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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1363 - (show annotations)
Tue Oct 1 16:54:40 2013 UTC (6 years ago) by ph10
File size: 252151 byte(s)
Refactored auto-possessification code.
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
92 Version 8.33 28-May-2013
93 ------------------------
94
95 1. Added 'U' to some constants that are compared to unsigned integers, to
96 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
97 variables that are added to signed variables, to ensure the result is
98 signed and can be negated.
99
100 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
101 CMake config files.
102
103 3. Revise the creation of config.h.generic so that all boolean macros are
104 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
105 overriding via -D on the command line possible.
106
107 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
108 to unsigned int is reported to make a quite noticeable speed difference in
109 a specific Windows environment. Testing on Linux did also appear to show
110 some benefit (and it is clearly not harmful). Also fixed the definition of
111 Xop which should be unsigned.
112
113 5. Related to (4), changing the definition of the intermediate variable cc
114 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
115 improvements.
116
117 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
118 unnecessary spaces.
119
120 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
121 and later.
122
123 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
124
125 9. Optimizing fast_forward_start_bits in JIT.
126
127 10. Adding support for callouts in JIT, and fixing some issues revealed
128 during this work. Namely:
129
130 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
131
132 (b) Minimum length was not checked before the matching is started.
133
134 11. The value of capture_last that is passed to callouts was incorrect in some
135 cases when there was a capture on one path that was subsequently abandoned
136 after a backtrack. Also, the capture_last value is now reset after a
137 recursion, since all captures are also reset in this case.
138
139 12. The interpreter no longer returns the "too many substrings" error in the
140 case when an overflowing capture is in a branch that is subsequently
141 abandoned after a backtrack.
142
143 13. In the pathological case when an offset vector of size 2 is used, pcretest
144 now prints out the matched string after a yield of 0 or 1.
145
146 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
147 Only supported by the JIT compiler at the moment.
148
149 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
150
151 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
152 offset of the starting point of the matching process, provided the offsets
153 vector is large enough.
154
155 17. The \A escape now records a lookbehind value of 1, though its execution
156 does not actually inspect the previous character. This is to ensure that,
157 in partial multi-segment matching, at least one character from the old
158 segment is retained when a new segment is processed. Otherwise, if there
159 are no lookbehinds in the pattern, \A might match incorrectly at the start
160 of a new segment.
161
162 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
163
164 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
165 gives some modest performance improvement in 8-bit mode.
166
167 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
168 be expressed in certain programming languages using Universal Character
169 Names.
170
171 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
172 which points out that "non characters" are not "characters that may not
173 appear in Unicode strings" but rather "characters that are reserved for
174 internal use and have only local meaning".
175
176 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
177 there was a conditional group that depended on an assertion, if the
178 assertion was false, the callout that immediately followed the alternation
179 in the condition was skipped when pcre_exec() was used for matching.
180
181 23. Allow an explicit callout to be inserted before an assertion that is the
182 condition for a conditional group, for compatibility with automatic
183 callouts, which always insert a callout at this point.
184
185 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
186 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
187
188 25. (*PRUNE) is now supported by the JIT compiler.
189
190 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
191
192 27. Fix the case where there are two or more SKIPs with arguments that may be
193 ignored.
194
195 28. (*SKIP) is now supported by the JIT compiler.
196
197 29. (*THEN) is now supported by the JIT compiler.
198
199 30. Update RunTest with additional test selector options.
200
201 31. The way PCRE handles backtracking verbs has been changed in two ways.
202
203 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
204 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
205 backtracking. In some cases this makes it more Perl-compatible, but Perl's
206 rather obscure rules do not always do the same thing.
207
208 (2) Previously, backtracking verbs were confined within assertions. This is
209 no longer the case for positive assertions, except for (*ACCEPT). Again,
210 this sometimes improves Perl compatibility, and sometimes does not.
211
212 32. A number of tests that were in test 2 because Perl did things differently
213 have been moved to test 1, because either Perl or PCRE has changed, and
214 these tests are now compatible.
215
216 32. Backtracking control verbs are now handled in the same way in JIT and
217 interpreter.
218
219 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
220 contained a forward subroutine reference caused a compile error.
221
222 34. Auto-detect and optimize limited repetitions in JIT.
223
224 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
225 blocking (*UTF) etc.
226
227 36. In the interpreter, maximizing pattern repetitions for characters and
228 character types now use tail recursion, which reduces stack usage.
229
230 37. The value of the max lookbehind was not correctly preserved if a compiled
231 and saved regex was reloaded on a host of different endianness.
232
233 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
234 of the compiled pattern block, expand the flags field from 16 to 32 bits
235 because it was almost full.
236
237 39. Try madvise first before posix_madvise.
238
239 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
240 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
241 original change locked out all matches of empty strings. This has been
242 changed so that one match of an empty string per line is recognized.
243 Subsequent searches on the same line (for colouring or for --only-matching,
244 for example) do not recognize empty strings.
245
246 41. Applied a user patch to fix a number of spelling mistakes in comments.
247
248 42. Data lines longer than 65536 caused pcretest to crash.
249
250 43. Clarified the data type for length and startoffset arguments for pcre_exec
251 and pcre_dfa_exec in the function-specific man pages, where they were
252 explicitly stated to be in bytes, never having been updated. I also added
253 some clarification to the pcreapi man page.
254
255 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
256 a segmentation fault.
257
258
259 Version 8.32 30-November-2012
260 -----------------------------
261
262 1. Improved JIT compiler optimizations for first character search and single
263 character iterators.
264
265 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
266 Patch by Daniel Richard G.
267
268 3. Single character iterator optimizations in the JIT compiler.
269
270 4. Improved JIT compiler optimizations for character ranges.
271
272 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
273 Reported by Giuseppe D'Angelo.
274
275 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
276 of a line, was being set incorrectly in cases where .* appeared inside
277 atomic brackets at the start of a pattern, or where there was a subsequent
278 *PRUNE or *SKIP.
279
280 7. Improved instruction cache flush for POWER/PowerPC.
281 Patch by Daniel Richard G.
282
283 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
284 grep:
285
286 (a) There is now no limit to the number of patterns to be matched.
287
288 (b) An error is given if a pattern is too long.
289
290 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
291 are now supported.
292
293 (d) --exclude-from and --include-from (multiple use) have been added.
294
295 (e) Exclusions and inclusions now apply to all files and directories, not
296 just to those obtained from scanning a directory recursively.
297
298 (f) Multiple uses of -f and --file-list are now supported.
299
300 (g) In a Windows environment, the default for -d has been changed from
301 "read" (the GNU grep default) to "skip", because otherwise the presence
302 of a directory in the file list provokes an error.
303
304 (h) The documentation has been revised and clarified in places.
305
306 9. Improve the matching speed of capturing brackets.
307
308 10. Changed the meaning of \X so that it now matches a Unicode extended
309 grapheme cluster.
310
311 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
312 out POSIX threads when JIT support is configured.
313
314 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
315
316 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
317 structure could go wrong in environments where size_t is not the same size
318 as int.
319
320 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
321 set.
322
323 15. The EBCDIC support had decayed; later updates to the code had included
324 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
325 general tidy up of EBCDIC-related issues, and the documentation was also
326 not quite right. There is now a test that can be run on ASCII systems to
327 check some of the EBCDIC-related things (but is it not a full test).
328
329 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
330 in a small tidy to the code.
331
332 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
333
334 18. If the --only-matching (-o) option in pcregrep is specified multiple
335 times, each one causes appropriate output. For example, -o1 -o2 outputs the
336 substrings matched by the 1st and 2nd capturing parentheses. A separating
337 string can be specified by --om-separator (default empty).
338
339 19. Improving the first n character searches.
340
341 20. Turn case lists for horizontal and vertical white space into macros so that
342 they are defined only once.
343
344 21. This set of changes together give more compatible Unicode case-folding
345 behaviour for characters that have more than one other case when UCP
346 support is available.
347
348 (a) The Unicode property table now has offsets into a new table of sets of
349 three or more characters that are case-equivalent. The MultiStage2.py
350 script that generates these tables (the pcre_ucd.c file) now scans
351 CaseFolding.txt instead of UnicodeData.txt for character case
352 information.
353
354 (b) The code for adding characters or ranges of characters to a character
355 class has been abstracted into a generalized function that also handles
356 case-independence. In UTF-mode with UCP support, this uses the new data
357 to handle characters with more than one other case.
358
359 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
360 whose other case is greater than 256 are now correctly matched
361 caselessly. Previously, the high codepoint matched the low one, but not
362 vice versa.
363
364 (d) The processing of \h, \H, \v, and \ in character classes now makes use
365 of the new class addition function, using character lists defined as
366 macros alongside the case definitions of 20 above.
367
368 (e) Caseless back references now work with characters that have more than
369 one other case.
370
371 (f) General caseless matching of characters with more than one other case
372 is supported.
373
374 22. Unicode character properties were updated from Unicode 6.2.0
375
376 23. Improved CMake support under Windows. Patch by Daniel Richard G.
377
378 24. Add support for 32-bit character strings, and UTF-32
379
380 25. Major JIT compiler update (code refactoring and bugfixing).
381 Experimental Sparc 32 support is added.
382
383 26. Applied a modified version of Daniel Richard G's patch to create
384 pcre.h.generic and config.h.generic by "make" instead of in the
385 PrepareRelease script.
386
387 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
388 pcre_compile.c when checking for a zero character.
389
390 28. Introducing a native interface for JIT. Through this interface, the compiled
391 machine code can be directly executed. The purpose of this interface is to
392 provide fast pattern matching, so several sanity checks are not performed.
393 However, feature tests are still performed. The new interface provides
394 1.4x speedup compared to the old one.
395
396 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
397 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
398 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
399
400 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
401 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
402 the "old" RFC 2279). Instead, it ended up passing a negative length to
403 pcre_exec().
404
405 31. Add support for GCC's visibility feature to hide internal functions.
406
407 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
408 "unknown -C option" after outputting 0 or 1.
409
410 33. There is now support for generating a code coverage report for the test
411 suite in environments where gcc is the compiler and lcov is installed. This
412 is mainly for the benefit of the developers.
413
414 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
415 unaddressable using valgrind annotations, allowing valgrind to detect
416 invalid memory accesses. This is mainly for the benefit of the developers.
417
418 25. (*UTF) can now be used to start a pattern in any of the three libraries.
419
420 26. Give configure error if --enable-cpp but no C++ compiler found.
421
422
423 Version 8.31 06-July-2012
424 -------------------------
425
426 1. Fixing a wrong JIT test case and some compiler warnings.
427
428 2. Removed a bashism from the RunTest script.
429
430 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
431 to unsigned type, result still unsigned" that was given by an MS compiler
432 on encountering the code "-sizeof(xxx)".
433
434 4. Partial matching support is added to the JIT compiler.
435
436 5. Fixed several bugs concerned with partial matching of items that consist
437 of more than one character:
438
439 (a) /^(..)\1/ did not partially match "aba" because checking references was
440 done on an "all or nothing" basis. This also applied to repeated
441 references.
442
443 (b) \R did not give a hard partial match if \r was found at the end of the
444 subject.
445
446 (c) \X did not give a hard partial match after matching one or more
447 characters at the end of the subject.
448
449 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
450 a partial match for the string "\r".
451
452 (e) When newline was set to CRLF, the metacharacter "." did not recognize
453 a partial match for a CR character at the end of the subject string.
454
455 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
456 running pcretest, the text "(JIT)" added to the output whenever JIT is
457 actually used to run the match.
458
459 7. Individual JIT compile options can be set in pcretest by following -s+[+]
460 or /S+[+] with a digit between 1 and 7.
461
462 8. OP_NOT now supports any UTF character not just single-byte ones.
463
464 9. (*MARK) control verb is now supported by the JIT compiler.
465
466 10. The command "./RunTest list" lists the available tests without actually
467 running any of them. (Because I keep forgetting what they all are.)
468
469 11. Add PCRE_INFO_MAXLOOKBEHIND.
470
471 12. Applied a (slightly modified) user-supplied patch that improves performance
472 when the heap is used for recursion (compiled with --disable-stack-for-
473 recursion). Instead of malloc and free for each heap frame each time a
474 logical recursion happens, frames are retained on a chain and re-used where
475 possible. This sometimes gives as much as 30% improvement.
476
477 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
478 call.
479
480 14. As documented, (*COMMIT) is now confined to within a positive assertion.
481
482 15. It is now possible to link pcretest with libedit as an alternative to
483 libreadline.
484
485 16. (*COMMIT) control verb is now supported by the JIT compiler.
486
487 17. The Unicode data tables have been updated to Unicode 6.1.0.
488
489 18. Added --file-list option to pcregrep.
490
491 19. Added binary file support to pcregrep, including the -a, --binary-files,
492 -I, and --text options.
493
494 20. The madvise function is renamed for posix_madvise for QNX compatibility
495 reasons. Fixed by Giuseppe D'Angelo.
496
497 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
498
499 22. Changed the option for creating symbolic links for 16-bit man pages from
500 -s to -sf so that re-installing does not cause issues.
501
502 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
503
504 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
505 matches in certain environments (the workspace was not being correctly
506 retained). Also added to pcre_dfa_exec() a simple plausibility check on
507 some of the workspace data at the beginning of a restart.
508
509 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
510 was not doing so when it should - probably a typo introduced by SVN 528
511 (change 8.10/14).
512
513 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
514 \w+ when the character tables indicated that \x{c4} was a word character.
515 There were several related cases, all because the tests for doing a table
516 lookup were testing for characters less than 127 instead of 255.
517
518 27. If a pattern contains capturing parentheses that are not used in a match,
519 their slots in the ovector are set to -1. For those that are higher than
520 any matched groups, this happens at the end of processing. In the case when
521 there were back references that the ovector was too small to contain
522 (causing temporary malloc'd memory to be used during matching), and the
523 highest capturing number was not used, memory off the end of the ovector
524 was incorrectly being set to -1. (It was using the size of the temporary
525 memory instead of the true size.)
526
527 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
528 ovector size, it uses memory at the end of the block that it has got.
529
530 29. Check for an overlong MARK name and give an error at compile time. The
531 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
532
533 30. JIT compiler update.
534
535 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
536 Rill for the patch.
537
538 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
539
540 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
541
542 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
543 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
544 was enabled.
545
546 35. Improve JIT code generation for greedy plus quantifier.
547
548 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
549 1 to "aa" instead of to an empty string. The bug affected repeated groups
550 that could potentially match an empty string.
551
552 37. Optimizing single character iterators in JIT.
553
554 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
555 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
556 codepoints that are too big for the mode are faulted, and in a UTF mode,
557 disallowed codepoints are also faulted.
558
559 39. If PCRE was compiled with UTF support, in three places in the DFA
560 matcher there was code that should only have been obeyed in UTF mode, but
561 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
562 processing when bytes with values greater than 127 were present. In 16-bit
563 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
564 both cases the values are those that cannot be the first data item in a UTF
565 character. The three items that might have provoked this were recursions,
566 possessively repeated groups, and atomic groups.
567
568 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
569 and pcregrep, because some OS require shared objects to be explicitly
570 passed to ld, causing the link step to fail if they are not.
571
572 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
573 mode, patterns that started with \h* or \R* might be incorrectly matched.
574
575
576 Version 8.30 04-February-2012
577 -----------------------------
578
579 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
580 name is defined in ctype.h.
581
582 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
583 only in quite long subpatterns.
584
585 3. Removed the function pcre_info(), which has been obsolete and deprecated
586 since it was replaced by pcre_fullinfo() in February 2000.
587
588 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
589 match a (*MARK), and the match failed at the start of the subject, a
590 reference to memory before the start of the subject could occur. This bug
591 was introduced by fix 17 of release 8.21.
592
593 5. A reference to an unset group with zero minimum repetition was giving
594 totally wrong answers (in non-JavaScript-compatibility mode). For example,
595 /(another)?(\1?)test/ matched against "hello world test". This bug was
596 introduced in release 8.13.
597
598 6. Add support for 16-bit character strings (a large amount of work involving
599 many changes and refactorings).
600
601 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
602 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
603 from a file.
604
605 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
606 rounding is not applied in this particular case).
607
608 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
609 if they appear, or are escaped, in patterns.
610
611 10. Get rid of a number of -Wunused-but-set-variable warnings.
612
613 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
614 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
615 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
616 also returns the mark "x". This bug applied to capturing parentheses,
617 non-capturing parentheses, and atomic parentheses. It also applied to some
618 assertions.
