
2652 The emulated memove() is provided for those environments that have neither
2653 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2654 is clearly not the case, as these two bugs were recently reported.
2655
2656 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2657 and Detrail to create the HTML documentation, the .txt form of the man
2658 pages, and it removes trailing spaces from listed files. It also creates
2659 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2660 case, it wraps all the #defines with #ifndefs. This script should be run
2661 before "make dist".
2662
2663 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2664 with Unicode property support.
2665
2666 (a) For a maximizing quantifier, if the two different cases of the
2667 character were of different lengths in their UTF-8 codings (there are
2668 some cases like this - I found 11), and the matching function had to
2669 back up over a mixture of the two cases, it incorrectly assumed they
2670 were both the same length.
2671
2672 (b) When PCRE was configured to use the heap rather than the stack for
2673 recursion during matching, it was not correctly preserving the data for
2674 the other case of a UTF-8 character when checking ahead for a match
2675 while processing a minimizing repeat. If the check also involved
2676 matching a wide character, but failed, corruption could cause an
2677 erroneous result when trying to check for a repeat of the original
2678 character.
2679
2680 13. Some tidying changes to the testing mechanism:
2681
2682 (a) The RunTest script now detects the internal link size and whether there
2683 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2684 values substituted by "configure". (The RunGrepTest script already did
2685 this for UTF-8.) The configure.ac script no longer substitutes the
2686 relevant variables.
2687
2688 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2689 with length and offset values. This means that the output is different
2690 for different internal link sizes. Test 2 is skipped for link sizes
2691 other than 2 because of this, bypassing the problem. Unfortunately,
2692 there was also a test in test 3 (the locale tests) that used /B and
2693 failed for link sizes other than 2. Rather than cut the whole test out,
2694 I have added a new /Z option to pcretest that replaces the length and
2695 offset values with spaces. This is now used to make test 3 independent
2696 of link size. (Test 2 will be tidied up later.)
2697
2698 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2699 segmentation fault instead of returning the appropriate error message.
2700
2701 15. In multiline mode when the newline sequence was set to "any", the pattern
2702 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2703 This doesn't seem right; it now treats the CRLF combination as the line
2704 ending, and so does not match in that case. It's only a pattern such as ^$
2705 that would hit this one: something like ^ABC$ would have failed after \r
2706 and then tried again after \r\n.
2707
2708 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2709 in an attempt to make files that differ only in their line terminators
2710 compare equal. This works on Linux.
2711
2712 17. Under certain error circumstances pcregrep might try to free random memory
2713 as it exited. This is now fixed, thanks to valgrind.
2714
2715 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2716 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2717 was because its rules for how to advance for /g after matching an empty
2718 string at the end of a line did not allow for this case. They now check for
2719 it specially.
2720
2721 20. pcretest is supposed to handle patterns and data of any length, by
2722 extending its buffers when necessary. It was getting this wrong when the
2723 buffer for a data line had to be extended.
2724
2725 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2726 CRLF as a newline sequence.
2727
2728 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2729 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2730 I have nevertheless tidied it up.
2731
2732 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2733
2734 24. Added a man page for pcre-config.
2735
2736
2737 Version 7.0 19-Dec-06
2738 ---------------------
2739
2740 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2741 moving to gcc 4.1.1.
2742
2743 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2744 sys/time.h, which is documented as needed for this function. It doesn't
2745 seem to matter on Linux, but it showed up on some releases of OS X.
2746
2747 3. It seems that there are systems where bytes whose values are greater than
2748 127 match isprint() in the "C" locale. The "C" locale should be the
2749 default when a C program starts up. In most systems, only ASCII printing
2750 characters match isprint(). This difference caused the output from pcretest
2751 to vary, making some of the tests fail. I have changed pcretest so that:
2752
2753 (a) When it is outputting text in the compiled version of a pattern, bytes
2754 other than 32-126 are always shown as hex escapes.
2755
2756 (b) When it is outputting text that is a matched part of a subject string,
2757 it does the same, unless a different locale has been set for the match
2758 (using the /L modifier). In this case, it uses isprint() to decide.
2759
2760 4. Fixed a major bug that caused incorrect computation of the amount of memory
2761 required for a compiled pattern when options that changed within the
2762 pattern affected the logic of the preliminary scan that determines the
2763 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2764 that the computed length was too small. The symptoms of this bug were
2765 either the PCRE error "internal error: code overflow" from pcre_compile(),
2766 or a glibc crash with a message such as "pcretest: free(): invalid next
2767 size (fast)". Examples of patterns that provoked this bug (shown in
2768 pcretest format) are:
2769
2770 /(?-x: )/x
2771 /(?x)(?-x: \s*#\s*)/
2772 /((?i)[\x{c0}])/8
2773 /(?i:[\x{c0}])/8
2774
2775 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2776 is now done differently.
2777
2778 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2779 wrapper classes; (b) implement a new function in the C++ scanner that is
2780 more efficient than the old way of doing things because it avoids levels of
2781 recursion in the regex matching; (c) add a paragraph to the documentation
2782 for the FullMatch() function.
2783
2784 6. The escape sequence \n was being treated as whatever was defined as
2785 "newline". Not only was this contrary to the documentation, which states
2786 that \n is character 10 (hex 0A), but it also went horribly wrong when
2787 "newline" was defined as CRLF. This has been fixed.
2788
2789 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2790 was being set to -1 for the "end of line" case (supposedly a value that no
2791 character can have). Though this value is never used (the check for end of
2792 line is "zero bytes in current character"), it caused compiler complaints.
2793 I've changed it to 0xffffffff.
2794
2795 8. In pcre_version.c, the version string was being built by a sequence of
2796 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2797 string (as it is for production releases) called a macro with an empty
2798 argument. The C standard says the result of this is undefined. The gcc
2799 compiler treats it as an empty string (which was what was wanted) but it is
2800 reported that Visual C gives an error. The source has been hacked around to
2801 avoid this problem.
2802
2803 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2804 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2805 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2806 of them did).
2807
2808 10. Originally, pcretest opened its input and output without "b"; then I was
2809 told that "b" was needed in some environments, so it was added for release
2810 5.0 to both the input and output. (It makes no difference on Unix-like
2811 systems.) Later I was told that it is wrong for the input on Windows. I've
2812 now abstracted the modes into two macros, to make it easier to fiddle with
2813 them, and removed "b" from the input mode under Windows.
2814
2815 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2816
2817 12. Added -help and --help to pcretest as an official way of being reminded
2818 of the options.
2819
2820 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2821 and pcrecpp.cc because they annoy compilers at high warning levels.
2822
2823 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2824
2825 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2826 scripts are not C programs :-) and which was not noticed because it works
2827 on Linux.
2828
2829 16. pcretest is supposed to handle any length of pattern and data line (as one
2830 line or as a continued sequence of lines) by extending its input buffer if
2831 necessary. This feature was broken for very long pattern lines, leading to
2832 a string of junk being passed to pcre_compile() if the pattern was longer
2833 than about 50K.
2834
2835 17. I have done a major re-factoring of the way pcre_compile() computes the
2836 amount of memory needed for a compiled pattern. Previously, there was code
2837 that made a preliminary scan of the pattern in order to do this. That was
2838 OK when PCRE was new, but as the facilities have expanded, it has become
2839 harder and harder to keep it in step with the real compile phase, and there
2840 have been a number of bugs (see for example, 4 above). I have now found a
2841 cunning way of running the real compile function in a "fake" mode that
2842 enables it to compute how much memory it would need, while actually only
2843 ever using a few hundred bytes of working memory and without too many
2844 tests of the mode. This should make future maintenance and development
2845 easier. A side effect of this work is that the limit of 200 on the nesting
2846 depth of parentheses has been removed (though this was never a serious
2847 limitation, I suspect). However, there is a downside: pcre_compile() now
2848 runs more slowly than before (30% or more, depending on the pattern). I
2849 hope this isn't a big issue. There is no effect on runtime performance.
2850
2851 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2852 newline (only possible for the last line of a file) and it was a
2853 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2854
2855 19. Added additional timing features to pcretest. (1) The -tm option now times
2856 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2857 separate command line item, by a number that specifies the number of
2858 repeats to use when timing. The default is 50000; this gives better
2859 precision, but takes uncomfortably long for very large patterns.
2860
2861 20. Extended pcre_study() to be more clever in cases where a branch of a
2862 subpattern has no definite first character. For example, (a*|b*)[cd] would
2863 previously give no result from pcre_study(). Now it recognizes that the
2864 first character must be a, b, c, or d.
2865
2866 21. There was an incorrect error "recursive call could loop indefinitely" if
2867 a subpattern (or the entire pattern) that was being tested for matching an
2868 empty string contained only one non-empty item after a nested subpattern.
2869 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2870 incorrectly, because the \d was being skipped in the check.
2871
2872 22. The pcretest program now has a new pattern option /B and a command line
2873 option -b, which is equivalent to adding /B to every pattern. This causes
2874 it to show the compiled bytecode, without the additional information that
2875 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2876 is the same as /B/I).
2877
2878 23. A new optimization is now able automatically to treat some sequences such
2879 as a*b as a*+b. More specifically, if something simple (such as a character
2880 or a simple class like \d) has an unlimited quantifier, and is followed by
2881 something that cannot possibly match the quantified thing, the quantifier
2882 is automatically "possessified".
2883
2884 24. A recursive reference to a subpattern whose number was greater than 39
2885 went wrong under certain circumstances in UTF-8 mode. This bug could also
2886 have affected the operation of pcre_study().
2887
2888 25. Realized that a little bit of performance could be had by replacing
2889 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2890
2891 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2892
2893 27. Possessive quantifiers such as a++ were previously implemented by turning
2894 them into atomic groups such as ($>a+). Now they have their own opcodes,
2895 which improves performance. This includes the automatically created ones
2896 from 23 above.
2897
2898 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2899 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2900 the first matched character to be a colon. This applied both to named and
2901 numbered groups.
2902
2903 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2904
2905 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2906 building PCRE on Windows easier, so I have included it in the distribution.
2907
2908 31. There is now a check in pcretest against a ridiculously large number being
2909 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2910 loop, the loop is abandoned.
2911
2912 32. Forward references to subpatterns in conditions such as (?(2)...) where
2913 subpattern 2 is defined later cause pcre_compile() to search forwards in
2914 the pattern for the relevant set of parentheses. This search went wrong
2915 when there were unescaped parentheses in a character class, parentheses
2916 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2917
2918 33. "Subroutine" calls and backreferences were previously restricted to
2919 referencing subpatterns earlier in the regex. This restriction has now
2920 been removed.
2921
2922 34. Added a number of extra features that are going to be in Perl 5.10. On the
2923 whole, these are just syntactic alternatives for features that PCRE had
2924 previously implemented using the Python syntax or my own invention. The
2925 other formats are all retained for compatibility.
2926
2927 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2928 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2929 also .NET compatible.
2930
2931 (b) A recursion or subroutine call to a named group can now be defined as
2932 (?&name) as well as (?P>name).
2933
2934 (c) A backreference to a named group can now be defined as \k<name> or
2935 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2936 5.10, are also .NET compatible.
2937
2938 (d) A conditional reference to a named group can now use the syntax
2939 (?(<name>) or (?('name') as well as (?(name).
2940
2941 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2942 groups (named and numbered) that are never evaluated inline, but can be
2943 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2944 is always false. There may be only one alternative in such a group.