619
620 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
621 information out of configure.ac instead of relying on pcre.h.generic, which
622 is not stored in the repository.
623
624 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
625 -lreadline.
626
627 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
628
629 15. Applied Graycode's patch to put the top-level frame on the stack rather
630 than the heap when not using the stack for recursion. This gives a
631 performance improvement in many cases when recursion is not deep.
632
633 16. Experimental code added to "pcretest -C" to output the stack frame size.
634
635
636 Version 8.21 12-Dec-2011
637 ------------------------
638
639 1. Updating the JIT compiler.
640
641 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
642 are added as well.
643
644 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
645 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
646 calling _pcre_jit_exec. Some extra comments are added.
647
648 4. (*MARK) settings inside atomic groups that do not contain any capturing
649 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
650 was introduced by change 18 for 8.20.
651
652 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
653 ECMA-262 standard.
654
655 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
656 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
657 This bug was probably introduced by change 9 of 8.13.
658
659 7. While fixing 6 above, I noticed that a number of other items were being
660 incorrectly rejected as "not fixed length". This arose partly because newer
661 opcodes had not been added to the fixed-length checking code. I have (a)
662 corrected the bug and added tests for these items, and (b) arranged for an
663 error to occur if an unknown opcode is encountered while checking for fixed
664 length instead of just assuming "not fixed length". The items that were
665 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
666 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
667 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
668
669 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
670 being incorrectly compiled and would have given unpredicatble results.
671
672 9. A possessively repeated subpattern with minimum repeat count greater than
673 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
674 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
675 the first (A) could occur when it should not.
676
677 10. Add a cast and remove a redundant test from the code.
678
679 11. JIT should use pcre_malloc/pcre_free for allocation.
680
681 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
682 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
683 is anything other than /usr, -L is still shown).
684
685 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
686
687 14. Perl does not support \N without a following name in a [] class; PCRE now
688 also gives an error.
689
690 15. If a forward reference was repeated with an upper limit of around 2000,
691 it caused the error "internal error: overran compiling workspace". The
692 maximum number of forward references (including repeats) was limited by the
693 internal workspace, and dependent on the LINK_SIZE. The code has been
694 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
695 the default depends on LINK_SIZE. There is a new upper limit (for safety)
696 of around 200,000 forward references. While doing this, I also speeded up
697 the filling in of repeated forward references.
698
699 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
700 incorrectly expecting the subject to contain another "a" after the start.
701
702 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
703 in the match, the SKIP should be ignored. This was not happening; instead
704 the SKIP was being treated as NOMATCH. For patterns such as
705 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
706 tested.
707
708 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
709 now much more compatible with Perl, in particular in cases where the result
710 is a non-match for a non-anchored pattern. For example, if
711 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
712 "m", where previously it did not return a name. A side effect of this
713 change is that for partial matches, the last encountered mark name is
714 returned, as for non matches. A number of tests that were previously not
715 Perl-compatible have been moved into the Perl-compatible test files. The
716 refactoring has had the pleasing side effect of removing one argument from
717 the match() function, thus reducing its stack requirements.
718
719 19. If the /S+ option was used in pcretest to study a pattern using JIT,
720 subsequent uses of /S (without +) incorrectly behaved like /S+.
721
722 21. Retrieve executable code size support for the JIT compiler and fixing
723 some warnings.
724
725 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
726 not work when the shorter character appeared right at the end of the
727 subject string.
728
729 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
730 systems.
731
732 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
733 output it when the /M option is used in pcretest.
734
735 25. The CheckMan script was not being included in the distribution. Also, added
736 an explicit "perl" to run Perl scripts from the PrepareRelease script
737 because this is reportedly needed in Windows.
738
739 26. If study data was being save in a file and studying had not found a set of
740 "starts with" bytes for the pattern, the data written to the file (though
741 never used) was taken from uninitialized memory and so caused valgrind to
742 complain.
743
744 27. Updated RunTest.bat as provided by Sheri Pierce.
745
746 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
747
748 29. Computation of memory usage for the table of capturing group names was
749 giving an unnecessarily large value.
750
751
752 Version 8.20 21-Oct-2011
753 ------------------------
754
755 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
756 a POSIX class. After further experiments with Perl, which convinced me that
757 Perl has bugs and confusions, a closing square bracket is no longer allowed
758 in a POSIX name. This bug also affected patterns with classes that started
759 with full stops.
760
761 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
762 captured substring, but while checking the failing first alternative,
763 substring 1 is temporarily captured. If the output vector supplied to
764 pcre_exec() was not big enough for this capture, the yield of the function
765 was still zero ("insufficient space for captured substrings"). This cannot
766 be totally fixed without adding another stack variable, which seems a lot
767 of expense for a edge case. However, I have improved the situation in cases
768 such as /(a)(b)x|abc/ matched against "abc", where the return code
769 indicates that fewer than the maximum number of slots in the ovector have
770 been set.
771
772 3. Related to (2) above: when there are more back references in a pattern than
773 slots in the output vector, pcre_exec() uses temporary memory during
774 matching, and copies in the captures as far as possible afterwards. It was
775 using the entire output vector, but this conflicts with the specification
776 that only 2/3 is used for passing back captured substrings. Now it uses
777 only the first 2/3, for compatibility. This is, of course, another edge
778 case.
779
780 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
781 main code base, and can be used by building with --enable-jit. When this is
782 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
783 runtime --no-jit option is given.
784
785 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
786 ovector, the return from the function was zero, implying that there were
787 other matches that did not fit. The correct "exactly full" value is now
788 returned.
789
790 6. If a subpattern that was called recursively or as a subroutine contained
791 (*PRUNE) or any other control that caused it to give a non-standard return,
792 invalid errors such as "Error -26 (nested recursion at the same subject
793 position)" or even infinite loops could occur.
794
795 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
796 computing the minimum length on reaching *ACCEPT, and so ended up with the
797 wrong value of 1 rather than 0. Further investigation indicates that
798 computing a minimum subject length in the presence of *ACCEPT is difficult
799 (think back references, subroutine calls), and so I have changed the code
800 so that no minimum is registered for a pattern that contains *ACCEPT.
801
802 8. If (*THEN) was present in the first (true) branch of a conditional group,
803 it was not handled as intended. [But see 16 below.]
804
805 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
806 Sheri Pierce.
807
808 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
809 the first byte in a match must be "a".
810
811 11. Change 17 for 8.13 increased the recursion depth for patterns like
812 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
813 pattern contains any instances of (*THEN). If it does not, the old
814 optimizations are restored. It would be nice to do this on a per-group
815 basis, but at the moment that is not feasible.
816
817 12. In some environments, the output of pcretest -C is CRLF terminated. This
818 broke RunTest's code that checks for the link size. A single white space
819 character after the value is now allowed for.
820
821 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
822 For "fr", it uses the Windows-specific input and output files.
823
824 14. If (*THEN) appeared in a group that was called recursively or as a
825 subroutine, it did not work as intended. [But see next item.]
826
827 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
828 pattern fragments (but not containing any | characters). If A and B are
829 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
830 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
831 D. In other words, Perl considers parentheses that do not contain any |
832 characters to be part of a surrounding alternative, whereas PCRE was
833 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
834 differently. PCRE now behaves in the same way as Perl, except in the case
835 of subroutine/recursion calls such as (?1) which have in any case always
836 been different (but PCRE had them first :-).
837
838 16. Related to 15 above: Perl does not treat the | in a conditional group as
839 creating alternatives. Such a group is treated in the same way as an
840 ordinary group without any | characters when processing (*THEN). PCRE has
841 been changed to match Perl's behaviour.
842
843 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
844 RunGrepTest script failed.
845
846 18. Change 22 for version 13 caused atomic groups to use more stack. This is
847 inevitable for groups that contain captures, but it can lead to a lot of
848 stack use in large patterns. The old behaviour has been restored for atomic
849 groups that do not contain any capturing parentheses.
850
851 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
852 suppress the check for a minimum subject length at run time. (If it was
853 given to pcre_exec() or pcre_dfa_exec() it did work.)
854
855 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
856 fail to work when decoding hex characters in data strings in EBCDIC
857 environments.
858
859 21. It appears that in at least one Mac OS environment, the isxdigit() function
860 is implemented as a macro that evaluates to its argument more than once,
861 contravening the C 90 Standard (I haven't checked a later standard). There
862 was an instance in pcretest which caused it to go wrong when processing
863 \x{...} escapes in subject strings. The has been rewritten to avoid using
864 things like p++ in the argument of isxdigit().
865
866
867 Version 8.13 16-Aug-2011
868 ------------------------
869
870 1. The Unicode data tables have been updated to Unicode 6.0.0.
871
872 2. Two minor typos in pcre_internal.h have been fixed.
873
874 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
875 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
876 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
877
878 4. There were a number of related bugs in the code for matching backrefences
879 caselessly in UTF-8 mode when codes for the characters concerned were
880 different numbers of bytes. For example, U+023A and U+2C65 are an upper
881 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
882 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
883 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
884 2-byte code at the end of the subject (it thought there wasn't enough data
885 left).
886
887 5. Comprehensive information about what went wrong is now returned by
888 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
889 as the output vector has at least 2 elements. The offset of the start of
890 the failing character and a reason code are placed in the vector.
891
892 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
893 now returned is for the first byte of the failing character, instead of the
894 last byte inspected. This is an incompatible change, but I hope it is small
895 enough not to be a problem. It makes the returned offset consistent with
896 pcre_exec() and pcre_dfa_exec().
897
898 7. pcretest now gives a text phrase as well as the error number when
899 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
900 failure, the offset and reason code are output.
901
902 8. When \R was used with a maximizing quantifier it failed to skip backwards
903 over a \r\n pair if the subsequent match failed. Instead, it just skipped
904 back over a single character (\n). This seems wrong (because it treated the
905 two characters as a single entity when going forwards), conflicts with the
906 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
907 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
908 has been changed.
909
910 9. Some internal refactoring has changed the processing so that the handling
911 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
912 time (the PCRE_DOTALL option was changed this way some time ago: version
913 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
914 which was always a bit of a fudge. It also means that there is one less
915 argument for the match() function, which reduces its stack requirements
916 slightly. This change also fixes an incompatibility with Perl: the pattern
917 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
918
919 10. More internal refactoring has drastically reduced the number of recursive
920 calls to match() for possessively repeated groups such as (abc)++ when
921 using pcre_exec().
922
923 11. While implementing 10, a number of bugs in the handling of groups were
924 discovered and fixed:
925
926 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
927 (a|)*(?1) gave a compile-time internal error.
928 ((a|)+)+ did not notice that the outer group could match an empty string.
929 (^a|^)+ was not marked as anchored.
930 (.*a|.*)+ was not marked as matching at start or after a newline.
931
932 12. Yet more internal refactoring has removed another argument from the match()
933 function. Special calls to this function are now indicated by setting a
934 value in a variable in the "match data" data block.
935
936 13. Be more explicit in pcre_study() instead of relying on "default" for
937 opcodes that mean there is no starting character; this means that when new
938 ones are added and accidentally left out of pcre_study(), testing should
939 pick them up.
940
941 14. The -s option of pcretest has been documented for ages as being an old
942 synonym of -m (show memory usage). I have changed it to mean "force study
943 for every regex", that is, assume /S for every regex. This is similar to -i
944 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
945 using it. It makes it easier to run collections of tests with and without
946 study enabled, and thereby test pcre_study() more easily. All the standard
947 tests are now run with and without -s (but some patterns can be marked as
948 "never study" - see 20 below).
949
950 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
951 restoration of the capturing data to the outer values was not happening
952 correctly.
953
954 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
955 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
956 pattern had matched an empty string, and so incorrectly returned a no
957 match.
958
959 17. There was optimizing code for the last branch of non-capturing parentheses,
960 and also for the obeyed branch of a conditional subexpression, which used
961 tail recursion to cut down on stack usage. Unfortunately, now that there is
962 the possibility of (*THEN) occurring in these branches, tail recursion is
963 no longer possible because the return has to be checked for (*THEN). These
964 two optimizations have therefore been removed. [But see 8.20/11 above.]
965
966 18. If a pattern containing \R was studied, it was assumed that \R always
967 matched two bytes, thus causing the minimum subject length to be
968 incorrectly computed because \R can also match just one byte.
969
970 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
971 was incorrectly computed.
972
973 20. If /S is present twice on a test pattern in pcretest input, it now
974 *disables* studying, thereby overriding the use of -s on the command line
975 (see 14 above). This is necessary for one or two tests to keep the output
976 identical in both cases.
977
978 21. When (*ACCEPT) was used in an assertion that matched an empty string and
979 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
980
981 22. When an atomic group that contained a capturing parenthesis was
982 successfully matched, but the branch in which it appeared failed, the
983 capturing was not being forgotten if a higher numbered group was later
984 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
985 group 1 to "a", when in fact it should be unset. This applied to multi-
986 branched capturing and non-capturing groups, repeated or not, and also to
987 positive assertions (capturing in negative assertions does not happen
988 in PCRE) and also to nested atomic groups.
989
990 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
991 subject after a captured substring, to make it easier to tell which of a
992 number of identical substrings has been captured.
993
994 24. The way atomic groups are processed by pcre_exec() has been changed so that
995 if they are repeated, backtracking one repetition now resets captured
996 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
997 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
998 "aaa". Previously, it would have been "a". As part of this code
999 refactoring, the way recursive calls are handled has also been changed.
1000
1001 25. If an assertion condition captured any substrings, they were not passed
1002 back unless some other capturing happened later. For example, if
1003 (?(?=(a))a) was matched against "a", no capturing was returned.
1004
1005 26. When studying a pattern that contained subroutine calls or assertions,
1006 the code for finding the minimum length of a possible match was handling
1007 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1008 group 1 called group 2 while simultaneously a separate group 2 called group
1009 1). A stack overflow occurred in this case. I have fixed this by limiting
1010 the recursion depth to 10.
1011
1012 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1013 Fortmann. This supports explicit test numbers on the command line, and has
1014 argument validation and error reporting.
1015
1016 28. An instance of \X with an unlimited repeat could fail if at any point the
1017 first character it looked at was a mark character.
1018
1019 29. Some minor code refactoring concerning Unicode properties and scripts
1020 should reduce the stack requirement of match() slightly.
1021
1022 30. Added the '=' option to pcretest to check the setting of unused capturing
1023 slots at the end of the pattern, which are documented as being -1, but are
1024 not included in the return count.
1025
1026 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1027 compiled something random. Now it gives a compile-time error (as does
1028 Perl).
1029
1030 32. A *MARK encountered during the processing of a positive assertion is now
1031 recorded and passed back (compatible with Perl).
1032
1033 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1034 had alternative anchored branches, the search for a second match in a line
1035 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1036 matched the line "0102" twice. The same bug affected patterns that started
1037 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1038 twice.
1039
1040 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1041 does, and because of capturing effects, quantifying parenthesized
1042 assertions may at times be useful. Quantifiers are now allowed for
1043 parenthesized assertions.
1044
1045 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1046
1047 36. \g was being checked for fancy things in a character class, when it should
1048 just be a literal "g".
1049
1050 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1051 appearance of a nested POSIX class supersedes an apparent external class.
1052 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1053 unescaped square brackets may also appear as part of class names. For
1054 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1055 more like Perl. (But see 8.20/1 above.)
1056
1057 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1058 was because it thought it was \N{name}, which is not supported).
1059
1060 39. Add minix to OS list not supporting the -S option in pcretest.
1061
1062 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1063 cannot analyze patterns in sufficient detail to catch mutual recursions
1064 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1065 subgroup is called recursively as a subpattern for a second time at the
1066 same position in the subject string. In previous releases this might have
1067 been caught by the recursion limit, or it might have run out of stack.
1068
1069 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1070 happen only once. PCRE was, however incorrectly giving a compile time error
1071 "recursive call could loop indefinitely" because it cannot analyze the
1072 pattern in sufficient detail. The compile time test no longer happens when
1073 PCRE is compiling a conditional subpattern, but actual runaway loops are
1074 now caught at runtime (see 40 above).
1075
1076 42. It seems that Perl allows any characters other than a closing parenthesis
1077 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1078 has been changed to be the same.
1079
1080 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1081 as not to get warnings when autogen.sh is called. Also changed
1082 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1083
1084 44. To help people who use pcregrep to scan files containing exceedingly long
1085 lines, the following changes have been made:
1086
1087 (a) The default value of the buffer size parameter has been increased from
1088 8K to 20K. (The actual buffer used is three times this size.)
1089
1090 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1091 PCRE is built.
1092
1093 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1094 to be set at run time.
1095
1096 (d) Numerical values in pcregrep options can be followed by K or M, for
1097 example --buffer-size=50K.
1098
1099 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1100 given and the return code is set to 2.
1101
1102 45. Add a pointer to the latest mark to the callout data block.
1103
1104 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1105 partial match of an empty string instead of no match. This was specific to
1106 the use of ".".
1107
1108 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1109 complete match instead of a partial match. This bug was dependent on both
1110 the PCRE_UTF8 and PCRE_DOTALL options being set.
1111
1112 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1113 starting byte set, because \b was not being ignored.
1114
1115
1116 Version 8.12 15-Jan-2011
1117 ------------------------
1118
1119 1. Fixed some typos in the markup of the man pages, and wrote a script that
1120 checks for such things as part of the documentation building process.
1121
1122 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1123 --match-limit and --recursion-limit options (added for 8.11). In
1124 particular, this made one of the standard tests fail. (The integer value
1125 went into the wrong half of a long int.)
1126
1127 3. If the --colour option was given to pcregrep with -v (invert match), it
1128 did strange things, either producing crazy output, or crashing. It should,
1129 of course, ignore a request for colour when reporting lines that do not
1130 match.
1131
1132 4. Another pcregrep bug caused similar problems if --colour was specified with
1133 -M (multiline) and the pattern match finished with a line ending.
1134
1135 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1136 matched in multiline mode, the following line was shown as part of the
1137 match. This seems wrong, so I have changed it.
1138
1139 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1140 the check for further matches in the same line (so they could be coloured)
1141 to overrun the end of the current line. If another match was found, it was
1142 incorrectly shown (and then shown again when found in the next line).
1143
1144 7. If pcregrep was compiled under Windows, there was a reference to the
1145 function pcregrep_exit() before it was defined. I am assuming this was
1146 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1147 reported by a user. I've moved the definition above the reference.
1148
1149
1150 Version 8.11 10-Dec-2010
1151 ------------------------
1152
1153 1. (*THEN) was not working properly if there were untried alternatives prior
1154 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1155 backtracked to try for "b" instead of moving to the next alternative branch
1156 at the same level (in this case, to look for "c"). The Perl documentation
1157 is clear that when (*THEN) is backtracked onto, it goes to the "next
1158 alternative in the innermost enclosing group".
1159
1160 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1161 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1162 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1163 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1164 (*THEN).
1165
1166 3. If \s appeared in a character class, it removed the VT character from
1167 the class, even if it had been included by some previous item, for example
1168 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1169 of \s, but is part of the POSIX "space" class.)
1170
1171 4. A partial match never returns an empty string (because you can always
1172 match an empty string at the end of the subject); however the checking for
1173 an empty string was starting at the "start of match" point. This has been
1174 changed to the "earliest inspected character" point, because the returned
1175 data for a partial match starts at this character. This means that, for
1176 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1177 (previously it gave "no match").
1178
1179 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1180 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1181 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1182 has an implication that the given string is incomplete (because a partial
1183 match is preferred over a full match). For this reason, these items now
1184 give a partial match in this situation. [Aside: previously, the one case
1185 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1186 match rather than a full match, which was wrong by the old rules, but is
1187 now correct.]
1188
1189 6. There was a bug in the handling of #-introduced comments, recognized when
1190 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1191 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1192 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1193 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1194 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1195 places in pcre_compile().
1196
1197 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1198 comments when looking ahead for named forward references to subpatterns,
1199 the only newline sequence it recognized was NL. It now handles newlines
1200 according to the set newline convention.
1201
1202 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1203 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1204 cater for a lack of strerror(). These oversights have been fixed.
1205
1206 9. Added --match-limit and --recursion-limit to pcregrep.
1207
1208 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1209
1210 11. When the -o option was used, pcregrep was setting a return code of 1, even
1211 when matches were found, and --line-buffered was not being honoured.
1212
1213 12. Added an optional parentheses number to the -o and --only-matching options
1214 of pcregrep.
1215
1216 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1217 can match an empty string. The code to do it in pcretest and pcredemo
1218 needed fixing:
1219
1220 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1221 only one byte after an empty string match just before CRLF (this case
1222 just got forgotten; "any" and "anycrlf" were OK).
1223
1224 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1225 mode when an empty string match preceded an ASCII character followed by
1226 a non-ASCII character. (The code for advancing by one character rather
1227 than one byte was nonsense.)
1228
1229 (c) The pcredemo.c sample program did not have any code at all to handle
1230 the cases when CRLF is a valid newline sequence.
1231
1232 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1233 as a starting offset was within the subject string. There is now a new
1234 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1235 negative or greater than the length of the string. In order to test this,
1236 pcretest is extended to allow the setting of negative starting offsets.
1237
1238 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1239 starting offset points to the beginning of a UTF-8 character was
1240 unnecessarily clumsy. I tidied it up.
1241
1242 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1243 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1244
1245 17. Nobody had reported that the --include_dir option, which was added in
1246 release 7.7 should have been called --include-dir (hyphen, not underscore)
1247 for compatibility with GNU grep. I have changed it to --include-dir, but
1248 left --include_dir as an undocumented synonym, and the same for
1249 --exclude-dir, though that is not available in GNU grep, at least as of
1250 release 2.5.4.
1251
1252 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1253 characters from a string of bytes) have been redefined so as not to use
1254 loops, in order to improve performance in some environments. At the same
1255 time, I abstracted some of the common code into auxiliary macros to save
1256 repetition (this should not affect the compiled code).
1257
1258 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1259 compile-time error is now given if \c is not followed by an ASCII
1260 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1261 different, and any byte value is allowed.)
1262
1263 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1264 START_OPTIMIZE option, which is now allowed at compile time - but just
1265 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1266 to pcregrep and other applications that have no direct access to PCRE
1267 options. The new /Y option in pcretest sets this option when calling
1268 pcre_compile().
1269
1270 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1271 back references. Groups containing recursive back references were forced to
1272 be atomic by that change, but in the case of named groups, the amount of
1273 memory required was incorrectly computed, leading to "Failed: internal
1274 error: code overflow". This has been fixed.
1275
1276 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1277 pcretest.c, to avoid build problems in some Borland environments.
1278
1279
1280 Version 8.10 25-Jun-2010
1281 ------------------------
1282
1283 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1284 THEN.
1285
1286 2. (*ACCEPT) was not working when inside an atomic group.
1287
1288 3. Inside a character class, \B is treated as a literal by default, but
1289 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1290 causes the error). The code is unchanged, but I tidied the documentation.
1291
1292 4. Inside a character class, PCRE always treated \R and \X as literals,
1293 whereas Perl faults them if its -w option is set. I have changed PCRE so
1294 that it faults them when PCRE_EXTRA is set.
1295
1296 5. Added support for \N, which always matches any character other than
1297 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1298
1299 6. When compiling pcregrep with newer versions of gcc which may have
1300 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1301 declared with attribute warn_unused_result" were given. Just casting the
1302 result to (void) does not stop the warnings; a more elaborate fudge is
1303 needed. I've used a macro to implement this.
1304
1305 7. Minor change to pcretest.c to avoid a compiler warning.
1306
1307 8. Added four artifical Unicode properties to help with an option to make
1308 \s etc use properties (see next item). The new properties are: Xan
1309 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1310
1311 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1312 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1313 this option. Modified pcretest to add /W to test this facility. Added
1314 REG_UCP to make it available via the POSIX interface.
1315
1316 10. Added --line-buffered to pcregrep.
1317
1318 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1319 studied, and the match started with a letter with a code point greater than
1320 127 whose first byte was different to the first byte of the other case of
1321 the letter, the other case of this starting letter was not recognized
1322 (#976).
1323
1324 12. If a pattern that was studied started with a repeated Unicode property
1325 test, for example, \p{Nd}+, there was the theoretical possibility of
1326 setting up an incorrect bitmap of starting bytes, but fortunately it could
1327 not have actually happened in practice until change 8 above was made (it
1328 added property types that matched character-matching opcodes).
1329
1330 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1331 possible starting bytes for non-anchored patterns.
1332
1333 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1334 \R, and also a number of cases that involve Unicode properties, both
1335 explicit and implicit when PCRE_UCP is set.
1336
1337 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1338 input, it could crash or give wrong results if characters with values
1339 greater than 0xc0 were present in the subject string. (Detail: it assumed
1340 UTF-8 input when processing these items.)
1341
1342 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1343 size_t is 64-bit (#991).
1344
1345 17. Added a check for running out of memory when PCRE is compiled with
1346 --disable-stack-for-recursion (#990).
1347
1348 18. If the last data line in a file for pcretest does not have a newline on
1349 the end, a newline was missing in the output.
1350
1351 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1352 less than 128) in its various bitmaps. However, there is a facility for
1353 generating tables according to the current locale when PCRE is compiled. It
1354 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1355 characters, are recognized by isspace() and therefore were getting set in
1356 these tables, and indeed these tables seem to approximate to ISO 8859. This
1357 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1358 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1359 which of course cannot start UTF-8 characters. I have changed the code so
1360 that only real ASCII characters (less than 128) and the correct starting
1361 bytes for UTF-8 encodings are set for characters greater than 127 when in
1362 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1363 altogether.)
1364
1365 20. Added the /T option to pcretest so as to be able to run tests with non-
1366 standard character tables, thus making it possible to include the tests
1367 used for 19 above in the standard set of tests.
1368
1369 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1370 reference to a subpattern the other side of a comment that contains an
1371 opening parenthesis caused either an internal compiling error, or a
1372 reference to the wrong subpattern.
1373
1374
1375 Version 8.02 19-Mar-2010
1376 ------------------------
1377
1378 1. The Unicode data tables have been updated to Unicode 5.2.0.
1379
1380 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1381 configured.
1382
1383 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1384 original author of that file, following a query about its status.
1385
1386 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1387 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1388
1389 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1390 quantifier applied to a forward-referencing subroutine call, could compile
1391 incorrect code or give the error "internal error: previously-checked
1392 referenced subpattern not found".
1393
1394 6. Both MS Visual Studio and Symbian OS have problems with initializing
1395 variables to point to external functions. For these systems, therefore,
1396 pcre_malloc etc. are now initialized to local functions that call the
1397 relevant global functions.
1398
1399 7. There were two entries missing in the vectors called coptable and poptable
1400 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1401 I've fixed the data, and added a kludgy way of testing at compile time that
1402 the lengths are correct (equal to the number of opcodes).
1403
1404 8. Following on from 7, I added a similar kludge to check the length of the
1405 eint vector in pcreposix.c.
1406
1407 9. Error texts for pcre_compile() are held as one long string to avoid too
1408 much relocation at load time. To find a text, the string is searched,
1409 counting zeros. There was no check for running off the end of the string,
1410 which could happen if a new error number was added without updating the
1411 string.
1412
1413 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1414
1415 11. \K was not working if it appeared in an atomic group or in a group that
1416 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1417 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1418 the assertion is positive, but not if it is negative.
1419
1420 12. Change 11 fortuitously reduced the size of the stack frame used in the
1421 "match()" function of pcre_exec.c by one pointer. Forthcoming
1422 implementation of support for (*MARK) will need an extra pointer on the
1423 stack; I have reserved it now, so that the stack frame size does not
1424 decrease.
1425
1426 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1427 item in branch that calls a recursion is a subroutine call - as in the
1428 second branch in the above example - was incorrectly given the compile-
1429 time error "recursive call could loop indefinitely" because pcre_compile()
1430 was not correctly checking the subroutine for matching a non-empty string.
1431
1432 14. The checks for overrunning compiling workspace could trigger after an
1433 overrun had occurred. This is a "should never occur" error, but it can be
1434 triggered by pathological patterns such as hundreds of nested parentheses.
1435 The checks now trigger 100 bytes before the end of the workspace.
1436
1437 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1438
1439
1440 Version 8.01 19-Jan-2010
1441 ------------------------
1442
1443 1. If a pattern contained a conditional subpattern with only one branch (in
1444 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1445 computed the wrong minimum data length (which is of course zero for such
1446 subpatterns). This could cause incorrect "no match" results.
1447
1448 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1449 the pattern is reset in the first branch, pcre_compile() failed with
1450 "internal error: code overflow at offset...". This happened only when
1451 the reset was to the original external option setting. (An optimization
1452 abstracts leading options settings into an external setting, which was the
1453 cause of this.)
1454
1455 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1456 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1457 assertion pattern did not match (meaning that the assertion was true), it
1458 was incorrectly treated as false if the SKIP had been reached during the
1459 matching. This also applied to assertions used as conditions.
1460
1461 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1462 assertion subpattern, including such a pattern used as a condition,
1463 unpredictable results occurred, instead of the error return
1464 PCRE_ERROR_DFA_UITEM.
1465
1466 5. The C++ GlobalReplace function was not working like Perl for the special
1467 situation when an empty string is matched. It now does the fancy magic
1468 stuff that is necessary.
1469
1470 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1471 removed. (These were left over from very, very early versions of PCRE.)
1472
1473 7. Some cosmetic changes to the code to make life easier when compiling it
1474 as part of something else:
1475
1476 (a) Change DEBUG to PCRE_DEBUG.
1477
1478 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1479 called "current" as "current_branch", to prevent a collision with the
1480 Linux macro when compiled as a kernel module.
1481
1482 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1483 prevent a collision with the Linux macro when compiled as a kernel
1484 module.
1485
1486 8. In pcre_compile() there are some checks for integer overflows that used to
1487 cast potentially large values to (double). This has been changed to that
1488 when building, a check for int64_t is made, and if it is found, it is used
1489 instead, thus avoiding the use of floating point arithmetic. (There is no
1490 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1491 double.
1492
1493 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1494 2005 (difference between two addresses compared to an unsigned value).
1495
1496 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1497 custom one, because of the following reported problem in Windows:
1498
1499 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1500 under Win32.
1501 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1502 therefore missing the function definition.
1503 - The compiler thus generates a "C" signature for the test function.
1504 - The linker fails to find the "C" function.
1505 - PCRE fails to configure if asked to do so against libbz2.
1506
1507 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1508 messages were output:
1509
1510 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1511 rerunning libtoolize, to keep the correct libtool macros in-tree.
1512 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1513
1514 I have done both of these things.
1515
1516 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1517 most of the time, it *can* run out if it is given a pattern that contains a
1518 runaway infinite recursion. I updated the discussion in the pcrestack man
1519 page.
1520
1521 13. Now that we have gone to the x.xx style of version numbers, the minor
1522 version may start with zero. Using 08 or 09 is a bad idea because users
1523 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1524 interpreted as invalid octal numbers. I've updated the previous comment in
1525 configure.ac, and also added a check that gives an error if 08 or 09 are
1526 used.
1527
1528 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1529 causing partial matching to fail when the end of the subject matched \W
1530 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1531
1532 15. There were some discrepancies between the declarations in pcre_internal.h
1533 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1534 their definitions. The declarations used "const uschar *" and the
1535 definitions used USPTR. Even though USPTR is normally defined as "const
1536 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1537 reported that: "This difference in casting confuses some C++ compilers, for
1538 example, SunCC recognizes above declarations as different functions and
1539 generates broken code for hbpcre." I have changed the declarations to use
1540 USPTR.
1541
1542 16. GNU libtool is named differently on some systems. The autogen.sh script now
1543 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1544 (FreeBSD).
1545
1546 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1547 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1548 comment: "Figure out how to create a longlong from a string: strtoll and
1549 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1550 instance, but it only takes 2 args instead of 3!"
1551
1552 18. A subtle bug concerned with back references has been fixed by a change of
1553 specification, with a corresponding code fix. A pattern such as
1554 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1555 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1556 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1557 same bug. Such groups have to be quantified to be useful, or contained
1558 inside another quantified group. (If there's no repetition, the reference
1559 can never match.) The problem arises because, having left the group and
1560 moved on to the rest of the pattern, a later failure that backtracks into
1561 the group uses the captured value from the final iteration of the group
1562 rather than the correct earlier one. I have fixed this in PCRE by forcing
1563 any group that contains a reference to itself to be an atomic group; that
1564 is, there cannot be any backtracking into it once it has completed. This is
1565 similar to recursive and subroutine calls.