2945
2946 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2947 as the simple (?(R). The condition is true only if the most recent
2948 recursion is that of the given number or name. It does not search out
2949 through the entire recursion stack.
2950
2951 (g) The escape \gN or \g{N} has been added, where N is a positive or
2952 negative number, specifying an absolute or relative reference.
2953
2954 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2955 some "unreachable code" warnings.
2956
2957 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2958 things, this adds five new scripts.
2959
2960 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2961 There were also incompatibilities regarding the handling of \Q..\E inside
2962 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2963 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2964
2965 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2966 matches an empty string, and forcibly breaks the loop. There were bugs in
2967 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2968 against aaaa the result was just "a" rather than "aaaa", for example. Two
2969 separate and independent bugs (that affected different cases) have been
2970 fixed.
2971
2972 39. Refactored the code to abolish the use of different opcodes for small
2973 capturing bracket numbers. This is a tidy that I avoided doing when I
2974 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2975 The new approach is not only tidier, it makes it possible to reduce the
2976 memory needed to fix the previous bug (38).
2977
2978 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2979 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2980 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2981 mode.
2982
2983 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2984 report.
2985
2986 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2987 copy construction and assignment in the C++ wrapper.
2988
2989 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2990 couple of bugs that could have given wrong results in the "--newline=crlf"
2991 case.
2992
2993 44. Added a number of casts and did some reorganization of signed/unsigned int
2994 variables following suggestions from Dair Grant. Also renamed the variable
2995 "this" as "item" because it is a C++ keyword.
2996
2997 45. Arranged for dftables to add
2998
2999 #include "pcre_internal.h"
3000
3001 to pcre_chartables.c because without it, gcc 4.x may remove the array
3002 definition from the final binary if PCRE is built into a static library and
3003 dead code stripping is activated.
3004
3005 46. For an unanchored pattern, if a match attempt fails at the start of a
3006 newline sequence, and the newline setting is CRLF or ANY, and the next two
3007 characters are CRLF, advance by two characters instead of one.
3008
3009
3010 Version 6.7 04-Jul-06
3011 ---------------------
3012
3013 1. In order to handle tests when input lines are enormously long, pcretest has
3014 been re-factored so that it automatically extends its buffers when
3015 necessary. The code is crude, but this _is_ just a test program. The
3016 default size has been increased from 32K to 50K.
3017
3018 2. The code in pcre_study() was using the value of the re argument before
3019 testing it for NULL. (Of course, in any sensible call of the function, it
3020 won't be NULL.)
3021
3022 3. The memmove() emulation function in pcre_internal.h, which is used on
3023 systems that lack both memmove() and bcopy() - that is, hardly ever -
3024 was missing a "static" storage class specifier.
3025
3026 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3027 containing an extended class (one that cannot be represented by a bitmap
3028 because it contains high-valued characters or Unicode property items, e.g.
3029 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3030 pattern, but PCRE should not loop if you do not (it no longer does).
3031 [Detail: two cases were found: (a) a repeated subpattern containing an
3032 extended class; (b) a recursive reference to a subpattern that followed a
3033 previous extended class. It wasn't skipping over the extended class
3034 correctly when UTF-8 mode was not set.]
3035
3036 5. A negated single-character class was not being recognized as fixed-length
3037 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3038 compile error "lookbehind assertion is not fixed length".
3039
3040 6. The RunPerlTest auxiliary script was showing an unexpected difference
3041 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3042 write a Perl script that can interpret lines of an input file either as
3043 byte characters or as UTF-8, which is what "perltest" was being required to
3044 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3045 can't do is switch easily at run time between having the "use utf8;" pragma
3046 or not. In the end, I fudged it by using the RunPerlTest script to insert
3047 "use utf8;" explicitly for the UTF-8 tests.
3048
3049 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3050 the end of the subject string, contrary to the documentation and to what
3051 Perl does. This was true of both matching functions. Now it matches only at
3052 the start of the subject and immediately after *internal* newlines.
3053
3054 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3055 a pointer to an int instead of a pointer to an unsigned long int. This
3056 caused problems on 64-bit systems.
3057
3058 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3059 instance of the 'standard' template library not being so standard".
3060
3061 10. There was no check on the number of named subpatterns nor the maximum
3062 length of a subpattern name. The product of these values is used to compute
3063 the size of the memory block for a compiled pattern. By supplying a very
3064 long subpattern name and a large number of named subpatterns, the size
3065 computation could be caused to overflow. This is now prevented by limiting
3066 the length of names to 32 characters, and the number of named subpatterns
3067 to 10,000.
3068
3069 11. Subpatterns that are repeated with specific counts have to be replicated in
3070 the compiled pattern. The size of memory for this was computed from the
3071 length of the subpattern and the repeat count. The latter is limited to
3072 65535, but there was no limit on the former, meaning that integer overflow
3073 could in principle occur. The compiled length of a repeated subpattern is
3074 now limited to 30,000 bytes in order to prevent this.
3075
3076 12. Added the optional facility to have named substrings with the same name.
3077
3078 13. Added the ability to use a named substring as a condition, using the
3079 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3080 are numbers (not recommended). Forward references are permitted.
3081
3082 14. Added forward references in named backreferences (if you see what I mean).
3083
3084 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3085 pattern could run off the end of the subject. For example, the pattern
3086 "(?s)(.{1,5})"8 did this with the subject "ab".
3087
3088 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3089 PCRE_CASELESS was set when matching characters that were quantified with ?
3090 or *.
3091
3092 17. A character class other than a single negated character that had a minimum
3093 but no maximum quantifier - for example [ab]{6,} - was not handled
3094 correctly by pce_dfa_exec(). It would match only one character.
3095
3096 18. A valid (though odd) pattern that looked like a POSIX character
3097 class but used an invalid character after [ (for example [[,abc,]]) caused
3098 pcre_compile() to give the error "Failed: internal error: code overflow" or
3099 in some cases to crash with a glibc free() error. This could even happen if
3100 the pattern terminated after [[ but there just happened to be a sequence of
3101 letters, a binary zero, and a closing ] in the memory that followed.
3102
3103 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3104 over the years. Originally (before any Unicode support), just the bottom 8
3105 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3106 output from "man perlunicode" includes this:
3107
3108 The regular expression compiler produces polymorphic opcodes. That
3109 is, the pattern adapts to the data and automatically switches to
3110 the Unicode character scheme when presented with Unicode data--or
3111 instead uses a traditional byte scheme when presented with byte
3112 data.
3113
3114 Sadly, a wide octal escape does not cause a switch, and in a string with
3115 no other multibyte characters, these octal escapes are treated as before.
3116 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3117 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3118 Unicode string.
3119
3120 I have not perpetrated such confusion in PCRE. Up till now, it took just
3121 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3122 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3123 translate to the appropriate multibyte character.
3124
3125 29. Applied some refactoring to reduce the number of warnings from Microsoft
3126 and Borland compilers. This has included removing the fudge introduced
3127 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3128 a warning about an unused variable.
3129
3130 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3131 characters since release 4.0, because Perl (from release 5.004) does not.
3132 [Or at least, is documented not to: some releases seem to be in conflict
3133 with the documentation.] However, when a pattern was studied with
3134 pcre_study() and all its branches started with \s, PCRE still included VT
3135 as a possible starting character. Of course, this did no harm; it just
3136 caused an unnecessary match attempt.
3137
3138 22. Removed a now-redundant internal flag bit that recorded the fact that case
3139 dependency changed within the pattern. This was once needed for "required
3140 byte" processing, but is no longer used. This recovers a now-scarce options
3141 bit. Also moved the least significant internal flag bit to the most-
3142 significant bit of the word, which was not previously used (hangover from
3143 the days when it was an int rather than a uint) to free up another bit for
3144 the future.
3145
3146 23. Added support for CRLF line endings as well as CR and LF. As well as the
3147 default being selectable at build time, it can now be changed at runtime
3148 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3149 specify that it is scanning data with non-default line endings.
3150
3151 24. Changed the definition of CXXLINK to make it agree with the definition of
3152 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3153
3154 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3155 recursions. This makes a big different to stack usage for some patterns.
3156
3157 26. If a subpattern containing a named recursion or subroutine reference such
3158 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3159 the space required for the compiled pattern went wrong and gave too small a
3160 value. Depending on the environment, this could lead to "Failed: internal
3161 error: code overflow at offset 49" or "glibc detected double free or
3162 corruption" errors.
3163
3164 27. Applied patches from Google (a) to support the new newline modes and (b) to
3165 advance over multibyte UTF-8 characters in GlobalReplace.
3166
3167 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3168 difference for some implementation of PCRE in some Windows version.
3169
3170 29. Added some extra testing facilities to pcretest:
3171
3172 \q<number> in a data line sets the "match limit" value
3173 \Q<number> in a data line sets the "match recursion limt" value
3174 -S <number> sets the stack size, where <number> is in megabytes
3175
3176 The -S option isn't available for Windows.
3177
3178
3179 Version 6.6 06-Feb-06
3180 ---------------------
3181
3182 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3183 in pcreposix.h. I have copied the definition from pcre.h.
3184
3185 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3186 because pcre.h is no longer a built file.
3187
3188 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3189 not normally included in the compiled code.
3190
3191
3192 Version 6.5 01-Feb-06
3193 ---------------------
3194
3195 1. When using the partial match feature with pcre_dfa_exec(), it was not
3196 anchoring the second and subsequent partial matches at the new starting
3197 point. This could lead to incorrect results. For example, with the pattern
3198 /1234/, partially matching against "123" and then "a4" gave a match.
3199
3200 2. Changes to pcregrep:
3201
3202 (a) All non-match returns from pcre_exec() were being treated as failures
3203 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3204 error message is output. Some extra information is given for the
3205 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3206 probably the only errors that are likely to be caused by users (by
3207 specifying a regex that has nested indefinite repeats, for instance).
3208 If there are more than 20 of these errors, pcregrep is abandoned.
3209
3210 (b) A binary zero was treated as data while matching, but terminated the
3211 output line if it was written out. This has been fixed: binary zeroes
3212 are now no different to any other data bytes.
3213
3214 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3215 used to set a locale for matching. The --locale=xxxx long option has
3216 been added (no short equivalent) to specify a locale explicitly on the
3217 pcregrep command, overriding the environment variables.
3218
3219 (d) When -B was used with -n, some line numbers in the output were one less
3220 than they should have been.
3221
3222 (e) Added the -o (--only-matching) option.
3223
3224 (f) If -A or -C was used with -c (count only), some lines of context were
3225 accidentally printed for the final match.
3226
3227 (g) Added the -H (--with-filename) option.
3228
3229 (h) The combination of options -rh failed to suppress file names for files
3230 that were found from directory arguments.
3231
3232 (i) Added the -D (--devices) and -d (--directories) options.
3233
3234 (j) Added the -F (--fixed-strings) option.
3235
3236 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3237
3238 (l) Added the --colo(u)r option.
3239
3240 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3241 is not present by default.
3242
3243 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3244 items such as (?R) or (?1), when the recursion could match a number of
3245 alternatives. If it matched one of the alternatives, but subsequently,
3246 outside the recursion, there was a failure, the code tried to back up into
3247 the recursion. However, because of the way PCRE is implemented, this is not
3248 possible, and the result was an incorrect result from the match.
3249
3250 In order to prevent this happening, the specification of recursion has
3251 been changed so that all such subpatterns are automatically treated as
3252 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3253
3254 4. I had overlooked the fact that, in some locales, there are characters for
3255 which isalpha() is true but neither isupper() nor islower() are true. In
3256 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3257 and ordfeminine) are like this. This affected the treatment of \w and \W
3258 when they appeared in character classes, but not when they appeared outside
3259 a character class. The bit map for "word" characters is now created
3260 separately from the results of isalnum() instead of just taking it from the
3261 upper, lower, and digit maps. (Plus the underscore character, of course.)
3262
3263 5. The above bug also affected the handling of POSIX character classes such as
3264 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3265 permanent tables. Instead, the bit maps for such a class were previously
3266 created as the appropriate unions of the upper, lower, and digit bitmaps.
3267 Now they are created by subtraction from the [[:word:]] class, which has
3268 its own bitmap.
3269
3270 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3271 It is created by subtracting the vertical space characters (\x09, \x0a,
3272 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3273 subtraction was done in the overall bitmap for a character class, meaning
3274 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3275 be recognized. This bug has been fixed.
3276
3277 7. Patches from the folks at Google:
3278
3279 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3280 real life, but is still worth protecting against".
3281
3282 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3283 regular expressions".
3284
3285 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3286 have it.
3287
3288 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3289 "configure" and the latter not, in order to fix a problem somebody had
3290 with compiling the Arg class on HP-UX.
3291
3292 (e) Improve the error-handling of the C++ wrapper a little bit.
3293
3294 (f) New tests for checking recursion limiting.
3295
3296 8. The pcre_memmove() function, which is used only if the environment does not
3297 have a standard memmove() function (and is therefore rarely compiled),
3298 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3299 returning a result (though PCRE never actually uses the result).
3300
3301 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3302 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3303 returned instead of calling malloc() with an overflowing number that would
3304 most likely cause subsequent chaos.
3305
3306 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3307
3308 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3309 with this option is matched, the nmatch and pmatch options of regexec() are
3310 ignored.
3311
3312 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3313 provided in case anyone wants to the the POSIX interface with UTF-8
3314 strings.
3315
3316 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3317 C++ linking (needed for some HP-UX environments).
3318
3319 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3320 (unused parameter) and in the pcre_printint() function (omitted "default"
3321 switch label when the default is to do nothing).
3322
3323 15. Added some code to make it possible, when PCRE is compiled as a C++
3324 library, to replace subject pointers for pcre_exec() with a smart pointer
3325 class, thus making it possible to process discontinuous strings.
3326
3327 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3328 much the same function. They were added by different people who were trying
3329 to make PCRE easy to compile on non-Unix systems. It has been suggested
3330 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3331 for compiling on Windows systems. I have therefore replaced it with
3332 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3333 defaults to "extern" for C or "extern C" for C++, which works fine on
3334 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3335 SCOPE with something explicit in config.h. In addition:
3336
3337 (a) pcreposix.h still had just "extern" instead of either of these macros;
3338 I have replaced it with PCRE_DATA_SCOPE.
3339
3340 (b) Functions such as _pcre_xclass(), which are internal to the library,
3341 but external in the C sense, all had PCRE_EXPORT in their definitions.
3342 This is apparently wrong for the Windows case, so I have removed it.
3343 (It makes no difference on Unix-like systems.)
3344
3345 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3346 of recursive calls to match(). This is different to MATCH_LIMIT because
3347 that limits the total number of calls to match(), not all of which increase
3348 the depth of recursion. Limiting the recursion depth limits the amount of
3349 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3350 when PCRE is compiled, and changed at run time. A patch from Google adds
3351 this functionality to the C++ interface.
3352
3353 18. Changes to the handling of Unicode character properties:
3354
3355 (a) Updated the table to Unicode 4.1.0.
3356
3357 (b) Recognize characters that are not in the table as "Cn" (undefined).
3358
3359 (c) I revised the way the table is implemented to a much improved format
3360 which includes recognition of ranges. It now supports the ranges that
3361 are defined in UnicodeData.txt, and it also amalgamates other
3362 characters into ranges. This has reduced the number of entries in the
3363 table from around 16,000 to around 3,000, thus reducing its size
3364 considerably. I realized I did not need to use a tree structure after
3365 all - a binary chop search is just as efficient. Having reduced the
3366 number of entries, I extended their size from 6 bytes to 8 bytes to
3367 allow for more data.
3368
3369 (d) Added support for Unicode script names via properties such as \p{Han}.
3370
3371 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3372 matching that character.
3373
3374 20. When matching a repeated Unicode property with a minimum greater than zero,
3375 (for example \pL{2,}), PCRE could look past the end of the subject if it
3376 reached it while seeking the minimum number of characters. This could
3377 happen only if some of the characters were more than one byte long, because
3378 there is a check for at least the minimum number of bytes.
3379
3380 21. Refactored the implementation of \p and \P so as to be more general, to
3381 allow for more different types of property in future. This has changed the
3382 compiled form incompatibly. Anybody with saved compiled patterns that use
3383 \p or \P will have to recompile them.
3384
3385 22. Added "Any" and "L&" to the supported property types.
3386
3387 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3388 but give a compile time error if the value is greater than 0xff.
3389
3390 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3391 accidentally not being installed or uninstalled.
3392
3393 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3394 made were to insert the current release number. This seemed silly, because
3395 it made things harder for people building PCRE on systems that don't run
3396 "configure". I have turned pcre.h into a distributed file, no longer built
3397 by "configure", with the version identification directly included. There is
3398 no longer a pcre.h.in file.
3399
3400 However, this change necessitated a change to the pcre-config script as
3401 well. It is built from pcre-config.in, and one of the substitutions was the
3402 release number. I have updated configure.ac so that ./configure now finds
3403 the release number by grepping pcre.h.
3404
3405 26. Added the ability to run the tests under valgrind.
3406
3407
3408 Version 6.4 05-Sep-05
3409 ---------------------
3410
3411 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3412 "--" to be printed when multiple files were scanned, even when none of the
3413 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3414 consider it to be a bug, and have restored the previous behaviour.
3415
3416 2. A couple of code tidies to get rid of compiler warnings.
3417
3418 3. The pcretest program used to cheat by referring to symbols in the library
3419 whose names begin with _pcre_. These are internal symbols that are not
3420 really supposed to be visible externally, and in some environments it is
3421 possible to suppress them. The cheating is now confined to including
3422 certain files from the library's source, which is a bit cleaner.
3423
3424 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3425 file's purpose clearer.
3426
3427 5. Reorganized pcre_ucp_findchar().
3428
3429
3430 Version 6.3 15-Aug-05
3431 ---------------------
3432
3433 1. The file libpcre.pc.in did not have general read permission in the tarball.
3434
3435 2. There were some problems when building without C++ support:
3436
3437 (a) If C++ support was not built, "make install" and "make test" still
3438 tried to test it.
3439
3440 (b) There were problems when the value of CXX was explicitly set. Some
3441 changes have been made to try to fix these, and ...
3442
3443 (c) --disable-cpp can now be used to explicitly disable C++ support.
3444
3445 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3446 backslash in a target when C++ was disabled. This confuses some
3447 versions of "make", apparently. Using an intermediate variable solves
3448 this. (Same for CPP_LOBJ.)
3449
3450 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3451 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3452 necessary on certain architectures.
3453
3454 4. Added a setting of -export-symbols-regex to the link command to remove
3455 those symbols that are exported in the C sense, but actually are local
3456 within the library, and not documented. Their names all begin with
3457 "_pcre_". This is not a perfect job, because (a) we have to except some
3458 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3459 available (and never for static libraries). I have made a note to try to
3460 find a way round (a) in the future.
3461
3462
3463 Version 6.2 01-Aug-05
3464 ---------------------
3465
3466 1. There was no test for integer overflow of quantifier values. A construction
3467 such as {1111111111111111} would give undefined results. What is worse, if
3468 a minimum quantifier for a parenthesized subpattern overflowed and became
3469 negative, the calculation of the memory size went wrong. This could have
3470 led to memory overwriting.
3471
3472 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3473
3474 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3475 operating environments where this matters.
3476
3477 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3478 PCRE options from within the C++ wrapper.
3479
3480 5. Named capturing subpatterns were not being correctly counted when a pattern
3481 was compiled. This caused two problems: (a) If there were more than 100
3482 such subpatterns, the calculation of the memory needed for the whole
3483 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3484 back references of the form \12, where the number was greater than 9, were
3485 not recognized as back references, even though there were sufficient
3486 previous subpatterns.
3487
3488 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3489 versions of gcc, e.g. 2.95.4.
3490
3491
3492 Version 6.1 21-Jun-05
3493 ---------------------
3494
3495 1. There was one reference to the variable "posix" in pcretest.c that was not
3496 surrounded by "#if !defined NOPOSIX".
3497
3498 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3499 the cross-check on the old pcre_info() function, for the benefit of the
3500 cut-down version of PCRE that is currently imported into Exim.
3501
3502 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3503 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3504 patterns that start (?i) but I don't think that matters. The use of (?i) is
3505 just an example; this all applies to the other options as well.
3506
3507 4. Since libtool seems to echo the compile commands it is issuing, the output
3508 from "make" can be reduced a bit by putting "@" in front of each libtool
3509 compile command.
3510
3511 5. Patch from the folks at Google for configure.in to be a bit more thorough
3512 in checking for a suitable C++ installation before trying to compile the
3513 C++ stuff. This should fix a reported problem when a compiler was present,
3514 but no suitable headers.
3515
3516 6. The man pages all had just "PCRE" as their title. I have changed them to
3517 be the relevant file name. I have also arranged that these names are
3518 retained in the file doc/pcre.txt, which is a concatenation in text format
3519 of all the man pages except the little individual ones for each function.
3520
3521 7. The NON-UNIX-USE file had not been updated for the different set of source
3522 files that come with release 6. I also added a few comments about the C++
3523 wrapper.
3524
3525
3526 Version 6.0 07-Jun-05
3527 ---------------------
3528
3529 1. Some minor internal re-organization to help with my DFA experiments.
3530
3531 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3532 didn't matter for the library itself when fully configured, but did matter
3533 when compiling without UCP support, or within Exim, where the ucp files are
3534 not imported.
3535
3536 3. Refactoring of the library code to split up the various functions into
3537 different source modules. The addition of the new DFA matching code (see
3538 below) to a single monolithic source would have made it really too
3539 unwieldy, quite apart from causing all the code to be include in a
3540 statically linked application, when only some functions are used. This is
3541 relevant even without the DFA addition now that patterns can be compiled in
3542 one application and matched in another.
3543
3544 The downside of splitting up is that there have to be some external
3545 functions and data tables that are used internally in different modules of
3546 the library but which are not part of the API. These have all had their
3547 names changed to start with "_pcre_" so that they are unlikely to clash
3548 with other external names.
3549
3550 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3551 a different (DFA) algorithm. Although it is slower than the original
3552 function, it does have some advantages for certain types of matching
3553 problem.
3554
3555 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3556 including restarting after a partial match.
3557
3558 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3559 defined when compiling for Windows was sent to me. I have put it into the
3560 code, though I have no means of testing or verifying it.
3561
3562 7. Added the pcre_refcount() auxiliary function.
3563
3564 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3565 match before or at the first newline in the subject string. In pcretest,
3566 the /f option on a pattern can be used to set this.
3567
3568 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3569 would behave wrongly. This has been present in PCRE since release 4.0.