1566
1567
1568 Version 8.00 19-Oct-09
1569 ----------------------
1570
1571 1. The table for translating pcre_compile() error codes into POSIX error codes
1572 was out-of-date, and there was no check on the pcre_compile() error code
1573 being within the table. This could lead to an OK return being given in
1574 error.
1575
1576 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1577 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1578 in a Windows environment.
1579
1580 3. The pcregrep --count option prints the count for each file even when it is
1581 zero, as does GNU grep. However, pcregrep was also printing all files when
1582 --files-with-matches was added. Now, when both options are given, it prints
1583 counts only for those files that have at least one match. (GNU grep just
1584 prints the file name in this circumstance, but including the count seems
1585 more useful - otherwise, why use --count?) Also ensured that the
1586 combination -clh just lists non-zero counts, with no names.
1587
1588 4. The long form of the pcregrep -F option was incorrectly implemented as
1589 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1590 but it seems right to fix it, and I didn't think it was worth preserving
1591 the old behaviour.
1592
1593 5. The command line items --regex=pattern and --regexp=pattern were not
1594 recognized by pcregrep, which required --regex pattern or --regexp pattern
1595 (with a space rather than an '='). The man page documented the '=' forms,
1596 which are compatible with GNU grep; these now work.
1597
1598 6. No libpcreposix.pc file was created for pkg-config; there was just
1599 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1600
1601 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1602 when UCP support is not needed, by modifying the Python script that
1603 generates it from Unicode data files. This should not matter if the module
1604 is correctly used as a library, but I received one complaint about 50K of
1605 unwanted data. My guess is that the person linked everything into his
1606 program rather than using a library. Anyway, it does no harm.
1607
1608 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1609 was a minimum greater than 1 for a wide character in a possessive
1610 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1611 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1612 character. Chaos in the form of incorrect output or a compiling loop could
1613 result.
1614
1615 9. The restrictions on what a pattern can contain when partial matching is
1616 requested for pcre_exec() have been removed. All patterns can now be
1617 partially matched by this function. In addition, if there are at least two
1618 slots in the offset vector, the offset of the earliest inspected character
1619 for the match and the offset of the end of the subject are set in them when
1620 PCRE_ERROR_PARTIAL is returned.
1621
1622 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1623 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1624 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1625 and may be more useful for multi-segment matching.
1626
1627 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1628 used to be given if ever the end of the subject was reached; now it is
1629 given only if matching could not proceed because another character was
1630 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1631 string "Z", which now yields "no match" instead of "partial match". In the
1632 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1633 final character ended with (*FAIL).
1634
1635 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1636 if the pattern had a "must contain" character that was already found in the
1637 earlier partial match, unless partial matching was again requested. For
1638 example, with the pattern /dog.(body)?/, the "must contain" character is
1639 "g". If the first part-match was for the string "dog", restarting with
1640 "sbody" failed. This bug has been fixed.
1641
1642 13. The string returned by pcre_dfa_exec() after a partial match has been
1643 changed so that it starts at the first inspected character rather than the
1644 first character of the match. This makes a difference only if the pattern
1645 starts with a lookbehind assertion or \b or \B (\K is not supported by
1646 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1647 matching functions compatible, and I think it's the right thing to do.
1648
1649 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1650 so that the demonstration program is easily available in environments where
1651 PCRE has not been installed from source.
1652
1653 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1654 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1655 library.
1656
1657 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1658 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1659 is not the first non-POSIX option to be added. Clearly some people find
1660 these options useful.
1661
1662 17. If a caller to the POSIX matching function regexec() passes a non-zero
1663 value for nmatch with a NULL value for pmatch, the value of
1664 nmatch is forced to zero.
1665
1666 18. RunGrepTest did not have a test for the availability of the -u option of
1667 the diff command, as RunTest does. It now checks in the same way as
1668 RunTest, and also checks for the -b option.
1669
1670 19. If an odd number of negated classes containing just a single character
1671 interposed, within parentheses, between a forward reference to a named
1672 subpattern and the definition of the subpattern, compilation crashed with
1673 an internal error, complaining that it could not find the referenced
1674 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1675 [The bug was that it was starting one character too far in when skipping
1676 over the character class, thus treating the ] as data rather than
1677 terminating the class. This meant it could skip too much.]
1678
1679 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1680 /g option in pcretest when the pattern contains \K, which makes it possible
1681 to have an empty string match not at the start, even when the pattern is
1682 anchored. Updated pcretest and pcredemo to use this option.
1683
1684 21. If the maximum number of capturing subpatterns in a recursion was greater
1685 than the maximum at the outer level, the higher number was returned, but
1686 with unset values at the outer level. The correct (outer level) value is
1687 now given.
1688
1689 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1690 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1691 make it do so. The string so far is captured, making this feature
1692 compatible with Perl.
1693
1694 23. The tests have been re-organized, adding tests 11 and 12, to make it
1695 possible to check the Perl 5.10 features against Perl 5.10.
1696
1697 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1698 pattern matches a fixed length string. PCRE did not allow this; now it
1699 does. Neither allows recursion.
1700
1701 25. I finally figured out how to implement a request to provide the minimum
1702 length of subject string that was needed in order to match a given pattern.
1703 (It was back references and recursion that I had previously got hung up
1704 on.) This code has now been added to pcre_study(); it finds a lower bound
1705 to the length of subject needed. It is not necessarily the greatest lower
1706 bound, but using it to avoid searching strings that are too short does give
1707 some useful speed-ups. The value is available to calling programs via
1708 pcre_fullinfo().
1709
1710 26. While implementing 25, I discovered to my embarrassment that pcretest had
1711 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1712 study optimizations had never been tested with that matching function.
1713 Oops. What is worse, even when it was passed study data, there was a bug in
1714 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1715 were also very few tests of studied patterns with pcre_dfa_exec().
1716
1717 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1718 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1719 on the other side of the coin, they are no longer allowed to have different
1720 names, because these cannot be distinguished in PCRE, and this has caused
1721 confusion. (This is a difference from Perl.)
1722
1723 28. When duplicate subpattern names are present (necessarily with different
1724 numbers, as required by 27 above), and a test is made by name in a
1725 conditional pattern, either for a subpattern having been matched, or for
1726 recursion in such a pattern, all the associated numbered subpatterns are
1727 tested, and the overall condition is true if the condition is true for any
1728 one of them. This is the way Perl works, and is also more like the way
1729 testing by number works.
1730
1731
1732 Version 7.9 11-Apr-09
1733 ---------------------
1734
1735 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1736 (pcretest), all targets were linked against these libraries. This included
1737 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1738 libraries. This caused unwanted dependencies to be created. This problem
1739 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1740 pcretest is linked with readline.
1741
1742 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1743 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1744 moved outside it again, because FALSE and TRUE are already defined in AIX,
1745 but BOOL is not.
1746
1747 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1748 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1749
1750 4. The pcregrep documentation said spaces were inserted as well as colons (or
1751 hyphens) following file names and line numbers when outputting matching
1752 lines. This is not true; no spaces are inserted. I have also clarified the
1753 wording for the --colour (or --color) option.
1754
1755 5. In pcregrep, when --colour was used with -o, the list of matching strings
1756 was not coloured; this is different to GNU grep, so I have changed it to be
1757 the same.
1758
1759 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1760 each matching line was coloured. Now it goes on to look for further matches
1761 of any of the test patterns, which is the same behaviour as GNU grep.
1762
1763 7. A pattern that could match an empty string could cause pcregrep to loop; it
1764 doesn't make sense to accept an empty string match in pcregrep, so I have
1765 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1766 seems to be how GNU grep behaves. [But see later change 40 for release
1767 8.33.]
1768
1769 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1770 start or after a newline", because the conditional assertion was not being
1771 correctly handled. The rule now is that both the assertion and what follows
1772 in the first alternative must satisfy the test.
1773
1774 9. If auto-callout was enabled in a pattern with a conditional group whose
1775 condition was an assertion, PCRE could crash during matching, both with
1776 pcre_exec() and pcre_dfa_exec().
1777
1778 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1779 used for matching.
1780
1781 11. Unicode property support in character classes was not working for
1782 characters (bytes) greater than 127 when not in UTF-8 mode.
1783
1784 12. Added the -M command line option to pcretest.
1785
1786 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1787
1788 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1789
1790 16. Added comments and documentation about mis-use of no_arg in the C++
1791 wrapper.
1792
1793 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1794 from Martin Jerabek that uses macro names for all relevant character and
1795 string constants.
1796
1797 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1798 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1799 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1800 these, but not everybody uses configure.
1801
1802 19. A conditional group that had only one branch was not being correctly
1803 recognized as an item that could match an empty string. This meant that an
1804 enclosing group might also not be so recognized, causing infinite looping
1805 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1806 with the subject "ab", where knowledge that the repeated group can match
1807 nothing is needed in order to break the loop.
1808
1809 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1810 exec(), but without supplying a callout function, matching went wrong.
1811
1812 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1813 leak if the size of the offset vector was greater than 30. When the vector
1814 is smaller, the saved offsets during recursion go onto a local stack
1815 vector, but for larger vectors malloc() is used. It was failing to free
1816 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1817 error, in fact).
1818
1819 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1820 heapframe that is used only when UTF-8 support is enabled. This caused no
1821 problem, but was untidy.
1822
1823 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1824 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1825 included within another project.
1826
1827 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1828 slightly modified by me:
1829
1830 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1831 not building pcregrep.
1832
1833 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1834 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1835
1836 25. Forward references, both numeric and by name, in patterns that made use of
1837 duplicate group numbers, could behave incorrectly or give incorrect errors,
1838 because when scanning forward to find the reference group, PCRE was not
1839 taking into account the duplicate group numbers. A pattern such as
1840 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1841
1842 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1843 the feature of a custom pointer more persuasive (as requested by a user).
1844
1845 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1846 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1847 pre-defined.
1848
1849 28. Added support for (*UTF8) at the start of a pattern.
1850
1851 29. Arrange for flags added by the "release type" setting in CMake to be shown
1852 in the configuration summary.
1853
1854
1855 Version 7.8 05-Sep-08
1856 ---------------------
1857
1858 1. Replaced UCP searching code with optimized version as implemented for Ad
1859 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1860 stage table and inline lookup instead of a function, giving speed ups of 2
1861 to 5 times on some simple patterns that I tested. Permission was given to
1862 distribute the MultiStage2.py script that generates the tables (it's not in
1863 the tarball, but is in the Subversion repository).
1864
1865 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1866 scripts.
1867
1868 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1869 a group with a zero qualifier. The result of the study could be incorrect,
1870 or the function might crash, depending on the pattern.
1871
1872 4. Caseless matching was not working for non-ASCII characters in back
1873 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1874 It now works when Unicode Property Support is available.
1875
1876 5. In pcretest, an escape such as \x{de} in the data was always generating
1877 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1878 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1879 truncation.
1880
1881 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1882
1883 7. Added two (int) casts to pcregrep when printing the difference of two
1884 pointers, in case they are 64-bit values.
1885
1886 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1887 test 2 if it fails.
1888
1889 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1890 and a #define of that name to empty if it is not externally set. This is to
1891 allow users of MSVC to set it if necessary.
1892
1893 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1894 the convenience functions in the pcre_get.c source file.
1895
1896 11. An option change at the start of a pattern that had top-level alternatives
1897 could cause overwriting and/or a crash. This command provoked a crash in
1898 some environments:
1899
1900 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1901
1902 This potential security problem was recorded as CVE-2008-2371.
1903
1904 12. For a pattern where the match had to start at the beginning or immediately
1905 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1906 pcre_dfa_exec() could read past the end of the passed subject if there was
1907 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1908 pcretest so that it places the subject at the end of its malloc-ed buffer.
1909
1910 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1911 exec() might read past the end of the data buffer in UTF-8 mode.
1912
1913 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1914 the data contained the byte 0x85 as part of a UTF-8 character within its
1915 first line. This applied both to normal and DFA matching.
1916
1917 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1918 /^[^d]*?$/8 failed to match "abc".
1919
1920 16. Added a missing copyright notice to pcrecpp_internal.h.
1921
1922 17. Make it more clear in the documentation that values returned from
1923 pcre_exec() in ovector are byte offsets, not character counts.
1924
1925 18. Tidied a few places to stop certain compilers from issuing warnings.
1926
1927 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1928 supplied by Stefan Weber. I made a further small update for 7.8 because
1929 there is a change of source arrangements: the pcre_searchfuncs.c module is
1930 replaced by pcre_ucd.c.
1931
1932
1933 Version 7.7 07-May-08
1934 ---------------------
1935
1936 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1937 a string to a long long, pretend we don't even have a long long." This is
1938 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1939
1940 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1941 pre-7.6 versions, which defined a global no_arg variable instead of putting
1942 it in the RE class. (See also #8 below.)
1943
1944 3. Remove a line of dead code, identified by coverity and reported by Nuno
1945 Lopes.
1946
1947 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1948
1949 (1) The include/exclude patterns were being applied to the whole pathnames
1950 of files, instead of just to the final components.
1951
1952 (2) If there was more than one level of directory, the subdirectories were
1953 skipped unless they satisfied the include/exclude conditions. This is
1954 inconsistent with GNU grep (and could even be seen as contrary to the
1955 pcregrep specification - which I improved to make it absolutely clear).
1956 The action now is always to scan all levels of directory, and just
1957 apply the include/exclude patterns to regular files.
1958
1959 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1960 --exclude_dir in the tests to avoid scanning .svn directories.
1961
1962 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1963 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1964 doesn't support NULs in patterns.
1965
1966 7. Added some missing "const"s to declarations of static tables in
1967 pcre_compile.c and pcre_dfa_exec.c.
1968
1969 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1970 caused by fix #2 above. (Subsequently also a second patch to fix the
1971 first patch. And a third patch - this was a messy problem.)
1972
1973 9. Applied Craig's patch to remove the use of push_back().
1974
1975 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1976 matching function regexec().
1977
1978 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1979 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1980 references. PCRE supports relative numbers with this syntax (I don't think
1981 Oniguruma does).
1982
1983 12. Previously, a group with a zero repeat such as (...){0} was completely
1984 omitted from the compiled regex. However, this means that if the group
1985 was called as a subroutine from elsewhere in the pattern, things went wrong
1986 (an internal error was given). Such groups are now left in the compiled
1987 pattern, with a new opcode that causes them to be skipped at execution
1988 time.
1989
1990 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1991 to the way PCRE behaves:
1992
1993 (a) A lone ] character is dis-allowed (Perl treats it as data).
1994
1995 (b) A back reference to an unmatched subpattern matches an empty string
1996 (Perl fails the current match path).
1997
1998 (c) A data ] in a character class must be notated as \] because if the
1999 first data character in a class is ], it defines an empty class. (In
2000 Perl it is not possible to have an empty class.) The empty class []
2001 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2002 The negative empty class [^] matches any one character, independently
2003 of the DOTALL setting.
2004
2005 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2006 non-existent subpattern following a character class starting with ']' and
2007 containing () gave an internal compiling error instead of "reference to
2008 non-existent subpattern". Fortunately, when the pattern did exist, the
2009 compiled code was correct. (When scanning forwards to check for the
2010 existence of the subpattern, it was treating the data ']' as terminating
2011 the class, so got the count wrong. When actually compiling, the reference
2012 was subsequently set up correctly.)
2013
2014 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2015 it was being rejected as not supported by pcre_dfa_exec(), even though
2016 other assertions are supported. I have made pcre_dfa_exec() support
2017 (*FAIL).
2018
2019 16. The implementation of 13c above involved the invention of a new opcode,
2020 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2021 cannot be changed at match time, I realized I could make a small
2022 improvement to matching performance by compiling OP_ALLANY instead of
2023 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2024 on the OP_ANY path.
2025
2026 17. Compiling pcretest on Windows with readline support failed without the
2027 following two fixes: (1) Make the unistd.h include conditional on
2028 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2029
2030 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2031 ncurses library to be included for pcretest when ReadLine support is
2032 requested, but also to allow for it to be overridden. This patch came from
2033 Daniel Bergström.
2034
2035 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2036 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2037 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2038 spotting this.
2039
2040
2041 Version 7.6 28-Jan-08
2042 ---------------------
2043
2044 1. A character class containing a very large number of characters with
2045 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2046 overflow.
2047
2048 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2049 HAVE_LONG_LONG is not defined.
2050
2051 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2052 bring them up to date and include new features. This patch includes:
2053
2054 - Fixed PH's badly added libz and libbz2 support.
2055 - Fixed a problem with static linking.