3570
3571 10. A number of changes to the pcregrep command:
3572
3573 (a) Refactored how -x works; insert ^(...)$ instead of setting
3574 PCRE_ANCHORED and checking the length, in preparation for adding
3575 something similar for -w.
3576
3577 (b) Added the -w (match as a word) option.
3578
3579 (c) Refactored the way lines are read and buffered so as to have more
3580 than one at a time available.
3581
3582 (d) Implemented a pcregrep test script.
3583
3584 (e) Added the -M (multiline match) option. This allows patterns to match
3585 over several lines of the subject. The buffering ensures that at least
3586 8K, or the rest of the document (whichever is the shorter) is available
3587 for matching (and similarly the previous 8K for lookbehind assertions).
3588
3589 (f) Changed the --help output so that it now says
3590
3591 -w, --word-regex(p)
3592
3593 instead of two lines, one with "regex" and the other with "regexp"
3594 because that confused at least one person since the short forms are the
3595 same. (This required a bit of code, as the output is generated
3596 automatically from a table. It wasn't just a text change.)
3597
3598 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3599 option but starts with a hyphen. Could be a pattern or a path name
3600 starting with a hyphen, for instance.
3601
3602 (h) "-" can be given as a file name to represent stdin.
3603
3604 (i) When file names are being printed, "(standard input)" is used for
3605 the standard input, for compatibility with GNU grep. Previously
3606 "<stdin>" was used.
3607
3608 (j) The option --label=xxx can be used to supply a name to be used for
3609 stdin when file names are being printed. There is no short form.
3610
3611 (k) Re-factored the options decoding logic because we are going to add
3612 two more options that take data. Such options can now be given in four
3613 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3614
3615 (l) Added the -A, -B, and -C options for requesting that lines of context
3616 around matches be printed.
3617
3618 (m) Added the -L option to print the names of files that do not contain
3619 any matching lines, that is, the complement of -l.
3620
3621 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3622 continue to scan other files.
3623
3624 (o) The -s option was incorrectly implemented. For compatibility with other
3625 greps, it now suppresses the error message for a non-existent or non-
3626 accessible file (but not the return code). There is a new option called
3627 -q that suppresses the output of matching lines, which was what -s was
3628 previously doing.
3629
3630 (p) Added --include and --exclude options to specify files for inclusion
3631 and exclusion when recursing.
3632
3633 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3634 Hopefully, it now does.
3635
3636 12. Missing cast in pcre_study().
3637
3638 13. Added an "uninstall" target to the makefile.
3639
3640 14. Replaced "extern" in the function prototypes in Makefile.in with
3641 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3642 world, but is set differently for Windows.
3643
3644 15. Added a second compiling function called pcre_compile2(). The only
3645 difference is that it has an extra argument, which is a pointer to an
3646 integer error code. When there is a compile-time failure, this is set
3647 non-zero, in addition to the error test pointer being set to point to an
3648 error message. The new argument may be NULL if no error number is required
3649 (but then you may as well call pcre_compile(), which is now just a
3650 wrapper). This facility is provided because some applications need a
3651 numeric error indication, but it has also enabled me to tidy up the way
3652 compile-time errors are handled in the POSIX wrapper.
3653
3654 16. Added VPATH=.libs to the makefile; this should help when building with one
3655 prefix path and installing with another. (Or so I'm told by someone who
3656 knows more about this stuff than I do.)
3657
3658 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3659 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3660 match everything, including newlines. This is not POSIX-compatible, but
3661 somebody wanted the feature. From pcretest it can be activated by using
3662 both the P and the s flags.
3663
3664 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3665
3666 19. libpcre.pc was being incorrectly installed as executable.
3667
3668 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3669 it now also looks for '\r' so that it will work unmodified on Windows.
3670
3671 21. Added Google's contributed C++ wrapper to the distribution.
3672
3673 22. Added some untidy missing memory free() calls in pcretest, to keep
3674 Electric Fence happy when testing.
3675
3676
3677
3678 Version 5.0 13-Sep-04
3679 ---------------------
3680
3681 1. Internal change: literal characters are no longer packed up into items
3682 containing multiple characters in a single byte-string. Each character
3683 is now matched using a separate opcode. However, there may be more than one
3684 byte in the character in UTF-8 mode.
3685
3686 2. The pcre_callout_block structure has two new fields: pattern_position and
3687 next_item_length. These contain the offset in the pattern to the next match
3688 item, and its length, respectively.
3689
3690 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3691 insertion of callouts before each pattern item. Added the /C option to
3692 pcretest to make use of this.
3693
3694 4. On the advice of a Windows user, the lines
3695
3696 #if defined(_WIN32) || defined(WIN32)
3697 _setmode( _fileno( stdout ), 0x8000 );
3698 #endif /* defined(_WIN32) || defined(WIN32) */
3699
3700 have been added to the source of pcretest. This apparently does useful
3701 magic in relation to line terminators.
3702
3703 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3704 for the benefit of those environments where the "b" makes a difference.
3705
3706 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3707 to know about it. I have put a hack into configure.in that adds in code
3708 to set GCC=yes if CC=icc. This seems to end up at a point in the
3709 generated configure script that is early enough to affect the setting of
3710 compiler options, which is what is needed, but I have no means of testing
3711 whether it really works. (The user who reported this had patched the
3712 generated configure script, which of course I cannot do.)
3713
3714 LATER: After change 22 below (new libtool files), the configure script
3715 seems to know about icc (and also ecc). Therefore, I have commented out
3716 this hack in configure.in.
3717
3718 7. Added support for pkg-config (2 patches were sent in).
3719
3720 8. Negated POSIX character classes that used a combination of internal tables
3721 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3722 [[:^ascii]]. Typically, they would match almost any characters. The other
3723 POSIX classes were not broken in this way.
3724
3725 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3726 to find the match, as PCRE was deluded into thinking that the match had to
3727 start at the start point or following a newline. The same bug applied to
3728 patterns with negative forward assertions or any backward assertions
3729 preceding ".*" at the start, unless the pattern required a fixed first
3730 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3731
3732 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3733 starting at the last subject character, bytes beyond the end of the subject
3734 string were read.
3735
3736 11. Renamed the variable "class" as "classbits" to make life easier for C++
3737 users. (Previously there was a macro definition, but it apparently wasn't
3738 enough.)
3739
3740 12. Added the new field "tables" to the extra data so that tables can be passed
3741 in at exec time, or the internal tables can be re-selected. This allows
3742 a compiled regex to be saved and re-used at a later time by a different
3743 program that might have everything at different addresses.
3744
3745 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3746 -R library as well as a -L library.
3747
3748 14. The debugging options of pcretest (-d on the command line or D on a
3749 pattern) showed incorrect output for anything following an extended class
3750 that contained multibyte characters and which was followed by a quantifier.
3751
3752 15. Added optional support for general category Unicode character properties
3753 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3754 support. It adds about 90K to the size of the library. The meanings of the
3755 inbuilt class escapes such as \d and \s have NOT been changed.
3756
3757 16. Updated pcredemo.c to include calls to free() to release the memory for the
3758 compiled pattern.
3759
3760 17. The generated file chartables.c was being created in the source directory
3761 instead of in the building directory. This caused the build to fail if the
3762 source directory was different from the building directory, and was
3763 read-only.
3764
3765 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3766 file. No doubt somebody will tell me if they don't make sense... Also added
3767 Dan Mooney's comments about building on OpenVMS.
3768
3769 19. Added support for partial matching via the PCRE_PARTIAL option for
3770 pcre_exec() and the \P data escape in pcretest.
3771
3772 20. Extended pcretest with 3 new pattern features:
3773
3774 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3775 write the compiled pattern to the file whose name is "rest-of-line".
3776 This is a straight binary dump of the data, with the saved pointer to
3777 the character tables forced to be NULL. The study data, if any, is
3778 written too. After writing, pcretest reads a new pattern.
3779
3780 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3781 compiled pattern from the given file. There must not be any
3782 occurrences of "<" in the file name (pretty unlikely); if there are,
3783 pcretest will instead treat the initial "<" as a pattern delimiter.
3784 After reading in the pattern, pcretest goes on to read data lines as
3785 usual.
3786
3787 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3788 and 16-bit fields in a compiled pattern, to simulate a pattern that
3789 was compiled on a host of opposite endianness.
3790
3791 21. The pcre-exec() function can now cope with patterns that were compiled on
3792 hosts of opposite endianness, with this restriction:
3793
3794 As for any compiled expression that is saved and used later, the tables
3795 pointer field cannot be preserved; the extra_data field in the arguments
3796 to pcre_exec() should be used to pass in a tables address if a value
3797 other than the default internal tables were used at compile time.
3798
3799 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3800 now diagnosed as an error. Previously, most of the time, a negative number
3801 would have been treated as zero, but if in addition "ovector" was passed as
3802 NULL, a crash could occur.
3803
3804 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3805 new versions from the libtool 1.5 distribution (the last one is a copy of
3806 a file called libtool.m4). This seems to have fixed the need to patch
3807 "configure" to support Darwin 1.3 (which I used to do). However, I still
3808 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3809 workstation).
3810
3811 24. Changed the PCRE licence to be the more standard "BSD" licence.
3812
3813
3814 Version 4.5 01-Dec-03
3815 ---------------------
3816
3817 1. There has been some re-arrangement of the code for the match() function so
3818 that it can be compiled in a version that does not call itself recursively.
3819 Instead, it keeps those local variables that need separate instances for
3820 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3821 needs to "recurse". Keeping track of where control must go is done by means
3822 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3823 hide most of the details from the main code, and operates only if
3824 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3825 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3826 operating.
3827
3828 To make it easier for callers to provide specially tailored get/free
3829 functions for this usage, two new functions, pcre_stack_malloc, and
3830 pcre_stack_free, are used. They are always called in strict stacking order,
3831 and the size of block requested is always the same.
3832
3833 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3834 PCRE has been compiled to use the stack or the heap for recursion. The
3835 -C option of pcretest uses this to show which version is compiled.
3836
3837 A new data escape \S, is added to pcretest; it causes the amounts of store
3838 obtained and freed by both kinds of malloc/free at match time to be added
3839 to the output.
3840
3841 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3842 what's available on my current Linux desktop machine.
3843
3844 3. When matching a UTF-8 string, the test for a valid string at the start has
3845 been extended. If start_offset is not zero, PCRE now checks that it points
3846 to a byte that is the start of a UTF-8 character. If not, it returns
3847 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3848 this is necessary because there may be backward assertions in the pattern.
3849 When matching the same subject several times, it may save resources to use
3850 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3851
3852 4. The code for checking the validity of UTF-8 strings has been tightened so
3853 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3854 containing "overlong sequences".
3855
3856 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3857 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3858 should have been "&", but it just so happened that all the cases this let
3859 through by mistake were picked up later in the function.
3860
3861 6. I had used a variable called "isblank" - this is a C99 function, causing
3862 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3863
3864 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3865 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3866 so the version is shown; (c) stop "make test" echoing "./RunTest".
3867
3868 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3869
3870 9. The prototype for memmove() for systems that don't have it was using
3871 size_t, but the inclusion of the header that defines size_t was later. I've
3872 moved the #includes for the C headers earlier to avoid this.
3873
3874 10. Added some adjustments to the code to make it easier to compiler on certain
3875 special systems:
3876
3877 (a) Some "const" qualifiers were missing.