2056 - Added pcredemo. [But later removed - see 7 below.]
2057 - Fixed dftables problem and added an option.
2058 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2059 HAVE_LONG_LONG.
2060 - Added readline support for pcretest.
2061 - Added an listing of the option settings after cmake has run.
2062
2063 4. A user submitted a patch to Makefile that makes it easy to create
2064 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2065 Makefile.am that cause it to include this special target, without
2066 affecting anything else. Note that the same mingw target plus all
2067 the other distribution libraries and programs are now supported
2068 when configuring with CMake (see 6 below) instead of with
2069 Configure/Make.
2070
2071 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2072 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2073 exported in the Windows port". It has not yet been confirmed that the patch
2074 solves the problem, but it does no harm.
2075
2076 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2077 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2078 with CMake, and also correct the comment about stack recursion.
2079
2080 7. Remove the automatic building of pcredemo from the ./configure system and
2081 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2082 of a program that users should build themselves after PCRE is installed, so
2083 building it automatically is not really right. What is more, it gave
2084 trouble in some build environments.
2085
2086 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2087
2088
2089 Version 7.5 10-Jan-08
2090 ---------------------
2091
2092 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2093 values in parens when parsing an RE using the C++ wrapper."
2094
2095 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2096 Characters greater than 255 were excluded from the class instead of being
2097 included.
2098
2099 3. The same bug as (2) above applied to negated POSIX classes such as
2100 [:^space:].
2101
2102 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2103 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2104 I have changed it.
2105
2106 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2107 first named subpattern) and a construct such as (?&a) would reference the
2108 first named subpattern whose name started with "a" (in other words, the
2109 length check was missing). Both these problems are fixed. "Subpattern name
2110 expected" is now given for (?&) (a zero-length name), and this patch also
2111 makes it give the same error for \k'' (previously it complained that that
2112 was a reference to a non-existent subpattern).
2113
2114 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2115 this is right because (?- can be followed by option settings as well as by
2116 digits. I have, however, made the messages clearer.
2117
2118 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2119 than the number used in the conditional) now cause a compile-time error.
2120 This is actually not compatible with Perl, which accepts such patterns, but
2121 treats the conditional as always being FALSE (as PCRE used to), but it
2122 seems to me that giving a diagnostic is better.
2123
2124 8. Change "alphameric" to the more common word "alphanumeric" in comments
2125 and messages.
2126
2127 9. Fix two occurrences of "backslash" in comments that should have been
2128 "backspace".
2129
2130 10. Remove two redundant lines of code that can never be obeyed (their function
2131 was moved elsewhere).
2132
2133 11. The program that makes PCRE's Unicode character property table had a bug
2134 which caused it to generate incorrect table entries for sequences of
2135 characters that have the same character type, but are in different scripts.
2136 It amalgamated them into a single range, with the script of the first of
2137 them. In other words, some characters were in the wrong script. There were
2138 thirteen such cases, affecting characters in the following ranges:
2139
2140 U+002b0 - U+002c1
2141 U+0060c - U+0060d
2142 U+0061e - U+00612
2143 U+0064b - U+0065e
2144 U+0074d - U+0076d
2145 U+01800 - U+01805
2146 U+01d00 - U+01d77
2147 U+01d9b - U+01dbf
2148 U+0200b - U+0200f
2149 U+030fc - U+030fe
2150 U+03260 - U+0327f
2151 U+0fb46 - U+0fbb1
2152 U+10450 - U+1049d
2153
2154 12. The -o option (show only the matching part of a line) for pcregrep was not
2155 compatible with GNU grep in that, if there was more than one match in a
2156 line, it showed only the first of them. It now behaves in the same way as
2157 GNU grep.
2158
2159 13. If the -o and -v options were combined for pcregrep, it printed a blank
2160 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2161 does the same. The return code can be used to tell if there were any
2162 non-matching lines.
2163
2164 14. Added --file-offsets and --line-offsets to pcregrep.
2165
2166 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2167 infinitely looping recursion. The bug was that positive lookaheads were not
2168 being skipped when checking for a possible empty match (negative lookaheads
2169 and both kinds of lookbehind were skipped).
2170
2171 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2172 inclusion of <windows.h> to before rather than after the definition of
2173 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2174
2175 17. Specifying a possessive quantifier with a specific limit for a Unicode
2176 character property caused pcre_compile() to compile bad code, which led at
2177 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2178 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2179 caused the error; without that there was no problem.
2180
2181 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2182
2183 19. Added --enable-pcretest-libreadline.
2184
2185 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2186 RE::GlobalReplace(). As a result, the number of replacements returned was
2187 double what it should be. I removed one of the increments, but Craig sent a
2188 later patch that removed the other one (the right fix) and added unit tests
2189 that check the return values (which was not done before).
2190
2191 21. Several CMake things:
2192
2193 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2194 the names libpcre and libpcreposix, not just pcre and pcreposix.
2195
2196 (2) The above change means that pcretest and pcregrep are now correctly
2197 linked with the newly-built libraries, not previously installed ones.
2198
2199 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2200
2201 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2202 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2203 UTF-8 newline character). The key issue is that the pattern starts .*;
2204 this means that the match must be either at the beginning, or after a
2205 newline. The bug was in the code for advancing after a failed match and
2206 checking that the new position followed a newline. It was not taking
2207 account of UTF-8 characters correctly.
2208
2209 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2210 character classes. PCRE was not treating the sequence [:...:] as a
2211 character class unless the ... were all letters. Perl, however, seems to
2212 allow any characters between [: and :], though of course it rejects as
2213 unknown any "names" that contain non-letters, because all the known class
2214 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2215 for example, whereas PCRE did not - it did not recognize a POSIX character
2216 class. This seemed a bit dangerous, so the code has been changed to be
2217 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2218 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2219 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2220 Perl does, and where it didn't before.
2221
2222 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2223 Windows environments %n is disabled by default.
2224
2225
2226 Version 7.4 21-Sep-07
2227 ---------------------
2228
2229 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2230 means that a class such as [\s] counted as "explicit reference to CR or
2231 LF". That isn't really right - the whole point of the change was to try to
2232 help when there was an actual mention of one of the two characters. So now
2233 the change happens only if \r or \n (or a literal CR or LF) character is
2234 encountered.
2235
2236 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2237 of both had grown to the point where there were only 3 bits left.
2238 Fortunately, there was spare space in the data structure, and so I have
2239 moved the internal flags into a new 16-bit field to free up more option
2240 bits.
2241
2242 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2243 but did not set the internal JCHANGED flag - either of these is enough to
2244 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2245 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2246 start sets both bits.
2247
2248 4. Added options (at build time, compile time, exec time) to change \R from
2249 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2250
2251 5. doc/pcresyntax.html was missing from the distribution.
2252
2253 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2254 compatibility, even though it is no longer used.
2255
2256 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2257 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2258 windows.h file is present (where different names are used). [This was
2259 reversed later after testing - see 16 below.]
2260
2261 8. Changed all #include <config.h> to #include "config.h". There were also
2262 some further <pcre.h> cases that I changed to "pcre.h".
2263
2264 9. When pcregrep was used with the --colour option, it missed the line ending
2265 sequence off the lines that it output.
2266
2267 10. It was pointed out to me that arrays of string pointers cause lots of
2268 relocations when a shared library is dynamically loaded. A technique of
2269 using a single long string with a table of offsets can drastically reduce
2270 these. I have refactored PCRE in four places to do this. The result is
2271 dramatic:
2272
2273 Originally: 290
2274 After changing UCP table: 187
2275 After changing error message table: 43
2276 After changing table of "verbs" 36
2277 After changing table of Posix names 22
2278
2279 Thanks to the folks working on Gregex for glib for this insight.
2280
2281 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2282 unicode-properties was also set.
2283
2284 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2285
2286 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2287 checked only for CRLF.
2288
2289 14. Added casts to pcretest.c to avoid compiler warnings.
2290
2291 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2292
2293 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2294 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2295 entirely. This removes changes made in 7 above.
2296
2297 17. The CMake files have been updated, and there is now more information about
2298 building with CMake in the NON-UNIX-USE document.
2299
2300
2301 Version 7.3 28-Aug-07
2302 ---------------------
2303
2304 1. In the rejigging of the build system that eventually resulted in 7.1, the
2305 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2306 brackets there is not right, since it causes compilers to look for an
2307 installed pcre.h, not the version that is in the source that is being
2308 compiled (which of course may be different). I have changed it back to:
2309
2310 #include "pcre.h"
2311
2312 I have a vague recollection that the change was concerned with compiling in
2313 different directories, but in the new build system, that is taken care of
2314 by the VPATH setting the Makefile.
2315
2316 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2317 when the subject happened to end in the byte 0x85 (e.g. if the last
2318 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2319 characters but of course it shouldn't be taken as a newline when it is part
2320 of another character. The bug was that, for an unlimited repeat of . in
2321 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2322 characters when looking for a newline.
2323
2324 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2325
2326 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2327 in debug output.
2328
2329 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2330 long printing in the pcrecpp unittest when running under MinGW.
2331
2332 6. ESC_K was left out of the EBCDIC table.
2333
2334 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2335 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2336 limit also applies to "virtual nesting" when a pattern is recursive, and in
2337 this case 1000 isn't so big. I have been able to remove this limit at the
2338 expense of backing off one optimization in certain circumstances. Normally,
2339 when pcre_exec() would call its internal match() function recursively and
2340 immediately return the result unconditionally, it uses a "tail recursion"
2341 feature to save stack. However, when a subpattern that can match an empty
2342 string has an unlimited repetition quantifier, it no longer makes this
2343 optimization. That gives it a stack frame in which to save the data for
2344 checking that an empty string has been matched. Previously this was taken
2345 from the 1000-entry workspace that had been reserved. So now there is no
2346 explicit limit, but more stack is used.
2347
2348 8. Applied Daniel's patches to solve problems with the import/export magic
2349 syntax that is required for Windows, and which was going wrong for the
2350 pcreposix and pcrecpp parts of the library. These were overlooked when this
2351 problem was solved for the main library.
2352
2353 9. There were some crude static tests to avoid integer overflow when computing
2354 the size of patterns that contain repeated groups with explicit upper
2355 limits. As the maximum quantifier is 65535, the maximum group length was
2356 set at 30,000 so that the product of these two numbers did not overflow a
2357 32-bit integer. However, it turns out that people want to use groups that
2358 are longer than 30,000 bytes (though not repeat them that many times).
2359 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2360 made it possible to implement the integer overflow checks in a much more
2361 dynamic way, which I have now done. The artificial limitation on group
2362 length has been removed - we now have only the limit on the total length of
2363 the compiled pattern, which depends on the LINK_SIZE setting.
2364
2365 10. Fixed a bug in the documentation for get/copy named substring when
2366 duplicate names are permitted. If none of the named substrings are set, the
2367 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2368 empty string.
2369
2370 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2371 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2372 because the ] is interpreted as the first data character and the
2373 terminating ] is not found. PCRE has been made compatible with Perl in this
2374 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2375 cause memory overwriting.
2376
2377 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2378 string has been matched (to stop an infinite loop). It was not recognizing
2379 a conditional subpattern that could match an empty string if that
2380 subpattern was within another subpattern. For example, it looped when
2381 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2382 condition was not nested. This bug has been fixed.
2383
2384 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2385 past the start of the subject in the presence of bytes with the top bit
2386 set, for example "\x8aBCD".
2387
2388 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2389 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2390
2391 14. Optimized (?!) to (*FAIL).
2392
2393 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2394 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2395 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2396 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2397 does: it's just the validity check that is more restrictive.
2398
2399 16. Inserted checks for integer overflows during escape sequence (backslash)
2400 processing, and also fixed erroneous offset values for syntax errors during
2401 backslash processing.
2402
2403 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2404 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2405
2406 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2407 caused an overrun.
2408
2409 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2410 something other than just ASCII characters) inside a group that had an
2411 unlimited repeat caused a loop at compile time (while checking to see
2412 whether the group could match an empty string).
2413
2414 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2415 [\P{Any}] did so. (Error in the code for printing property names.)
2416
2417 21. An orphan \E inside a character class could cause a crash.
2418
2419 22. A repeated capturing bracket such as (A)? could cause a wild memory
2420 reference during compilation.
2421
2422 23. There are several functions in pcre_compile() that scan along a compiled
2423 expression for various reasons (e.g. to see if it's fixed length for look
2424 behind). There were bugs in these functions when a repeated \p or \P was
2425 present in the pattern. These operators have additional parameters compared
2426 with \d, etc, and these were not being taken into account when moving along
2427 the compiled data. Specifically:
2428
2429 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2430 length.
2431
2432 (b) An item such as \pL+ within a repeated group could cause crashes or
2433 loops.
2434
2435 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2436 "reference to non-existent subpattern" error.
2437
2438 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2439
2440 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2441 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2442
2443 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2444
2445 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2446 character were causing crashes (broken optimization).
2447
2448 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2449 \p or \P) caused a compile-time loop.
2450
2451 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2452 break. For example, the unstudied pattern [\r\n]A does not match the string
2453 "\r\nA" because change 7.0/46 below moves the current point on by two
2454 characters after failing to match at the start. However, the pattern \nA
2455 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2456 the same is true. There doesn't seem any very clean way out of this, but
2457 what I have chosen to do makes the common cases work: PCRE now takes note
2458 of whether there can be an explicit match for \r or \n anywhere in the
2459 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2460 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2461 pattern has explicit CR or LF references.
2462
2463 29. Added (*CR) etc for changing newline setting at start of pattern.
2464
2465
2466 Version 7.2 19-Jun-07
2467 ---------------------
2468
2469 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2470 which is apparently normally available under Windows.
2471
2472 2. Re-jig the pcregrep tests with different newline settings in an attempt
2473 to make them independent of the local environment's newline setting.
2474
2475 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2476
2477 4. Some of the "internals" tests were previously cut out when the link size
2478 was not 2, because the output contained actual offsets. The recent new
2479 "Z" feature of pcretest means that these can be cut out, making the tests
2480 usable with all link sizes.
2481
2482 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2483 stack recursion. This gives a massive performance boost under BSD, but just
2484 a small improvement under Linux. However, it saves one field in the frame
2485 in all cases.
2486
2487 6. Added more features from the forthcoming Perl 5.10:
2488
2489 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2490 recursion call. It refers to the nth most recently opened parentheses.
2491
2492 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2493 to be opened parentheses.
2494
2495 (c) Conditions that refer to capturing parentheses can be specified
2496 relatively, for example, (?(-2)... or (?(+3)...
2497
2498 (d) \K resets the start of the current match so that everything before
2499 is not part of it.
2500
2501 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2502
2503 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2504 reference syntax.
2505
2506 (g) (?| introduces a group in which the numbering of parentheses in each
2507 alternative starts with the same number.
2508
2509 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2510
2511 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2512 PCRE_INFO_JCHANGED.
2513
2514 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2515 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2516 for detecting groups that can match an empty string.
2517
2518 9. A pattern with a very large number of alternatives (more than several
2519 hundred) was running out of internal workspace during the pre-compile
2520 phase, where pcre_compile() figures out how much memory will be needed. A
2521 bit of new cunning has reduced the workspace needed for groups with
2522 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2523 workspace instead of running out of the 4096 that are available.
2524
2525 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2526
2527 11. Applied patch from Google to remove an optimization that didn't quite work.
2528 The report of the bug said:
2529
2530 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2531 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2532 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2533
2534 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2535 it matched the wrong number of bytes.
2536
2537
2538 Version 7.1 24-Apr-07
2539 ---------------------
2540
2541 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2542 that is more "standard", making use of automake and other Autotools. There
2543 is some re-arrangement of the files and adjustment of comments consequent
2544 on this.
2545
2546 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2547 for recursive directory scanning broke on some systems because the files
2548 are not scanned in any specific order and on different systems the order
2549 was different. A call to "sort" has been inserted into RunGrepTest for the
2550 approprate test as a short-term fix. In the longer term there may be an
2551 alternative.
2552
2553 3. I had an email from Eric Raymond about problems translating some of PCRE's
2554 man pages to HTML (despite the fact that I distribute HTML pages, some
2555 people do their own conversions for various reasons). The problems
2556 concerned the use of low-level troff macros .br and .in. I have therefore
2557 removed all such uses from the man pages (some were redundant, some could
2558 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2559 HTML has been updated to handle .nf/.fi and to complain if it encounters
2560 .br or .in.
2561
2562 4. Updated comments in configure.ac that get placed in config.h.in and also
2563 arranged for config.h to be included in the distribution, with the name
2564 config.h.generic, for the benefit of those who have to compile without
2565 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2566
2567 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2568 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2569 makevp.bat for the current PCRE, using the additional files
2570 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2571
2572 6. A Windows user reported a minor discrepancy with test 2, which turned out
2573 to be caused by a trailing space on an input line that had got lost in his
2574 copy. The trailing space was an accident, so I've just removed it.
2575
2576 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2577 that is needed.
2578
2579 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2580 as "const" (a) because they are and (b) because it helps the PHP
2581 maintainers who have recently made a script to detect big data structures
2582 in the php code that should be moved to the .rodata section. I remembered
2583 to update Builducptable as well, so it won't revert if ucptable.h is ever
2584 re-created.
2585
2586 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2587 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2588 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2589 support is not required. This saves 1.5-2K of code, which is important in
2590 some applications.
2591
2592 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2593 so as not to refer to the tables, even though these functions will never be
2594 called when UTF-8 support is disabled. Otherwise there are problems with a
2595 shared library.
2596
2597 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2598
2599 (a) It was defining its arguments as char * instead of void *.
2600
2601 (b) It was assuming that all moves were upwards in memory; this was true
2602 a long time ago when I wrote it, but is no longer the case.
2603
2604 The emulated memove() is provided for those environments that have neither
2605 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2606 is clearly not the case, as these two bugs were recently reported.
2607
2608 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2609 and Detrail to create the HTML documentation, the .txt form of the man
2610 pages, and it removes trailing spaces from listed files. It also creates
2611 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2612 case, it wraps all the #defines with #ifndefs. This script should be run
2613 before "make dist".
2614
2615 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2616 with Unicode property support.
2617
2618 (a) For a maximizing quantifier, if the two different cases of the
2619 character were of different lengths in their UTF-8 codings (there are
2620 some cases like this - I found 11), and the matching function had to
2621 back up over a mixture of the two cases, it incorrectly assumed they
2622 were both the same length.
2623
2624 (b) When PCRE was configured to use the heap rather than the stack for
2625 recursion during matching, it was not correctly preserving the data for
2626 the other case of a UTF-8 character when checking ahead for a match
2627 while processing a minimizing repeat. If the check also involved
2628 matching a wide character, but failed, corruption could cause an
2629 erroneous result when trying to check for a repeat of the original
2630 character.
2631
2632 13. Some tidying changes to the testing mechanism:
2633
2634 (a) The RunTest script now detects the internal link size and whether there
2635 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2636 values substituted by "configure". (The RunGrepTest script already did
2637 this for UTF-8.) The configure.ac script no longer substitutes the
2638 relevant variables.
2639
2640 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2641 with length and offset values. This means that the output is different
2642 for different internal link sizes. Test 2 is skipped for link sizes
2643 other than 2 because of this, bypassing the problem. Unfortunately,
2644 there was also a test in test 3 (the locale tests) that used /B and
2645 failed for link sizes other than 2. Rather than cut the whole test out,
2646 I have added a new /Z option to pcretest that replaces the length and
2647 offset values with spaces. This is now used to make test 3 independent
2648 of link size. (Test 2 will be tidied up later.)
2649
2650 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2651 segmentation fault instead of returning the appropriate error message.
2652
2653 15. In multiline mode when the newline sequence was set to "any", the pattern
2654 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2655 This doesn't seem right; it now treats the CRLF combination as the line
2656 ending, and so does not match in that case. It's only a pattern such as ^$
2657 that would hit this one: something like ^ABC$ would have failed after \r
2658 and then tried again after \r\n.
2659
2660 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2661 in an attempt to make files that differ only in their line terminators
2662 compare equal. This works on Linux.
2663
2664 17. Under certain error circumstances pcregrep might try to free random memory
2665 as it exited. This is now fixed, thanks to valgrind.
2666
2667 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2668 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2669 was because its rules for how to advance for /g after matching an empty
2670 string at the end of a line did not allow for this case. They now check for
2671 it specially.
2672
2673 20. pcretest is supposed to handle patterns and data of any length, by
2674 extending its buffers when necessary. It was getting this wrong when the
2675 buffer for a data line had to be extended.
2676
2677 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2678 CRLF as a newline sequence.
2679
2680 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2681 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2682 I have nevertheless tidied it up.
2683
2684 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2685
2686 24. Added a man page for pcre-config.
2687
2688
2689 Version 7.0 19-Dec-06
2690 ---------------------
2691
2692 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2693 moving to gcc 4.1.1.
2694
2695 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2696 sys/time.h, which is documented as needed for this function. It doesn't
2697 seem to matter on Linux, but it showed up on some releases of OS X.
2698
2699 3. It seems that there are systems where bytes whose values are greater than
2700 127 match isprint() in the "C" locale. The "C" locale should be the
2701 default when a C program starts up. In most systems, only ASCII printing
2702 characters match isprint(). This difference caused the output from pcretest
2703 to vary, making some of the tests fail. I have changed pcretest so that:
2704
2705 (a) When it is outputting text in the compiled version of a pattern, bytes
2706 other than 32-126 are always shown as hex escapes.
2707
2708 (b) When it is outputting text that is a matched part of a subject string,
2709 it does the same, unless a different locale has been set for the match
2710 (using the /L modifier). In this case, it uses isprint() to decide.
2711
2712 4. Fixed a major bug that caused incorrect computation of the amount of memory
2713 required for a compiled pattern when options that changed within the
2714 pattern affected the logic of the preliminary scan that determines the
2715 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2716 that the computed length was too small. The symptoms of this bug were
2717 either the PCRE error "internal error: code overflow" from pcre_compile(),
2718 or a glibc crash with a message such as "pcretest: free(): invalid next
2719 size (fast)". Examples of patterns that provoked this bug (shown in
2720 pcretest format) are:
2721
2722 /(?-x: )/x
2723 /(?x)(?-x: \s*#\s*)/
2724 /((?i)[\x{c0}])/8
2725 /(?i:[\x{c0}])/8
2726
2727 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2728 is now done differently.
2729
2730 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2731 wrapper classes; (b) implement a new function in the C++ scanner that is
2732 more efficient than the old way of doing things because it avoids levels of
2733 recursion in the regex matching; (c) add a paragraph to the documentation
2734 for the FullMatch() function.
2735
2736 6. The escape sequence \n was being treated as whatever was defined as
2737 "newline". Not only was this contrary to the documentation, which states
2738 that \n is character 10 (hex 0A), but it also went horribly wrong when
2739 "newline" was defined as CRLF. This has been fixed.
2740
2741 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2742 was being set to -1 for the "end of line" case (supposedly a value that no
2743 character can have). Though this value is never used (the check for end of
2744 line is "zero bytes in current character"), it caused compiler complaints.
2745 I've changed it to 0xffffffff.
2746
2747 8. In pcre_version.c, the version string was being built by a sequence of
2748 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2749 string (as it is for production releases) called a macro with an empty
2750 argument. The C standard says the result of this is undefined. The gcc
2751 compiler treats it as an empty string (which was what was wanted) but it is
2752 reported that Visual C gives an error. The source has been hacked around to
2753 avoid this problem.
2754
2755 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2756 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2757 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2758 of them did).
2759
2760 10. Originally, pcretest opened its input and output without "b"; then I was
2761 told that "b" was needed in some environments, so it was added for release
2762 5.0 to both the input and output. (It makes no difference on Unix-like
2763 systems.) Later I was told that it is wrong for the input on Windows. I've
2764 now abstracted the modes into two macros, to make it easier to fiddle with
2765 them, and removed "b" from the input mode under Windows.
2766
2767 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2768
2769 12. Added -help and --help to pcretest as an official way of being reminded
2770 of the options.
2771
2772 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2773 and pcrecpp.cc because they annoy compilers at high warning levels.
2774
2775 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2776
2777 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2778 scripts are not C programs :-) and which was not noticed because it works
2779 on Linux.
2780
2781 16. pcretest is supposed to handle any length of pattern and data line (as one
2782 line or as a continued sequence of lines) by extending its input buffer if
2783 necessary. This feature was broken for very long pattern lines, leading to
2784 a string of junk being passed to pcre_compile() if the pattern was longer
2785 than about 50K.
2786
2787 17. I have done a major re-factoring of the way pcre_compile() computes the
2788 amount of memory needed for a compiled pattern. Previously, there was code
2789 that made a preliminary scan of the pattern in order to do this. That was
2790 OK when PCRE was new, but as the facilities have expanded, it has become
2791 harder and harder to keep it in step with the real compile phase, and there
2792 have been a number of bugs (see for example, 4 above). I have now found a
2793 cunning way of running the real compile function in a "fake" mode that
2794 enables it to compute how much memory it would need, while actually only
2795 ever using a few hundred bytes of working memory and without too many
2796 tests of the mode. This should make future maintenance and development
2797 easier. A side effect of this work is that the limit of 200 on the nesting
2798 depth of parentheses has been removed (though this was never a serious
2799 limitation, I suspect). However, there is a downside: pcre_compile() now
2800 runs more slowly than before (30% or more, depending on the pattern). I
2801 hope this isn't a big issue. There is no effect on runtime performance.
2802
2803 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2804 newline (only possible for the last line of a file) and it was a
2805 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2806
2807 19. Added additional timing features to pcretest. (1) The -tm option now times
2808 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2809 separate command line item, by a number that specifies the number of
2810 repeats to use when timing. The default is 50000; this gives better
2811 precision, but takes uncomfortably long for very large patterns.
2812
2813 20. Extended pcre_study() to be more clever in cases where a branch of a
2814 subpattern has no definite first character. For example, (a*|b*)[cd] would
2815 previously give no result from pcre_study(). Now it recognizes that the
2816 first character must be a, b, c, or d.
2817
2818 21. There was an incorrect error "recursive call could loop indefinitely" if
2819 a subpattern (or the entire pattern) that was being tested for matching an
2820 empty string contained only one non-empty item after a nested subpattern.
2821 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2822 incorrectly, because the \d was being skipped in the check.
2823
2824 22. The pcretest program now has a new pattern option /B and a command line
2825 option -b, which is equivalent to adding /B to every pattern. This causes
2826 it to show the compiled bytecode, without the additional information that
2827 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2828 is the same as /B/I).
2829
2830 23. A new optimization is now able automatically to treat some sequences such
2831 as a*b as a*+b. More specifically, if something simple (such as a character
2832 or a simple class like \d) has an unlimited quantifier, and is followed by
2833 something that cannot possibly match the quantified thing, the quantifier
2834 is automatically "possessified".
2835
2836 24. A recursive reference to a subpattern whose number was greater than 39
2837 went wrong under certain circumstances in UTF-8 mode. This bug could also
2838 have affected the operation of pcre_study().
2839
2840 25. Realized that a little bit of performance could be had by replacing
2841 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2842
2843 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2844
2845 27. Possessive quantifiers such as a++ were previously implemented by turning
2846 them into atomic groups such as ($>a+). Now they have their own opcodes,
2847 which improves performance. This includes the automatically created ones
2848 from 23 above.
2849
2850 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2851 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2852 the first matched character to be a colon. This applied both to named and
2853 numbered groups.
2854
2855 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2856
2857 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2858 building PCRE on Windows easier, so I have included it in the distribution.
2859
2860 31. There is now a check in pcretest against a ridiculously large number being
2861 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2862 loop, the loop is abandoned.
2863
2864 32. Forward references to subpatterns in conditions such as (?(2)...) where
2865 subpattern 2 is defined later cause pcre_compile() to search forwards in
2866 the pattern for the relevant set of parentheses. This search went wrong
2867 when there were unescaped parentheses in a character class, parentheses
2868 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2869
2870 33. "Subroutine" calls and backreferences were previously restricted to
2871 referencing subpatterns earlier in the regex. This restriction has now
2872 been removed.
2873
2874 34. Added a number of extra features that are going to be in Perl 5.10. On the
2875 whole, these are just syntactic alternatives for features that PCRE had
2876 previously implemented using the Python syntax or my own invention. The
2877 other formats are all retained for compatibility.
2878
2879 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2880 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2881 also .NET compatible.
2882
2883 (b) A recursion or subroutine call to a named group can now be defined as
2884 (?&name) as well as (?P>name).
2885
2886 (c) A backreference to a named group can now be defined as \k<name> or
2887 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2888 5.10, are also .NET compatible.
2889
2890 (d) A conditional reference to a named group can now use the syntax
2891 (?(<name>) or (?('name') as well as (?(name).
2892
2893 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2894 groups (named and numbered) that are never evaluated inline, but can be
2895 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2896 is always false. There may be only one alternative in such a group.
2897
2898 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2899 as the simple (?(R). The condition is true only if the most recent
2900 recursion is that of the given number or name. It does not search out
2901 through the entire recursion stack.
2902
2903 (g) The escape \gN or \g{N} has been added, where N is a positive or
2904 negative number, specifying an absolute or relative reference.
2905
2906 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2907 some "unreachable code" warnings.
2908
2909 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2910 things, this adds five new scripts.
2911
2912 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2913 There were also incompatibilities regarding the handling of \Q..\E inside
2914 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2915 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2916
2917 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2918 matches an empty string, and forcibly breaks the loop. There were bugs in
2919 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2920 against aaaa the result was just "a" rather than "aaaa", for example. Two
2921 separate and independent bugs (that affected different cases) have been
2922 fixed.
2923
2924 39. Refactored the code to abolish the use of different opcodes for small
2925 capturing bracket numbers. This is a tidy that I avoided doing when I
2926 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2927 The new approach is not only tidier, it makes it possible to reduce the
2928 memory needed to fix the previous bug (38).
2929
2930 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2931 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2932 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2933 mode.
2934
2935 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2936 report.
2937
2938 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2939 copy construction and assignment in the C++ wrapper.
2940
2941 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2942 couple of bugs that could have given wrong results in the "--newline=crlf"
2943 case.
2944
2945 44. Added a number of casts and did some reorganization of signed/unsigned int
2946 variables following suggestions from Dair Grant. Also renamed the variable
2947 "this" as "item" because it is a C++ keyword.
2948
2949 45. Arranged for dftables to add
2950
2951 #include "pcre_internal.h"
2952
2953 to pcre_chartables.c because without it, gcc 4.x may remove the array
2954 definition from the final binary if PCRE is built into a static library and
2955 dead code stripping is activated.
2956
2957 46. For an unanchored pattern, if a match attempt fails at the start of a
2958 newline sequence, and the newline setting is CRLF or ANY, and the next two
2959 characters are CRLF, advance by two characters instead of one.
2960
2961
2962 Version 6.7 04-Jul-06
2963 ---------------------
2964
2965 1. In order to handle tests when input lines are enormously long, pcretest has
2966 been re-factored so that it automatically extends its buffers when
2967 necessary. The code is crude, but this _is_ just a test program. The
2968 default size has been increased from 32K to 50K.
2969
2970 2. The code in pcre_study() was using the value of the re argument before
2971 testing it for NULL. (Of course, in any sensible call of the function, it
2972 won't be NULL.)
2973
2974 3. The memmove() emulation function in pcre_internal.h, which is used on
2975 systems that lack both memmove() and bcopy() - that is, hardly ever -
2976 was missing a "static" storage class specifier.
2977
2978 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2979 containing an extended class (one that cannot be represented by a bitmap
2980 because it contains high-valued characters or Unicode property items, e.g.
2981 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2982 pattern, but PCRE should not loop if you do not (it no longer does).
2983 [Detail: two cases were found: (a) a repeated subpattern containing an
2984 extended class; (b) a recursive reference to a subpattern that followed a
2985 previous extended class. It wasn't skipping over the extended class
2986 correctly when UTF-8 mode was not set.]
2987
2988 5. A negated single-character class was not being recognized as fixed-length
2989 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2990 compile error "lookbehind assertion is not fixed length".
2991
2992 6. The RunPerlTest auxiliary script was showing an unexpected difference
2993 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2994 write a Perl script that can interpret lines of an input file either as
2995 byte characters or as UTF-8, which is what "perltest" was being required to
2996 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2997 can't do is switch easily at run time between having the "use utf8;" pragma
2998 or not. In the end, I fudged it by using the RunPerlTest script to insert
2999 "use utf8;" explicitly for the UTF-8 tests.
3000
3001 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3002 the end of the subject string, contrary to the documentation and to what
3003 Perl does. This was true of both matching functions. Now it matches only at
3004 the start of the subject and immediately after *internal* newlines.