3878 (b) Added the macro EXPORT before all exported functions; by default this
3879 is defined to be empty.
3880 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3881 that it reads its output file name as an argument instead of writing
3882 to the standard output and assuming this can be redirected.
3883
3884 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3885 class containing characters with values greater than 255, PCRE compilation
3886 went into a loop.
3887
3888 12. A recursive reference to a subpattern that was within another subpattern
3889 that had a minimum quantifier of zero caused PCRE to crash. For example,
3890 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3891 recursion. If the recursively-called subpattern itself had a zero repeat,
3892 that was OK.
3893
3894 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3895 buffer into which it was copied (for escape processing) was still set at
3896 1024, so long lines caused crashes.
3897
3898 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3899 "internal error: code overflow...". This applied to any character class
3900 that was followed by a possessive quantifier.
3901
3902 15. Modified the Makefile to add libpcre.la as a prerequisite for
3903 libpcreposix.la because I was told this is needed for a parallel build to
3904 work.
3905
3906 16. If a pattern that contained .* following optional items at the start was
3907 studied, the wrong optimizing data was generated, leading to matching
3908 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3909 matching string must start with a or b or c. The correct conclusion for
3910 this pattern is that a match can start with any character.
3911
3912
3913 Version 4.4 13-Aug-03
3914 ---------------------
3915
3916 1. In UTF-8 mode, a character class containing characters with values between
3917 127 and 255 was not handled correctly if the compiled pattern was studied.
3918 In fixing this, I have also improved the studying algorithm for such
3919 classes (slightly).
3920
3921 2. Three internal functions had redundant arguments passed to them. Removal
3922 might give a very teeny performance improvement.
3923
3924 3. Documentation bug: the value of the capture_top field in a callout is *one
3925 more than* the number of the hightest numbered captured substring.
3926
3927 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3928 in incorrectly linking with a previously installed version. They now link
3929 explicitly with libpcre.la.
3930
3931 5. configure.in no longer needs to recognize Cygwin specially.
3932
3933 6. A problem in pcre.in for Windows platforms is fixed.
3934
3935 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3936 pcretest, it used to include the size of the study block as part of its
3937 output. Unfortunately, the structure contains a field that has a different
3938 size on different hardware architectures. This meant that the tests that
3939 showed this size failed. As the block is currently always of a fixed size,
3940 this information isn't actually particularly useful in pcretest output, so
3941 I have just removed it.
3942
3943 8. Three pre-processor statements accidentally did not start in column 1.
3944 Sadly, there are *still* compilers around that complain, even though
3945 standard C has not required this for well over a decade. Sigh.
3946
3947 9. In pcretest, the code for checking callouts passed small integers in the
3948 callout_data field, which is a void * field. However, some picky compilers
3949 complained about the casts involved for this on 64-bit systems. Now
3950 pcretest passes the address of the small integer instead, which should get
3951 rid of the warnings.
3952
3953 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3954 both compile and run time, and gives an error if an invalid UTF-8 sequence
3955 is found. There is a option for disabling this check in cases where the
3956 string is known to be correct and/or the maximum performance is wanted.
3957
3958 11. In response to a bug report, I changed one line in Makefile.in from
3959
3960 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3961 to
3962 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3963
3964 to look similar to other lines, but I have no way of telling whether this
3965 is the right thing to do, as I do not use Windows. No doubt I'll get told
3966 if it's wrong...
3967
3968
3969 Version 4.3 21-May-03
3970 ---------------------
3971
3972 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3973 Makefile.
3974
3975 2. Some refactoring to improve the quality of the code:
3976
3977 (i) The utf8_table... variables are now declared "const".
3978
3979 (ii) The code for \cx, which used the "case flipping" table to upper case
3980 lower case letters, now just substracts 32. This is ASCII-specific,
3981 but the whole concept of \cx is ASCII-specific, so it seems
3982 reasonable.
3983
3984 (iii) PCRE was using its character types table to recognize decimal and
3985 hexadecimal digits in the pattern. This is silly, because it handles
3986 only 0-9, a-f, and A-F, but the character types table is locale-
3987 specific, which means strange things might happen. A private
3988 table is now used for this - though it costs 256 bytes, a table is
3989 much faster than multiple explicit tests. Of course, the standard
3990 character types table is still used for matching digits in subject
3991 strings against \d.
3992
3993 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3994 ending in _t are). So I've renamed it as ESC_tee.
3995
3996 3. The first argument for regexec() in the POSIX wrapper should have been
3997 defined as "const".
3998
3999 4. Changed pcretest to use malloc() for its buffers so that they can be
4000 Electric Fenced for debugging.
4001
4002 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4003 to read one or more bytes before the start of the subject string. Often this
4004 had no effect on PCRE's behaviour, but in some circumstances it could
4005 provoke a segmentation fault.
4006
4007 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4008 to try to read one or more bytes before the start of the subject string.
4009
4010 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4011 UTF-8 support could misbehave in various ways if the subject string
4012 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4013 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4014 back over UTF-8 characters.)
4015
4016
4017 Version 4.2 14-Apr-03
4018 ---------------------
4019
4020 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4021
4022 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4023 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4024 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4025 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4026 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4027 and BUILD_EXEEXT
4028 Note: automatic setting of the BUILD variables is not yet working
4029 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4030 compile-time but not at link-time
4031 [LINK]: use for linking executables only
4032 make different versions for Windows and non-Windows
4033 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4034 libraries
4035 [LINK_FOR_BUILD]: new variable
4036 [OBJEXT]: use throughout
4037 [EXEEXT]: use throughout
4038 <winshared>: new target
4039 <wininstall>: new target
4040 <dftables.o>: use native compiler
4041 <dftables>: use native linker
4042 <install>: handle Windows platform correctly
4043 <clean>: ditto
4044 <check>: ditto
4045 copy DLL to top builddir before testing
4046
4047 As part of these changes, -no-undefined was removed again. This was reported
4048 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4049 in any case.
4050
4051 3. Some tidies to get rid of compiler warnings:
4052
4053 . In the match_data structure, match_limit was an unsigned long int, whereas
4054 match_call_count was an int. I've made them both unsigned long ints.
4055
4056 . In pcretest the fact that a const uschar * doesn't automatically cast to
4057 a void * provoked a warning.
4058
4059 . Turning on some more compiler warnings threw up some "shadow" variables
4060 and a few more missing casts.
4061
4062 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4063 option, a class that contained a single character with a value between 128
4064 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4065
4066 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4067 option, a class that contained several characters, but with at least one
4068 whose value was between 128 and 255 caused PCRE to crash.
4069
4070
4071 Version 4.1 12-Mar-03
4072 ---------------------
4073
4074 1. Compiling with gcc -pedantic found a couple of places where casts were
4075 needed, and a string in dftables.c that was longer than standard compilers are
4076 required to support.
4077
4078 2. Compiling with Sun's compiler found a few more places where the code could
4079 be tidied up in order to avoid warnings.
4080
4081 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4082 first of these names is deprecated in the latest Autoconf in favour of the name
4083 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4084 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4085 analogy I have changed it to CFLAGS_FOR_BUILD.
4086
4087 4. Added -no-undefined to the linking command in the Makefile, because this is
4088 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4089 linking step for the pcreposix library.
4090
4091 5. PCRE was failing to diagnose the case of two named groups with the same
4092 name.
4093
4094 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4095 literal character that is needed in the subject for a match, and scans along to
4096 ensure that it is present before embarking on the full matching process. This
4097 saves time in cases of nested unlimited repeats that are never going to match.
4098 Problem: the scan can take a lot of time if the subject is very long (e.g.
4099 megabytes), thus penalizing straightforward matches. It is now done only if the
4100 amount of subject to be scanned is less than 1000 bytes.
4101
4102 7. A lesser problem with the same optimization is that it was recording the
4103 first character of an anchored pattern as "needed", thus provoking a search
4104 right along the subject, even when the first match of the pattern was going to
4105 fail. The "needed" character is now not set for anchored patterns, unless it
4106 follows something in the pattern that is of non-fixed length. Thus, it still
4107 fulfils its original purpose of finding quick non-matches in cases of nested
4108 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4109
4110
4111 Version 4.0 17-Feb-03
4112 ---------------------
4113
4114 1. If a comment in an extended regex that started immediately after a meta-item
4115 extended to the end of string, PCRE compiled incorrect data. This could lead to
4116 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4117
4118 2. Moved to autoconf 2.53 and libtool 1.4.2.
4119
4120 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4121 the special perltest8 script is no longer needed - all the tests can be run
4122 from a single perltest script.
4123
4124 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4125 by \s. It has now been removed in PCRE. This means it isn't recognized as
4126 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4127 class [:space:] *does* include VT, thereby creating a mess.
4128
4129 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4130 space and tab.
4131
4132 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4133 its new features into the main test script, reducing the number of scripts.
4134
4135 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4136 were backward compatible, and made the (?i) apply to the whole pattern, as if
4137 /i were given. Now it behaves more logically, and applies the option setting
4138 only to what follows. PCRE has been changed to follow suit. However, if it
4139 finds options settings right at the start of the pattern, it extracts them into
4140 the global options, as before. Thus, they show up in the info data.
4141
4142 8. Added support for the \Q...\E escape sequence. Characters in between are
4143 treated as literals. This is slightly different from Perl in that $ and @ are
4144 also handled as literals inside the quotes. In Perl, they will cause variable
4145 interpolation. Note the following examples:
4146
4147 Pattern PCRE matches Perl matches
4148
4149 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4150 \Qabc\$xyz\E abc\$xyz abc\$xyz
4151 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4152
4153 For compatibility with Perl, \Q...\E sequences are recognized inside character
4154 classes as well as outside them.
4155
4156 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4157 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4158 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4159 signed/unsigned warnings.
4160
4161 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4162 option for pcretest, so I've replaced it by a simple function that does just
4163 that job.
4164
4165 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4166 "pcregrep -".
4167
4168 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4169 Java package. This provides some syntactic sugar for simple cases of what my
4170 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4171 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4172 item, you can use this simplified notation. Note that only makes sense with
4173 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4174 greediness, whatever the setting of the PCRE_UNGREEDY option.
4175
4176 13. A change of greediness default within a pattern was not taking effect at
4177 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4178 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4179 was abstracted outside.
4180
4181 14. PCRE now supports the \G assertion. It is true when the current matching
4182 position is at the start point of the match. This differs from \A when the
4183 starting offset is non-zero. Used with the /g option of pcretest (or similar
4184 code), it works in the same way as it does for Perl's /g option. If all
4185 alternatives of a regex begin with \G, the expression is anchored to the start
4186 match position, and the "anchored" flag is set in the compiled expression.
4187
4188 15. Some bugs concerning the handling of certain option changes within patterns
4189 have been fixed. These applied to options other than (?ims). For example,
4190 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4191 been the other way round. Some of this was related to change 7 above.
4192
4193 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4194 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4195 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4196 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4197
4198 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4199 mode. Unlike ".", it always matches newline, whatever the setting of
4200 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4201 assertions. Perl allows it, but it doesn't (in general) work because it can't
4202 calculate the length of the lookbehind. At least, that's the case for Perl
4203 5.8.0 - I've been told they are going to document that it doesn't work in
4204 future.
4205
4206 18. Added an error diagnosis for escapes that PCRE does not support: these are
4207 \L, \l, \N, \P, \p, \U, \u, and \X.