3005
3006 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3007 a pointer to an int instead of a pointer to an unsigned long int. This
3008 caused problems on 64-bit systems.
3009
3010 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3011 instance of the 'standard' template library not being so standard".
3012
3013 10. There was no check on the number of named subpatterns nor the maximum
3014 length of a subpattern name. The product of these values is used to compute
3015 the size of the memory block for a compiled pattern. By supplying a very
3016 long subpattern name and a large number of named subpatterns, the size
3017 computation could be caused to overflow. This is now prevented by limiting
3018 the length of names to 32 characters, and the number of named subpatterns
3019 to 10,000.
3020
3021 11. Subpatterns that are repeated with specific counts have to be replicated in
3022 the compiled pattern. The size of memory for this was computed from the
3023 length of the subpattern and the repeat count. The latter is limited to
3024 65535, but there was no limit on the former, meaning that integer overflow
3025 could in principle occur. The compiled length of a repeated subpattern is
3026 now limited to 30,000 bytes in order to prevent this.
3027
3028 12. Added the optional facility to have named substrings with the same name.
3029
3030 13. Added the ability to use a named substring as a condition, using the
3031 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3032 are numbers (not recommended). Forward references are permitted.
3033
3034 14. Added forward references in named backreferences (if you see what I mean).
3035
3036 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3037 pattern could run off the end of the subject. For example, the pattern
3038 "(?s)(.{1,5})"8 did this with the subject "ab".
3039
3040 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3041 PCRE_CASELESS was set when matching characters that were quantified with ?
3042 or *.
3043
3044 17. A character class other than a single negated character that had a minimum
3045 but no maximum quantifier - for example [ab]{6,} - was not handled
3046 correctly by pce_dfa_exec(). It would match only one character.
3047
3048 18. A valid (though odd) pattern that looked like a POSIX character
3049 class but used an invalid character after [ (for example [[,abc,]]) caused
3050 pcre_compile() to give the error "Failed: internal error: code overflow" or
3051 in some cases to crash with a glibc free() error. This could even happen if
3052 the pattern terminated after [[ but there just happened to be a sequence of
3053 letters, a binary zero, and a closing ] in the memory that followed.
3054
3055 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3056 over the years. Originally (before any Unicode support), just the bottom 8
3057 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3058 output from "man perlunicode" includes this:
3059
3060 The regular expression compiler produces polymorphic opcodes. That
3061 is, the pattern adapts to the data and automatically switches to
3062 the Unicode character scheme when presented with Unicode data--or
3063 instead uses a traditional byte scheme when presented with byte
3064 data.
3065
3066 Sadly, a wide octal escape does not cause a switch, and in a string with
3067 no other multibyte characters, these octal escapes are treated as before.
3068 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3069 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3070 Unicode string.
3071
3072 I have not perpetrated such confusion in PCRE. Up till now, it took just
3073 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3074 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3075 translate to the appropriate multibyte character.
3076
3077 29. Applied some refactoring to reduce the number of warnings from Microsoft
3078 and Borland compilers. This has included removing the fudge introduced
3079 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3080 a warning about an unused variable.
3081
3082 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3083 characters since release 4.0, because Perl (from release 5.004) does not.
3084 [Or at least, is documented not to: some releases seem to be in conflict
3085 with the documentation.] However, when a pattern was studied with
3086 pcre_study() and all its branches started with \s, PCRE still included VT
3087 as a possible starting character. Of course, this did no harm; it just
3088 caused an unnecessary match attempt.
3089
3090 22. Removed a now-redundant internal flag bit that recorded the fact that case
3091 dependency changed within the pattern. This was once needed for "required
3092 byte" processing, but is no longer used. This recovers a now-scarce options
3093 bit. Also moved the least significant internal flag bit to the most-
3094 significant bit of the word, which was not previously used (hangover from
3095 the days when it was an int rather than a uint) to free up another bit for
3096 the future.
3097
3098 23. Added support for CRLF line endings as well as CR and LF. As well as the
3099 default being selectable at build time, it can now be changed at runtime
3100 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3101 specify that it is scanning data with non-default line endings.
3102
3103 24. Changed the definition of CXXLINK to make it agree with the definition of
3104 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3105
3106 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3107 recursions. This makes a big different to stack usage for some patterns.
3108
3109 26. If a subpattern containing a named recursion or subroutine reference such
3110 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3111 the space required for the compiled pattern went wrong and gave too small a
3112 value. Depending on the environment, this could lead to "Failed: internal
3113 error: code overflow at offset 49" or "glibc detected double free or
3114 corruption" errors.
3115
3116 27. Applied patches from Google (a) to support the new newline modes and (b) to
3117 advance over multibyte UTF-8 characters in GlobalReplace.
3118
3119 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3120 difference for some implementation of PCRE in some Windows version.
3121
3122 29. Added some extra testing facilities to pcretest:
3123
3124 \q<number> in a data line sets the "match limit" value
3125 \Q<number> in a data line sets the "match recursion limt" value
3126 -S <number> sets the stack size, where <number> is in megabytes
3127
3128 The -S option isn't available for Windows.
3129
3130
3131 Version 6.6 06-Feb-06
3132 ---------------------
3133
3134 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3135 in pcreposix.h. I have copied the definition from pcre.h.
3136
3137 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3138 because pcre.h is no longer a built file.
3139
3140 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3141 not normally included in the compiled code.
3142
3143
3144 Version 6.5 01-Feb-06
3145 ---------------------
3146
3147 1. When using the partial match feature with pcre_dfa_exec(), it was not
3148 anchoring the second and subsequent partial matches at the new starting
3149 point. This could lead to incorrect results. For example, with the pattern
3150 /1234/, partially matching against "123" and then "a4" gave a match.
3151
3152 2. Changes to pcregrep:
3153
3154 (a) All non-match returns from pcre_exec() were being treated as failures
3155 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3156 error message is output. Some extra information is given for the
3157 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3158 probably the only errors that are likely to be caused by users (by
3159 specifying a regex that has nested indefinite repeats, for instance).
3160 If there are more than 20 of these errors, pcregrep is abandoned.
3161
3162 (b) A binary zero was treated as data while matching, but terminated the
3163 output line if it was written out. This has been fixed: binary zeroes
3164 are now no different to any other data bytes.
3165
3166 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3167 used to set a locale for matching. The --locale=xxxx long option has
3168 been added (no short equivalent) to specify a locale explicitly on the
3169 pcregrep command, overriding the environment variables.
3170
3171 (d) When -B was used with -n, some line numbers in the output were one less
3172 than they should have been.
3173
3174 (e) Added the -o (--only-matching) option.
3175
3176 (f) If -A or -C was used with -c (count only), some lines of context were
3177 accidentally printed for the final match.
3178
3179 (g) Added the -H (--with-filename) option.
3180
3181 (h) The combination of options -rh failed to suppress file names for files
3182 that were found from directory arguments.
3183
3184 (i) Added the -D (--devices) and -d (--directories) options.
3185
3186 (j) Added the -F (--fixed-strings) option.
3187
3188 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3189
3190 (l) Added the --colo(u)r option.
3191
3192 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3193 is not present by default.
3194
3195 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3196 items such as (?R) or (?1), when the recursion could match a number of
3197 alternatives. If it matched one of the alternatives, but subsequently,
3198 outside the recursion, there was a failure, the code tried to back up into
3199 the recursion. However, because of the way PCRE is implemented, this is not
3200 possible, and the result was an incorrect result from the match.
3201
3202 In order to prevent this happening, the specification of recursion has
3203 been changed so that all such subpatterns are automatically treated as
3204 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3205
3206 4. I had overlooked the fact that, in some locales, there are characters for
3207 which isalpha() is true but neither isupper() nor islower() are true. In
3208 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3209 and ordfeminine) are like this. This affected the treatment of \w and \W
3210 when they appeared in character classes, but not when they appeared outside
3211 a character class. The bit map for "word" characters is now created
3212 separately from the results of isalnum() instead of just taking it from the
3213 upper, lower, and digit maps. (Plus the underscore character, of course.)
3214
3215 5. The above bug also affected the handling of POSIX character classes such as
3216 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3217 permanent tables. Instead, the bit maps for such a class were previously
3218 created as the appropriate unions of the upper, lower, and digit bitmaps.
3219 Now they are created by subtraction from the [[:word:]] class, which has
3220 its own bitmap.
3221
3222 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3223 It is created by subtracting the vertical space characters (\x09, \x0a,
3224 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3225 subtraction was done in the overall bitmap for a character class, meaning
3226 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3227 be recognized. This bug has been fixed.
3228
3229 7. Patches from the folks at Google:
3230
3231 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3232 real life, but is still worth protecting against".
3233
3234 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3235 regular expressions".
3236
3237 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3238 have it.
3239
3240 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3241 "configure" and the latter not, in order to fix a problem somebody had
3242 with compiling the Arg class on HP-UX.
3243
3244 (e) Improve the error-handling of the C++ wrapper a little bit.
3245
3246 (f) New tests for checking recursion limiting.
3247
3248 8. The pcre_memmove() function, which is used only if the environment does not
3249 have a standard memmove() function (and is therefore rarely compiled),
3250 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3251 returning a result (though PCRE never actually uses the result).
3252
3253 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3254 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3255 returned instead of calling malloc() with an overflowing number that would
3256 most likely cause subsequent chaos.
3257
3258 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3259
3260 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3261 with this option is matched, the nmatch and pmatch options of regexec() are
3262 ignored.
3263
3264 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3265 provided in case anyone wants to the the POSIX interface with UTF-8
3266 strings.
3267
3268 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3269 C++ linking (needed for some HP-UX environments).
3270
3271 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3272 (unused parameter) and in the pcre_printint() function (omitted "default"
3273 switch label when the default is to do nothing).
3274
3275 15. Added some code to make it possible, when PCRE is compiled as a C++
3276 library, to replace subject pointers for pcre_exec() with a smart pointer
3277 class, thus making it possible to process discontinuous strings.
3278
3279 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3280 much the same function. They were added by different people who were trying
3281 to make PCRE easy to compile on non-Unix systems. It has been suggested
3282 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3283 for compiling on Windows systems. I have therefore replaced it with
3284 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3285 defaults to "extern" for C or "extern C" for C++, which works fine on
3286 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3287 SCOPE with something explicit in config.h. In addition:
3288
3289 (a) pcreposix.h still had just "extern" instead of either of these macros;
3290 I have replaced it with PCRE_DATA_SCOPE.
3291
3292 (b) Functions such as _pcre_xclass(), which are internal to the library,
3293 but external in the C sense, all had PCRE_EXPORT in their definitions.
3294 This is apparently wrong for the Windows case, so I have removed it.
3295 (It makes no difference on Unix-like systems.)
3296
3297 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3298 of recursive calls to match(). This is different to MATCH_LIMIT because
3299 that limits the total number of calls to match(), not all of which increase
3300 the depth of recursion. Limiting the recursion depth limits the amount of
3301 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3302 when PCRE is compiled, and changed at run time. A patch from Google adds
3303 this functionality to the C++ interface.
3304
3305 18. Changes to the handling of Unicode character properties:
3306
3307 (a) Updated the table to Unicode 4.1.0.
3308
3309 (b) Recognize characters that are not in the table as "Cn" (undefined).
3310
3311 (c) I revised the way the table is implemented to a much improved format
3312 which includes recognition of ranges. It now supports the ranges that
3313 are defined in UnicodeData.txt, and it also amalgamates other
3314 characters into ranges. This has reduced the number of entries in the
3315 table from around 16,000 to around 3,000, thus reducing its size
3316 considerably. I realized I did not need to use a tree structure after
3317 all - a binary chop search is just as efficient. Having reduced the
3318 number of entries, I extended their size from 6 bytes to 8 bytes to
3319 allow for more data.
3320
3321 (d) Added support for Unicode script names via properties such as \p{Han}.
3322
3323 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3324 matching that character.
3325
3326 20. When matching a repeated Unicode property with a minimum greater than zero,
3327 (for example \pL{2,}), PCRE could look past the end of the subject if it
3328 reached it while seeking the minimum number of characters. This could
3329 happen only if some of the characters were more than one byte long, because
3330 there is a check for at least the minimum number of bytes.
3331
3332 21. Refactored the implementation of \p and \P so as to be more general, to
3333 allow for more different types of property in future. This has changed the
3334 compiled form incompatibly. Anybody with saved compiled patterns that use
3335 \p or \P will have to recompile them.
3336
3337 22. Added "Any" and "L&" to the supported property types.
3338
3339 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3340 but give a compile time error if the value is greater than 0xff.
3341
3342 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3343 accidentally not being installed or uninstalled.
3344
3345 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3346 made were to insert the current release number. This seemed silly, because
3347 it made things harder for people building PCRE on systems that don't run
3348 "configure". I have turned pcre.h into a distributed file, no longer built
3349 by "configure", with the version identification directly included. There is
3350 no longer a pcre.h.in file.
3351
3352 However, this change necessitated a change to the pcre-config script as
3353 well. It is built from pcre-config.in, and one of the substitutions was the
3354 release number. I have updated configure.ac so that ./configure now finds
3355 the release number by grepping pcre.h.
3356
3357 26. Added the ability to run the tests under valgrind.
3358
3359
3360 Version 6.4 05-Sep-05
3361 ---------------------
3362
3363 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3364 "--" to be printed when multiple files were scanned, even when none of the
3365 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3366 consider it to be a bug, and have restored the previous behaviour.
3367
3368 2. A couple of code tidies to get rid of compiler warnings.
3369
3370 3. The pcretest program used to cheat by referring to symbols in the library
3371 whose names begin with _pcre_. These are internal symbols that are not
3372 really supposed to be visible externally, and in some environments it is
3373 possible to suppress them. The cheating is now confined to including
3374 certain files from the library's source, which is a bit cleaner.
3375
3376 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3377 file's purpose clearer.
3378
3379 5. Reorganized pcre_ucp_findchar().
3380
3381
3382 Version 6.3 15-Aug-05
3383 ---------------------
3384
3385 1. The file libpcre.pc.in did not have general read permission in the tarball.
3386
3387 2. There were some problems when building without C++ support:
3388
3389 (a) If C++ support was not built, "make install" and "make test" still
3390 tried to test it.
3391
3392 (b) There were problems when the value of CXX was explicitly set. Some
3393 changes have been made to try to fix these, and ...
3394
3395 (c) --disable-cpp can now be used to explicitly disable C++ support.
3396
3397 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3398 backslash in a target when C++ was disabled. This confuses some
3399 versions of "make", apparently. Using an intermediate variable solves
3400 this. (Same for CPP_LOBJ.)
3401
3402 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3403 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3404 necessary on certain architectures.
3405
3406 4. Added a setting of -export-symbols-regex to the link command to remove
3407 those symbols that are exported in the C sense, but actually are local
3408 within the library, and not documented. Their names all begin with
3409 "_pcre_". This is not a perfect job, because (a) we have to except some
3410 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3411 available (and never for static libraries). I have made a note to try to
3412 find a way round (a) in the future.
3413
3414
3415 Version 6.2 01-Aug-05
3416 ---------------------
3417
3418 1. There was no test for integer overflow of quantifier values. A construction
3419 such as {1111111111111111} would give undefined results. What is worse, if
3420 a minimum quantifier for a parenthesized subpattern overflowed and became
3421 negative, the calculation of the memory size went wrong. This could have
3422 led to memory overwriting.
3423
3424 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3425
3426 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3427 operating environments where this matters.
3428
3429 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3430 PCRE options from within the C++ wrapper.
3431
3432 5. Named capturing subpatterns were not being correctly counted when a pattern
3433 was compiled. This caused two problems: (a) If there were more than 100
3434 such subpatterns, the calculation of the memory needed for the whole
3435 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3436 back references of the form \12, where the number was greater than 9, were
3437 not recognized as back references, even though there were sufficient
3438 previous subpatterns.
3439
3440 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3441 versions of gcc, e.g. 2.95.4.
3442
3443
3444 Version 6.1 21-Jun-05
3445 ---------------------
3446
3447 1. There was one reference to the variable "posix" in pcretest.c that was not
3448 surrounded by "#if !defined NOPOSIX".
3449
3450 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3451 the cross-check on the old pcre_info() function, for the benefit of the
3452 cut-down version of PCRE that is currently imported into Exim.
3453
3454 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3455 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3456 patterns that start (?i) but I don't think that matters. The use of (?i) is
3457 just an example; this all applies to the other options as well.