4208
4209 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4210 reading past the end of the pattern in cases such as /[abcd/.
4211
4212 20. PCRE was getting more memory than necessary for patterns with classes that
4213 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4214
4215 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4216 compiling PCRE for use with Virtual Pascal.
4217
4218 22. Small fix to the Makefile to make it work properly if the build is done
4219 outside the source tree.
4220
4221 23. Added a new extension: a condition to go with recursion. If a conditional
4222 subpattern starts with (?(R) the "true" branch is used if recursion has
4223 happened, whereas the "false" branch is used only at the top level.
4224
4225 24. When there was a very long string of literal characters (over 255 bytes
4226 without UTF support, over 250 bytes with UTF support), the computation of how
4227 much memory was required could be incorrect, leading to segfaults or other
4228 strange effects.
4229
4230 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4231 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4232 there was a subsequent back reference to those brackets. This meant that, for
4233 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4234 possible to check for precisely this case. All we can do is abandon the
4235 optimization if .* occurs inside capturing brackets when there are any back
4236 references whatsoever. (See below for a better fix that came later.)
4237
4238 26. The handling of the optimization for finding the first character of a
4239 non-anchored pattern, and for finding a character that is required later in the
4240 match were failing in some cases. This didn't break the matching; it just
4241 failed to optimize when it could. The way this is done has been re-implemented.
4242
4243 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4244
4245 28. Added a new feature that provides some of the functionality that Perl
4246 provides with (?{...}). The facility is termed a "callout". The way it is done
4247 in PCRE is for the caller to provide an optional function, by setting
4248 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4249 global variable. By default it is unset, which disables all calling out. To get
4250 the function called, the regex must include (?C) at appropriate points. This
4251 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4252 This provides a means of identifying different callout points. When PCRE
4253 reaches such a point in the regex, if pcre_callout has been set, the external
4254 function is called. It is provided with data in a structure called
4255 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4256 matching continues; if it returns a non-zero value, the match at the current
4257 point fails. However, backtracking will occur if possible. [This was changed
4258 later and other features added - see item 49 below.]
4259
4260 29. pcretest is upgraded to test the callout functionality. It provides a
4261 callout function that displays information. By default, it shows the start of
4262 the match and the current position in the text. There are some new data escapes
4263 to vary what happens:
4264
4265 \C+ in addition, show current contents of captured substrings
4266 \C- do not supply a callout function
4267 \C!n return 1 when callout number n is reached
4268 \C!n!m return 1 when callout number n is reached for the mth time
4269
4270 30. If pcregrep was called with the -l option and just a single file name, it
4271 output "<stdin>" if a match was found, instead of the file name.
4272
4273 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4274 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4275 pcre_exec(). This saves a malloc/free per call. The default value of
4276 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4277 when configuring.
4278
4279 32. The default maximum size of a compiled pattern is 64K. There have been a
4280 few cases of people hitting this limit. The code now uses macros to handle the
4281 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4282 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4283 configuring. Tests 2 and 5 work only with 2-byte links because they output
4284 debugging information about compiled patterns.
4285
4286 33. Internal code re-arrangements:
4287
4288 (a) Moved the debugging function for printing out a compiled regex into
4289 its own source file (printint.c) and used #include to pull it into
4290 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4291 separate copies.
4292
4293 (b) Defined the list of op-code names for debugging as a macro in
4294 internal.h so that it is next to the definition of the opcodes.
4295
4296 (c) Defined a table of op-code lengths for simpler skipping along compiled
4297 code. This is again a macro in internal.h so that it is next to the
4298 definition of the opcodes.
4299
4300 34. Added support for recursive calls to individual subpatterns, along the
4301 lines of Robin Houston's patch (but implemented somewhat differently).
4302
4303 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4304 allow it to read and process whole directories in Win32. This code was
4305 contributed by Lionel Fourquaux; it has not been tested by me.
4306
4307 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4308 used to name a group. Names consist of alphanumerics and underscores, and must
4309 be unique. Back references use the syntax (?P=name) and recursive calls use
4310 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4311 numbers. The function pcre_fullinfo() can be used after compilation to extract
4312 a name/number map. There are three relevant calls:
4313
4314 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4315 PCRE_INFO_NAMECOUNT yields the number of entries
4316 PCRE_INFO_NAMETABLE yields a pointer to the map.
4317
4318 The map is a vector of fixed-size entries. The size of each entry depends on
4319 the length of the longest name used. The first two bytes of each entry are the
4320 group number, most significant byte first. There follows the corresponding
4321 name, zero terminated. The names are in alphabetical order.
4322
4323 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4324 case instead of 255. Making it the same both with and without UTF-8 support
4325 means that the same test output works with both.
4326
4327 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4328 calling malloc() with a zero argument.
4329
4330 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4331 optimization. I've improved things by keeping a bitmap of backreferences with
4332 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4333 fact referenced, the optimization can be applied. It is unlikely that a
4334 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4335 the match to follow \n) will appear inside brackets with a number greater than
4336 31, but if it does, any back reference > 31 suppresses the optimization.
4337
4338 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4339 of disabling numbered capturing parentheses. Any opening parenthesis that is
4340 not followed by ? behaves as if it were followed by ?: but named parentheses
4341 can still be used for capturing (and they will acquire numbers in the usual
4342 way).
4343
4344 41. Redesigned the return codes from the match() function into yes/no/error so
4345 that errors can be passed back from deep inside the nested calls. A malloc
4346 failure while inside a recursive subpattern call now causes the
4347 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4348
4349 42. It is now possible to set a limit on the number of times the match()
4350 function is called in a call to pcre_exec(). This facility makes it possible to
4351 limit the amount of recursion and backtracking, though not in a directly
4352 obvious way, because the match() function is used in a number of different
4353 circumstances. The count starts from zero for each position in the subject
4354 string (for non-anchored patterns). The default limit is, for compatibility, a
4355 large number, namely 10 000 000. You can change this in two ways:
4356
4357 (a) When configuring PCRE before making, you can use --with-match-limit=n
4358 to set a default value for the compiled library.
4359
4360 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4361 a different value is set. See 45 below.
4362
4363 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4364
4365 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4366 of things that can be changed at compile time. The first argument specifies
4367 what is wanted and the second points to where the information is to be placed.
4368 The current list of available information is:
4369
4370 PCRE_CONFIG_UTF8
4371
4372 The output is an integer that is set to one if UTF-8 support is available;
4373 otherwise it is set to zero.
4374
4375 PCRE_CONFIG_NEWLINE
4376
4377 The output is an integer that it set to the value of the code that is used for
4378 newline. It is either LF (10) or CR (13).
4379
4380 PCRE_CONFIG_LINK_SIZE
4381
4382 The output is an integer that contains the number of bytes used for internal
4383 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4384
4385 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4386
4387 The output is an integer that contains the threshold above which the POSIX
4388 interface uses malloc() for output vectors. See item 31 above.
4389
4390 PCRE_CONFIG_MATCH_LIMIT
4391
4392 The output is an unsigned integer that contains the default limit of the number
4393 of match() calls in a pcre_exec() execution. See 42 above.
4394
4395 44. pcretest has been upgraded by the addition of the -C option. This causes it
4396 to extract all the available output from the new pcre_config() function, and to
4397 output it. The program then exits immediately.
4398
4399 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4400 order to support additional features. One way would have been to define
4401 pcre_exec2() (for example) with extra arguments, but this would not have been
4402 extensible, and would also have required all calls to the original function to
4403 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4404 is used for passing in "extra" data from pcre_study().
4405
4406 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4407 contains the following fields:
4408
4409 flags a bitmap indicating which of the following fields are set
4410 study_data opaque data from pcre_study()
4411 match_limit a way of specifying a limit on match() calls for a specific
4412 call to pcre_exec()
4413 callout_data data for callouts (see 49 below)
4414
4415 The flag bits are also defined in pcre.h, and are
4416
4417 PCRE_EXTRA_STUDY_DATA
4418 PCRE_EXTRA_MATCH_LIMIT
4419 PCRE_EXTRA_CALLOUT_DATA
4420
4421 The pcre_study() function now returns one of these new pcre_extra blocks, with
4422 the actual study data pointed to by the study_data field, and the
4423 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4424 before. That is, this change is entirely upwards-compatible and requires no
4425 change to existing code.
4426
4427 If you want to pass in additional data to pcre_exec(), you can either place it
4428 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4429 block.
4430
4431 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4432 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4433 times with different match limits, until it finds the minimum value needed for
4434 pcre_exec() to complete. The value is then output. This can be instructive; for
4435 most simple matches the number is quite small, but for pathological cases it
4436 gets very large very quickly.
4437
4438 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4439 returns the size of the data block pointed to by the study_data field in a
4440 pcre_extra block, that is, the value that was passed as the argument to
4441 pcre_malloc() when PCRE was getting memory in which to place the information
4442 created by pcre_study(). The fourth argument should point to a size_t variable.
4443 pcretest has been extended so that this information is shown after a successful
4444 pcre_study() call when information about the compiled regex is being displayed.
4445
4446 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4447 because what follows is always an absolute path. (Later: it turns out that this
4448 is more than cosmetic for MinGW, because it doesn't like empty path
4449 components.)
4450
4451 49. Some changes have been made to the callout feature (see 28 above):
4452
4453 (i) A callout function now has three choices for what it returns:
4454
4455 0 => success, carry on matching
4456 > 0 => failure at this point, but backtrack if possible
4457 < 0 => serious error, return this value from pcre_exec()
4458
4459 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4460 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4461 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4462 use by callout functions. It will never be used by PCRE itself.
4463
4464 (ii) The pcre_extra structure (see 45 above) has a void * field called
4465 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4466 pcre_callout_block structure has a field of the same name. The contents of
4467 the field passed in the pcre_extra structure are passed to the callout
4468 function in the corresponding field in the callout block. This makes it
4469 easier to use the same callout-containing regex from multiple threads. For
4470 testing, the pcretest program has a new data escape
4471
4472 \C*n pass the number n (may be negative) as callout_data
4473
4474 If the callout function in pcretest receives a non-zero value as
4475 callout_data, it returns that value.
4476
4477 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4478 there were some redundant $(CFLAGS) in commands that are now specified as
4479 $(LINK), which already includes $(CFLAGS).
4480
4481 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4482 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4483 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4484 one-byte characters throughout. Note that case-insensitive matching applies
4485 only to characters whose values are less than 256. PCRE doesn't support the
4486 notion of cases for higher-valued characters.
4487
4488 (i) A character class whose characters are all within 0-255 is handled as
4489 a bit map, and the map is inverted for negative classes. Previously, a
4490 character > 255 always failed to match such a class; however it should
4491 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4492
4493 (ii) A negated character class with a single character < 255 is coded as
4494 "not this character" (OP_NOT). This wasn't working properly when the test
4495 character was multibyte, either singly or repeated.
4496
4497 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4498 mode, for example: \x{100}{2,3}.
4499
4500 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4501 singly or repeated) now correctly test multibyte characters. However,
4502 PCRE doesn't recognize any characters with values greater than 255 as
4503 digits, spaces, or word characters. Such characters always match \D, \S,
4504 and \W, and never match \d, \s, or \w.
4505
4506 (v) Classes may now contain characters and character ranges with values
4507 greater than 255. For example: [ab\x{100}-\x{400}].
4508
4509 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4510 PCRE in UTF-8 mode.
4511
4512 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4513 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4514 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4515 value.)
4516
4517 53. The single man page has become too large. I have therefore split it up into
4518 a number of separate man pages. These also give rise to individual HTML pages;
4519 these are now put in a separate directory, and there is an index.html page that
4520 lists them all. Some hyperlinking between the pages has been installed.
4521
4522 54. Added convenience functions for handling named capturing parentheses.
4523
4524 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4525 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4526 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4527 are faulted.
4528
4529 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4530 calling configure. These values are used when compiling the dftables.c program
4531 which is run to generate the source of the default character tables. They
4532 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4533 you will need to set these values.
4534
4535 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4536
4537
4538 Version 3.9 02-Jan-02
4539 ---------------------
4540
4541 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4542
4543 2. If --disable-static was given, the building process failed when trying to
4544 build pcretest and pcregrep. (For some reason it was using libtool to compile
4545 them, which is not right, as they aren't part of the library.)
4546
4547
4548 Version 3.8 18-Dec-01
4549 ---------------------
4550
4551 1. The experimental UTF-8 code was completely screwed up. It was packing the
4552 bytes in the wrong order. How dumb can you get?
4553
4554
4555 Version 3.7 29-Oct-01
4556 ---------------------
4557
4558 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4559 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4560 this didn't happen under Solaris 8, where I normally test things.
4561
4562 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4563 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4564 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4565 this fix an hour or so after the initial 3.7 release.)
4566
4567
4568 Version 3.6 23-Oct-01
4569 ---------------------
4570
4571 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4572 offsets passed as NULL with zero offset count.
4573
4574 2. The config.guess and config.sub files had not been updated when I moved to
4575 the latest autoconf.
4576
4577
4578 Version 3.5 15-Aug-01
4579 ---------------------
4580
4581 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4582 had been forgotten.
4583
4584 2. By using declared but undefined structures, we can avoid using "void"
4585 definitions in pcre.h while keeping the internal definitions of the structures
4586 private.
4587
4588 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4589 user point of view, this means that both static and shared libraries are built
4590 by default, but this can be individually controlled. More of the work of
4591 handling this static/shared cases is now inside libtool instead of PCRE's make
4592 file.
4593
4594 4. The pcretest utility is now installed along with pcregrep because it is
4595 useful for users (to test regexs) and by doing this, it automatically gets
4596 relinked by libtool. The documentation has been turned into a man page, so
4597 there are now .1, .txt, and .html versions in /doc.
4598
4599 5. Upgrades to pcregrep:
4600 (i) Added long-form option names like gnu grep.
4601 (ii) Added --help to list all options with an explanatory phrase.
4602 (iii) Added -r, --recursive to recurse into sub-directories.
4603 (iv) Added -f, --file to read patterns from a file.
4604
4605 6. pcre_exec() was referring to its "code" argument before testing that
4606 argument for NULL (and giving an error if it was NULL).
4607
4608 7. Upgraded Makefile.in to allow for compiling in a different directory from
4609 the source directory.
4610
4611 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4612 options bits, the pointer it was passed was to an int instead of to an unsigned
4613 long int. This mattered only on 64-bit systems.
4614
4615 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4616 generated) instead of pcre.in, which it its source. Also made the same change
4617 in several of the .c files.
4618
4619 10. A new release of gcc defines printf() as a macro, which broke pcretest
4620 because it had an ifdef in the middle of a string argument for printf(). Fixed
4621 by using separate calls to printf().
4622
4623 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4624 script, to force use of CR or LF instead of \n in the source. On non-Unix
4625 systems, the value can be set in config.h.
4626
4627 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4628 absolute limit. Changed the text of the error message to make this clear, and
4629 likewise updated the man page.
4630
4631 13. The limit of 99 on the number of capturing subpatterns has been removed.
4632 The new limit is 65535, which I hope will not be a "real" limit.
4633
4634
4635 Version 3.4 22-Aug-00
4636 ---------------------
4637
4638 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4639
4640 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4641
4642
4643 Version 3.3 01-Aug-00
4644 ---------------------
4645
4646 1. If an octal character was given, but the value was greater than \377, it
4647 was not getting masked to the least significant bits, as documented. This could
4648 lead to crashes in some systems.
4649
4650 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4651 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4652
4653 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4654 These just pass their arguments on to (pcre_free)(), but they are provided
4655 because some uses of PCRE bind it to non-C systems that can call its functions,
4656 but cannot call free() or pcre_free() directly.
4657
4658 4. Add "make test" as a synonym for "make check". Corrected some comments in
4659 the Makefile.
4660
4661 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4662 Makefile.
4663
4664 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4665 command called pgrep for grepping around the active processes.
4666
4667 7. Added the beginnings of support for UTF-8 character strings.
4668
4669 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4670 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4671 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4672 out for the ar command.)
4673
4674
4675 Version 3.2 12-May-00
4676 ---------------------
4677
4678 This is purely a bug fixing release.
4679
4680 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4681 of ZA. This was just one example of several cases that could provoke this bug,
4682 which was introduced by change 9 of version 2.00. The code for breaking
4683 infinite loops after an iteration that matches an empty string was't working
4684 correctly.
4685
4686 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4687 when matched against abbab (for example). After matching an empty string, it
4688 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4689 caused it to match further down the string than it should.
4690
4691 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4692 was there because it doesn't seem to be needed, and it causes trouble on some
4693 systems, as it is not a Standard C header. It has been removed.
4694
4695 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4696 were reported on the Macintosh. The changes were from
4697
4698 while ((c = *(++ptr)) != 0 && c != '\n');
4699 to
4700 while ((c = *(++ptr)) != 0 && c != '\n') ;
4701
4702 Totally extraordinary, but if that's what it takes...
4703
4704 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4705 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4706 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4707 assumes the way PCRE uses memmove() (always moving upwards).
4708
4709 6. PCRE is being used in one environment where strchr() is not available. There
4710 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4711 faster code anyway.
4712
4713
4714 Version 3.1 09-Feb-00
4715 ---------------------
4716
4717 The only change in this release is the fixing of some bugs in Makefile.in for
4718 the "install" target:
4719
4720 (1) It was failing to install pcreposix.h.
4721
4722 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4723
4724
4725 Version 3.0 01-Feb-00
4726 ---------------------
4727
4728 1. Add support for the /+ modifier to perltest (to output $` like it does in
4729 pcretest).
4730
4731 2. Add support for the /g modifier to perltest.
4732
4733 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4734 matches null strings.
4735
4736 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4737 pattern. Perl treats empty patterns specially - it reuses the most recent
4738 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4739 effect.
4740
4741 5. The POSIX interface was broken in that it was just handing over the POSIX
4742 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4743 required a bigger vector, with some working space on the end. This means that
4744 the POSIX wrapper now has to get and free some memory, and copy the results.
4745
4746 6. Added some simple autoconf support, placing the test data and the
4747 documentation in separate directories, re-organizing some of the
4748 information files, and making it build pcre-config (a GNU standard). Also added
4749 libtool support for building PCRE as a shared library, which is now the
4750 default.
4751
4752 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4753 09 are not valid octal constants. Single digits will be used for minor values
4754 less than 10.
4755
4756 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4757 existing programs that set these in the POSIX interface can use PCRE without
4758 modification.
4759
4760 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4761 return all that pcre_info() returns, plus additional data. The pcre_info()
4762 function is retained for compatibility, but is considered to be obsolete.
4763
4764 10. Added experimental recursion feature (?R) to handle one common case that
4765 Perl 5.6 will be able to do with (?p{...}).
4766
4767 11. Added support for POSIX character classes like [:alpha:], which Perl is
4768 adopting.
4769
4770
4771 Version 2.08 31-Aug-99
4772 ----------------------
4773
4774 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4775 trying to match at the startoffset position, but instead was moving forward to
4776 the next newline as if a previous match had failed.
4777
4778 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4779 and could get into a loop if a null string was matched other than at the start
4780 of the subject.
4781
4782 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4783 be distinguished at compile time, and for completeness also added PCRE_DATE.
4784
4785 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4786 in GnuWin32 environments.
4787
4788
4789 Version 2.07 29-Jul-99
4790 ----------------------
4791
4792 1. The documentation is now supplied in plain text form and HTML as well as in
4793 the form of man page sources.
4794
4795 2. C++ compilers don't like assigning (void *) values to other pointer types.
4796 In particular this affects malloc(). Although there is no problem in Standard
4797 C, I've put in casts to keep C++ compilers happy.
4798
4799 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4800 should be (const char *).
4801
4802 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4803 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4804 However, I haven't made this a standard facility. The documentation doesn't
4805 mention it, and the Makefile doesn't support it.
4806
4807 5. The Makefile now contains an "install" target, with editable destinations at
4808 the top of the file. The pcretest program is not installed.
4809
4810 6. pgrep -V now gives the PCRE version number and date.
4811
4812 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4813 causing the entire string to be ignored, instead of just the last character.
4814
4815 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4816 non-matching string, it can take a very, very long time, even for strings of
4817 quite modest length, because of the nested recursion. PCRE now does better in
4818 some of these cases. It does this by remembering the last required literal
4819 character in the pattern, and pre-searching the subject to ensure it is present
4820 before running the real match. In other words, it applies a heuristic to detect
4821 some types of certain failure quickly, and in the above example, if presented
4822 with a string that has no trailing " it gives "no match" very quickly.
4823
4824 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4825 other alternatives are tried instead.
4826
4827
4828 Version 2.06 09-Jun-99
4829 ----------------------
4830
4831 1. Change pcretest's output for amount of store used to show just the code
4832 space, because the remainder (the data block) varies in size between 32-bit and
4833 64-bit systems.
4834
4835 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4836 start matching at. This allows lookbehinds to work when searching for multiple
4837 occurrences in a string.
4838
4839 3. Added additional options to pcretest for testing multiple occurrences:
4840
4841 /+ outputs the rest of the string that follows a match
4842 /g loops for multiple occurrences, using the new startoffset argument
4843 /G loops for multiple occurrences by passing an incremented pointer
4844
4845 4. PCRE wasn't doing the "first character" optimization for patterns starting
4846 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4847 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4848 the letter 'x'. On long subject strings, this gives a significant speed-up.
4849
4850
4851 Version 2.05 21-Apr-99
4852 ----------------------
4853
4854 1. Changed the type of magic_number from int to long int so that it works
4855 properly on 16-bit systems.
4856
4857 2. Fixed a bug which caused patterns starting with .* not to work correctly
4858 when the subject string contained newline characters. PCRE was assuming
4859 anchoring for such patterns in all cases, which is not correct because .* will
4860 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4861 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4862 must be retried after every newline in the subject.
4863
4864
4865 Version 2.04 18-Feb-99
4866 ----------------------
4867
4868 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4869 computation of the store needed to hold the pattern was incorrect (too large).
4870 If such patterns were nested a few deep, this could multiply and become a real
4871 problem.
4872
4873 2. Added /M option to pcretest to show the memory requirement of a specific
4874 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4875
4876 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4877 compiled in such a way that the backtracking after subsequent failure was
4878 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4879 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4880
4881
4882 Version 2.03 02-Feb-99
4883 ----------------------
4884
4885 1. Fixed typo and small mistake in man page.
4886
4887 2. Added 4th condition (GPL supersedes if conflict) and created separate
4888 LICENCE file containing the conditions.
4889
4890 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4891 Perl, that is the internal \ allows the delimiter to be included in the
4892 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4893 the final delimiter, add \ to the end of the pattern (to test the error).
4894
4895 4. Added the convenience functions for extracting substrings after a successful
4896 match. Updated pcretest to make it able to test these functions.
4897
4898
4899 Version 2.02 14-Jan-99
4900 ----------------------
4901
4902 1. Initialized the working variables associated with each extraction so that
4903 their saving and restoring doesn't refer to uninitialized store.
4904
4905 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4906 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4907 fix the problem.
4908
4909 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4910 calls, and wasn't printing the correct value for compiling calls. Increased the
4911 default value of LOOPREPEAT, and the number of significant figures in the
4912 times.
4913
4914 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4915
4916 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4917 a building problem on Windows NT with a FAT file system.
4918
4919
4920 Version 2.01 21-Oct-98
4921 ----------------------
4922
4923 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4924 to character tables built by pcre_maketables() in the current locale. If NULL
4925 is passed, the default tables are used.
4926
4927
4928 Version 2.00 24-Sep-98
4929 ----------------------
4930
4931 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4932 it any more.
4933
4934 2. Allow quantification of (?>) groups, and make it work correctly.
4935
4936 3. The first character computation wasn't working for (?>) groups.
4937
4938 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4939 end of the subject) and add 5.005's \z, which really does match only at the
4940 very end of the subject.
4941
4942 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4943
4944 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4945 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4946 localized options. All options to pcre_study() were also removed.
4947
4948 7. Add other new features from 5.005:
4949
4950 $(?<= positive lookbehind
4951 $(?<! negative lookbehind
4952 (?imsx-imsx) added the unsetting capability
4953 such a setting is global if at outer level; local otherwise
4954 (?imsx-imsx:) non-capturing groups with option setting
4955 (?(cond)re|re) conditional pattern matching
4956
4957 A backreference to itself in a repeated group matches the previous
4958 captured string.
4959
4960 8. General tidying up of studying (both automatic and via "study")
4961 consequential on the addition of new assertions.
4962
4963 9. As in 5.005, unlimited repeated groups that could match an empty substring
4964 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4965 runtime if any iteration does actually match an empty substring.
4966
4967 10. Include the RunTest script in the distribution.
4968
4969 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4970 discrepancies, some of which were old and were also with respect to 5.004. They
4971 have now been fixed.
4972
4973
4974 Version 1.09 28-Apr-98
4975 ----------------------
4976
4977 1. A negated single character class followed by a quantifier with a minimum
4978 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4979 program crashes, or just wrong answers. This did not apply to negated classes
4980 containing more than one character, or to minima other than one.
4981
4982
4983 Version 1.08 27-Mar-98
4984 ----------------------
4985
4986 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4987
4988 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4989 latter must appear before anything that relies on it in the pattern.
4990
4991
4992 Version 1.07 16-Feb-98
4993 ----------------------
4994
4995 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4996 repeat of a potentially empty string).
4997
4998
4999 Version 1.06 23-Jan-98
5000 ----------------------
5001
5002 1. Added Markus Oberhumer's little patches for C++.
5003
5004 2. Literal strings longer than 255 characters were broken.
5005
5006
5007 Version 1.05 23-Dec-97
5008 ----------------------
5009
5010 1. Negated character classes containing more than one character were failing if
5011 PCRE_CASELESS was set at run time.
5012
5013
5014 Version 1.04 19-Dec-97
5015 ----------------------
5016
5017 1. Corrected the man page, where some "const" qualifiers had been omitted.
5018
5019 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5020 input syntax.
5021
5022 3. Fixed memory leak which occurred when a regex with back references was
5023 matched with an offsets vector that wasn't big enough. The temporary memory
5024 that is used in this case wasn't being freed if the match failed.
5025
5026 4. Tidied pcretest to ensure it frees memory that it gets.
5027
5028 5. Temporary memory was being obtained in the case where the passed offsets
5029 vector was exactly big enough.
5030
5031 6. Corrected definition of offsetof() from change 5 below.
5032
5033 7. I had screwed up change 6 below and broken the rules for the use of
5034 setjmp(). Now fixed.
5035
5036
5037 Version 1.03 18-Dec-97
5038 ----------------------
5039
5040 1. A erroneous regex with a missing opening parenthesis was correctly
5041 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5042 on some systems.
5043
5044 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5045 it was reported that one broken compiler failed on the former because "code" is
5046 also an independent variable.
5047
5048 3. The erroneous regex a[]b caused an array overrun reference.
5049
5050 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5051 fail on data ending with that character. (It was going on too far, and checking
5052 the next character, typically a binary zero.) This was specific to the
5053 optimized code for single-character negative classes.
5054
5055 5. Added a contributed patch from the TIN world which does the following:
5056
5057 + Add an undef for memmove, in case the the system defines a macro for it.
5058
5059 + Add a definition of offsetof(), in case there isn't one. (I don't know
5060 the reason behind this - offsetof() is part of the ANSI standard - but
5061 it does no harm).
5062
5063 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5064 most of the places where whitespace preceded '#'. I have given up and
5065 allowed the remaining 2 cases to be at the margin.
5066
5067 + Rename some variables in pcre to eliminate shadowing. This seems very
5068 pedantic, but does no harm, of course.
5069
5070 6. Moved the call to setjmp() into its own function, to get rid of warnings
5071 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5072
5073 7. Constructs such as \d{8,} were compiling into the equivalent of
5074 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5075 outcome, but in this particular case used more store than had been allocated,
5076 which caused the bug to be discovered because it threw up an internal error.
5077
5078 8. The debugging code in both pcre and pcretest for outputting the compiled
5079 form of a regex was going wrong in the case of back references followed by
5080 curly-bracketed repeats.
5081
5082
5083 Version 1.02 12-Dec-97
5084 ----------------------
5085
5086 1. Typos in pcre.3 and comments in the source fixed.
5087
5088 2. Applied a contributed patch to get rid of places where it used to remove
5089 'const' from variables, and fixed some signed/unsigned and uninitialized
5090 variable warnings.
5091
5092 3. Added the "runtest" target to Makefile.
5093
5094 4. Set default compiler flag to -O2 rather than just -O.
5095
5096
5097 Version 1.01 19-Nov-97
5098 ----------------------
5099
5100 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5101 like /([ab]*)*/, that is, for classes with more than one character in them.
5102
5103 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5104 as /((?>a*))*/ (a PCRE_EXTRA facility).
5105
5106
5107 Version 1.00 18-Nov-97
5108 ----------------------
5109
5110 1. Added compile-time macros to support systems such as SunOS4 which don't have
5111 memmove() or strerror() but have other things that can be used instead.
5112
5113 2. Arranged that "make clean" removes the executables.
5114
5115
5116 Version 0.99 27-Oct-97
5117 ----------------------
5118
5119 1. Fixed bug in code for optimizing classes with only one character. It was
5120 initializing a 32-byte map regardless, which could cause it to run off the end
5121 of the memory it had got.
5122
5123 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5124
5125
5126 Version 0.98 22-Oct-97
5127 ----------------------
5128
5129 1. Fixed bug in code for handling temporary memory usage when there are more
5130 back references than supplied space in the ovector. This could cause segfaults.
5131
5132
5133 Version 0.97 21-Oct-97
5134 ----------------------
5135
5136 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5137
5138 2. Optimized negated single characters not to use a bit map.
5139
5140 3. Brought error texts together as macro definitions; clarified some of them;
5141 fixed one that was wrong - it said "range out of order" when it meant "invalid
5142 escape sequence".
5143
5144 4. Changed some char * arguments to const char *.
5145
5146 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5147
5148 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5149 pcretest.
5150
5151
5152 Version 0.96 16-Oct-97
5153 ----------------------
5154
5155 1. Added a simple "pgrep" utility to the distribution.
5156
5157 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5158 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5159 where "ddd" means "one or more decimal digits".
5160
5161 3. Fixed serious bug. If a pattern had a back reference, but the call to
5162 pcre_exec() didn't supply a large enough ovector to record the related
5163 identifying subpattern, the match always failed. PCRE now remembers the number
5164 of the largest back reference, and gets some temporary memory in which to save
5165 the offsets during matching if necessary, in order to ensure that
5166 backreferences always work.
5167
5168 4. Increased the compatibility with Perl in a number of ways:
5169
5170 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5171 to request this handling. The option can be set at compile or exec time.
5172
5173 (b) $ matches before a terminating newline by default; an option
5174 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5175 mode). The option can be set at compile or exec time.
5176
5177 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5178 the same as Perl's. If the decimal number it represents is less than 10
5179 or there aren't that many previous left capturing parentheses, an octal
5180 escape is read. Inside a character class, it's always an octal escape,
5181 even if it is a single digit.
5182
5183 (d) An escaped but undefined alphabetic character is taken as a literal,
5184 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5185 escapes.
5186
5187 (e) {0} is now permitted. (The previous item is removed from the compiled
5188 pattern).
5189
5190 5. Changed all the names of code files so that the basic parts are no longer
5191 than 10 characters, and abolished the teeny "globals.c" file.
5192
5193 6. Changed the handling of character classes; they are now done with a 32-byte
5194 bit map always.
5195
5196 7. Added the -d and /D options to pcretest to make it possible to look at the
5197 internals of compilation without having to recompile pcre.
5198
5199
5200 Version 0.95 23-Sep-97
5201 ----------------------
5202
5203 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5204 \x20 at the start of a run of normal characters. These were being treated as
5205 real characters, instead of the source characters being re-checked.
5206
5207
5208 Version 0.94 18-Sep-97
5209 ----------------------
5210
5211 1. The functions are now thread-safe, with the caveat that the global variables
5212 containing pointers to malloc() and free() or alternative functions are the
5213 same for all threads.
5214
5215 2. Get pcre_study() to generate a bitmap of initial characters for non-
5216 anchored patterns when this is possible, and use it if passed to pcre_exec().
5217
5218
5219 Version 0.93 15-Sep-97
5220 ----------------------
5221
5222 1. /(b)|(:+)/ was computing an incorrect first character.
5223
5224 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5225 but not actually doing anything yet.
5226
5227 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5228 as Perl does (e.g. [-az] or [az-]).
5229
5230 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5231 all possible positions.
5232
5233 5. Split up into different modules to avoid including unneeded functions in a
5234 compiled binary. However, compile and exec are still in one module. The "study"
5235 function is split off.
5236
5237 6. The character tables are now in a separate module whose source is generated
5238 by an auxiliary program - but can then be edited by hand if required. There are
5239 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5240 toupper() in the code.
5241
5242 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5243 make them global. Abolish the function for setting them, as the caller can now
5244 set them directly.
5245
5246
5247 Version 0.92 11-Sep-97
5248 ----------------------
5249
5250 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5251 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5252
5253 2. Caseless matching was not working in character classes if the characters in
5254 the pattern were in upper case.
5255
5256 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5257
5258 4. Make PCRE_ANCHORED public and accept as a compile option.
5259
5260 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5261 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5262 pass them.
5263
5264 6. Give an error if bad option bits passed at compile or run time.
5265
5266 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5267 pcretest to cause it to pass that flag.
5268
5269 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5270 options, and the first character, if set.
5271
5272 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5273
5274
5275 Version 0.91 10-Sep-97
5276 ----------------------
5277
5278 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5279 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5280
5281 2. PCRE was looping on encountering an indefinitely repeated back reference to
5282 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5283 Perl does - treats the match as successful.
5284
5285 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5