3458
3459 4. Since libtool seems to echo the compile commands it is issuing, the output
3460 from "make" can be reduced a bit by putting "@" in front of each libtool
3461 compile command.
3462
3463 5. Patch from the folks at Google for configure.in to be a bit more thorough
3464 in checking for a suitable C++ installation before trying to compile the
3465 C++ stuff. This should fix a reported problem when a compiler was present,
3466 but no suitable headers.
3467
3468 6. The man pages all had just "PCRE" as their title. I have changed them to
3469 be the relevant file name. I have also arranged that these names are
3470 retained in the file doc/pcre.txt, which is a concatenation in text format
3471 of all the man pages except the little individual ones for each function.
3472
3473 7. The NON-UNIX-USE file had not been updated for the different set of source
3474 files that come with release 6. I also added a few comments about the C++
3475 wrapper.
3476
3477
3478 Version 6.0 07-Jun-05
3479 ---------------------
3480
3481 1. Some minor internal re-organization to help with my DFA experiments.
3482
3483 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3484 didn't matter for the library itself when fully configured, but did matter
3485 when compiling without UCP support, or within Exim, where the ucp files are
3486 not imported.
3487
3488 3. Refactoring of the library code to split up the various functions into
3489 different source modules. The addition of the new DFA matching code (see
3490 below) to a single monolithic source would have made it really too
3491 unwieldy, quite apart from causing all the code to be include in a
3492 statically linked application, when only some functions are used. This is
3493 relevant even without the DFA addition now that patterns can be compiled in
3494 one application and matched in another.
3495
3496 The downside of splitting up is that there have to be some external
3497 functions and data tables that are used internally in different modules of
3498 the library but which are not part of the API. These have all had their
3499 names changed to start with "_pcre_" so that they are unlikely to clash
3500 with other external names.
3501
3502 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3503 a different (DFA) algorithm. Although it is slower than the original
3504 function, it does have some advantages for certain types of matching
3505 problem.
3506
3507 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3508 including restarting after a partial match.
3509
3510 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3511 defined when compiling for Windows was sent to me. I have put it into the
3512 code, though I have no means of testing or verifying it.
3513
3514 7. Added the pcre_refcount() auxiliary function.
3515
3516 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3517 match before or at the first newline in the subject string. In pcretest,
3518 the /f option on a pattern can be used to set this.
3519
3520 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3521 would behave wrongly. This has been present in PCRE since release 4.0.
3522
3523 10. A number of changes to the pcregrep command:
3524
3525 (a) Refactored how -x works; insert ^(...)$ instead of setting
3526 PCRE_ANCHORED and checking the length, in preparation for adding
3527 something similar for -w.
3528
3529 (b) Added the -w (match as a word) option.
3530
3531 (c) Refactored the way lines are read and buffered so as to have more
3532 than one at a time available.
3533
3534 (d) Implemented a pcregrep test script.
3535
3536 (e) Added the -M (multiline match) option. This allows patterns to match
3537 over several lines of the subject. The buffering ensures that at least
3538 8K, or the rest of the document (whichever is the shorter) is available
3539 for matching (and similarly the previous 8K for lookbehind assertions).
3540
3541 (f) Changed the --help output so that it now says
3542
3543 -w, --word-regex(p)
3544
3545 instead of two lines, one with "regex" and the other with "regexp"
3546 because that confused at least one person since the short forms are the
3547 same. (This required a bit of code, as the output is generated
3548 automatically from a table. It wasn't just a text change.)
3549
3550 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3551 option but starts with a hyphen. Could be a pattern or a path name
3552 starting with a hyphen, for instance.
3553
3554 (h) "-" can be given as a file name to represent stdin.
3555
3556 (i) When file names are being printed, "(standard input)" is used for
3557 the standard input, for compatibility with GNU grep. Previously
3558 "<stdin>" was used.
3559
3560 (j) The option --label=xxx can be used to supply a name to be used for
3561 stdin when file names are being printed. There is no short form.
3562
3563 (k) Re-factored the options decoding logic because we are going to add
3564 two more options that take data. Such options can now be given in four
3565 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3566
3567 (l) Added the -A, -B, and -C options for requesting that lines of context
3568 around matches be printed.
3569
3570 (m) Added the -L option to print the names of files that do not contain
3571 any matching lines, that is, the complement of -l.
3572
3573 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3574 continue to scan other files.
3575
3576 (o) The -s option was incorrectly implemented. For compatibility with other
3577 greps, it now suppresses the error message for a non-existent or non-
3578 accessible file (but not the return code). There is a new option called
3579 -q that suppresses the output of matching lines, which was what -s was
3580 previously doing.
3581
3582 (p) Added --include and --exclude options to specify files for inclusion
3583 and exclusion when recursing.
3584
3585 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3586 Hopefully, it now does.
3587
3588 12. Missing cast in pcre_study().
3589
3590 13. Added an "uninstall" target to the makefile.
3591
3592 14. Replaced "extern" in the function prototypes in Makefile.in with
3593 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3594 world, but is set differently for Windows.
3595
3596 15. Added a second compiling function called pcre_compile2(). The only
3597 difference is that it has an extra argument, which is a pointer to an
3598 integer error code. When there is a compile-time failure, this is set
3599 non-zero, in addition to the error test pointer being set to point to an
3600 error message. The new argument may be NULL if no error number is required
3601 (but then you may as well call pcre_compile(), which is now just a
3602 wrapper). This facility is provided because some applications need a
3603 numeric error indication, but it has also enabled me to tidy up the way
3604 compile-time errors are handled in the POSIX wrapper.
3605
3606 16. Added VPATH=.libs to the makefile; this should help when building with one
3607 prefix path and installing with another. (Or so I'm told by someone who
3608 knows more about this stuff than I do.)
3609
3610 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3611 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3612 match everything, including newlines. This is not POSIX-compatible, but
3613 somebody wanted the feature. From pcretest it can be activated by using
3614 both the P and the s flags.
3615
3616 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3617
3618 19. libpcre.pc was being incorrectly installed as executable.
3619
3620 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3621 it now also looks for '\r' so that it will work unmodified on Windows.
3622
3623 21. Added Google's contributed C++ wrapper to the distribution.
3624
3625 22. Added some untidy missing memory free() calls in pcretest, to keep
3626 Electric Fence happy when testing.
3627
3628
3629
3630 Version 5.0 13-Sep-04
3631 ---------------------
3632
3633 1. Internal change: literal characters are no longer packed up into items
3634 containing multiple characters in a single byte-string. Each character
3635 is now matched using a separate opcode. However, there may be more than one
3636 byte in the character in UTF-8 mode.
3637
3638 2. The pcre_callout_block structure has two new fields: pattern_position and
3639 next_item_length. These contain the offset in the pattern to the next match
3640 item, and its length, respectively.
3641
3642 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3643 insertion of callouts before each pattern item. Added the /C option to
3644 pcretest to make use of this.
3645
3646 4. On the advice of a Windows user, the lines
3647
3648 #if defined(_WIN32) || defined(WIN32)
3649 _setmode( _fileno( stdout ), 0x8000 );
3650 #endif /* defined(_WIN32) || defined(WIN32) */
3651
3652 have been added to the source of pcretest. This apparently does useful
3653 magic in relation to line terminators.
3654
3655 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3656 for the benefit of those environments where the "b" makes a difference.
3657
3658 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3659 to know about it. I have put a hack into configure.in that adds in code
3660 to set GCC=yes if CC=icc. This seems to end up at a point in the
3661 generated configure script that is early enough to affect the setting of
3662 compiler options, which is what is needed, but I have no means of testing
3663 whether it really works. (The user who reported this had patched the
3664 generated configure script, which of course I cannot do.)
3665
3666 LATER: After change 22 below (new libtool files), the configure script
3667 seems to know about icc (and also ecc). Therefore, I have commented out
3668 this hack in configure.in.
3669
3670 7. Added support for pkg-config (2 patches were sent in).
3671
3672 8. Negated POSIX character classes that used a combination of internal tables
3673 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3674 [[:^ascii]]. Typically, they would match almost any characters. The other
3675 POSIX classes were not broken in this way.
3676
3677 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3678 to find the match, as PCRE was deluded into thinking that the match had to
3679 start at the start point or following a newline. The same bug applied to
3680 patterns with negative forward assertions or any backward assertions
3681 preceding ".*" at the start, unless the pattern required a fixed first
3682 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3683
3684 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3685 starting at the last subject character, bytes beyond the end of the subject
3686 string were read.
3687
3688 11. Renamed the variable "class" as "classbits" to make life easier for C++
3689 users. (Previously there was a macro definition, but it apparently wasn't
3690 enough.)
3691
3692 12. Added the new field "tables" to the extra data so that tables can be passed
3693 in at exec time, or the internal tables can be re-selected. This allows
3694 a compiled regex to be saved and re-used at a later time by a different
3695 program that might have everything at different addresses.
3696
3697 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3698 -R library as well as a -L library.
3699
3700 14. The debugging options of pcretest (-d on the command line or D on a
3701 pattern) showed incorrect output for anything following an extended class
3702 that contained multibyte characters and which was followed by a quantifier.
3703
3704 15. Added optional support for general category Unicode character properties
3705 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3706 support. It adds about 90K to the size of the library. The meanings of the
3707 inbuilt class escapes such as \d and \s have NOT been changed.
3708
3709 16. Updated pcredemo.c to include calls to free() to release the memory for the
3710 compiled pattern.
3711
3712 17. The generated file chartables.c was being created in the source directory
3713 instead of in the building directory. This caused the build to fail if the
3714 source directory was different from the building directory, and was
3715 read-only.
3716
3717 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3718 file. No doubt somebody will tell me if they don't make sense... Also added
3719 Dan Mooney's comments about building on OpenVMS.
3720
3721 19. Added support for partial matching via the PCRE_PARTIAL option for
3722 pcre_exec() and the \P data escape in pcretest.
3723
3724 20. Extended pcretest with 3 new pattern features:
3725
3726 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3727 write the compiled pattern to the file whose name is "rest-of-line".
3728 This is a straight binary dump of the data, with the saved pointer to
3729 the character tables forced to be NULL. The study data, if any, is
3730 written too. After writing, pcretest reads a new pattern.
3731
3732 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3733 compiled pattern from the given file. There must not be any
3734 occurrences of "<" in the file name (pretty unlikely); if there are,
3735 pcretest will instead treat the initial "<" as a pattern delimiter.
3736 After reading in the pattern, pcretest goes on to read data lines as
3737 usual.
3738
3739 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3740 and 16-bit fields in a compiled pattern, to simulate a pattern that
3741 was compiled on a host of opposite endianness.
3742
3743 21. The pcre-exec() function can now cope with patterns that were compiled on
3744 hosts of opposite endianness, with this restriction:
3745
3746 As for any compiled expression that is saved and used later, the tables
3747 pointer field cannot be preserved; the extra_data field in the arguments
3748 to pcre_exec() should be used to pass in a tables address if a value
3749 other than the default internal tables were used at compile time.
3750
3751 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3752 now diagnosed as an error. Previously, most of the time, a negative number
3753 would have been treated as zero, but if in addition "ovector" was passed as
3754 NULL, a crash could occur.
3755
3756 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3757 new versions from the libtool 1.5 distribution (the last one is a copy of
3758 a file called libtool.m4). This seems to have fixed the need to patch
3759 "configure" to support Darwin 1.3 (which I used to do). However, I still
3760 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3761 workstation).
3762
3763 24. Changed the PCRE licence to be the more standard "BSD" licence.
3764
3765
3766 Version 4.5 01-Dec-03
3767 ---------------------
3768
3769 1. There has been some re-arrangement of the code for the match() function so
3770 that it can be compiled in a version that does not call itself recursively.
3771 Instead, it keeps those local variables that need separate instances for
3772 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3773 needs to "recurse". Keeping track of where control must go is done by means
3774 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3775 hide most of the details from the main code, and operates only if
3776 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3777 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3778 operating.
3779
3780 To make it easier for callers to provide specially tailored get/free
3781 functions for this usage, two new functions, pcre_stack_malloc, and
3782 pcre_stack_free, are used. They are always called in strict stacking order,
3783 and the size of block requested is always the same.
3784
3785 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3786 PCRE has been compiled to use the stack or the heap for recursion. The
3787 -C option of pcretest uses this to show which version is compiled.
3788
3789 A new data escape \S, is added to pcretest; it causes the amounts of store
3790 obtained and freed by both kinds of malloc/free at match time to be added
3791 to the output.
3792
3793 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3794 what's available on my current Linux desktop machine.
3795
3796 3. When matching a UTF-8 string, the test for a valid string at the start has
3797 been extended. If start_offset is not zero, PCRE now checks that it points
3798 to a byte that is the start of a UTF-8 character. If not, it returns
3799 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3800 this is necessary because there may be backward assertions in the pattern.
3801 When matching the same subject several times, it may save resources to use
3802 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3803
3804 4. The code for checking the validity of UTF-8 strings has been tightened so
3805 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3806 containing "overlong sequences".
3807
3808 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3809 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3810 should have been "&", but it just so happened that all the cases this let
3811 through by mistake were picked up later in the function.
3812
3813 6. I had used a variable called "isblank" - this is a C99 function, causing
3814 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3815
3816 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3817 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3818 so the version is shown; (c) stop "make test" echoing "./RunTest".
3819
3820 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3821
3822 9. The prototype for memmove() for systems that don't have it was using
3823 size_t, but the inclusion of the header that defines size_t was later. I've
3824 moved the #includes for the C headers earlier to avoid this.
3825
3826 10. Added some adjustments to the code to make it easier to compiler on certain
3827 special systems:
3828
3829 (a) Some "const" qualifiers were missing.
3830 (b) Added the macro EXPORT before all exported functions; by default this
3831 is defined to be empty.
3832 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3833 that it reads its output file name as an argument instead of writing
3834 to the standard output and assuming this can be redirected.
3835
3836 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3837 class containing characters with values greater than 255, PCRE compilation
3838 went into a loop.
3839
3840 12. A recursive reference to a subpattern that was within another subpattern
3841 that had a minimum quantifier of zero caused PCRE to crash. For example,
3842 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3843 recursion. If the recursively-called subpattern itself had a zero repeat,
3844 that was OK.
3845
3846 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3847 buffer into which it was copied (for escape processing) was still set at
3848 1024, so long lines caused crashes.
3849
3850 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3851 "internal error: code overflow...". This applied to any character class
3852 that was followed by a possessive quantifier.
3853
3854 15. Modified the Makefile to add libpcre.la as a prerequisite for
3855 libpcreposix.la because I was told this is needed for a parallel build to
3856 work.
3857
3858 16. If a pattern that contained .* following optional items at the start was
3859 studied, the wrong optimizing data was generated, leading to matching
3860 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3861 matching string must start with a or b or c. The correct conclusion for
3862 this pattern is that a match can start with any character.
3863
3864
3865 Version 4.4 13-Aug-03
3866 ---------------------
3867
3868 1. In UTF-8 mode, a character class containing characters with values between
3869 127 and 255 was not handled correctly if the compiled pattern was studied.
3870 In fixing this, I have also improved the studying algorithm for such
3871 classes (slightly).
3872
3873 2. Three internal functions had redundant arguments passed to them. Removal
3874 might give a very teeny performance improvement.
3875
3876 3. Documentation bug: the value of the capture_top field in a callout is *one
3877 more than* the number of the hightest numbered captured substring.
3878
3879 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3880 in incorrectly linking with a previously installed version. They now link
3881 explicitly with libpcre.la.
3882
3883 5. configure.in no longer needs to recognize Cygwin specially.
3884
3885 6. A problem in pcre.in for Windows platforms is fixed.
3886
3887 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3888 pcretest, it used to include the size of the study block as part of its
3889 output. Unfortunately, the structure contains a field that has a different
3890 size on different hardware architectures. This meant that the tests that
3891 showed this size failed. As the block is currently always of a fixed size,
3892 this information isn't actually particularly useful in pcretest output, so
3893 I have just removed it.
3894
3895 8. Three pre-processor statements accidentally did not start in column 1.
3896 Sadly, there are *still* compilers around that complain, even though
3897 standard C has not required this for well over a decade. Sigh.
3898
3899 9. In pcretest, the code for checking callouts passed small integers in the
3900 callout_data field, which is a void * field. However, some picky compilers
3901 complained about the casts involved for this on 64-bit systems. Now
3902 pcretest passes the address of the small integer instead, which should get
3903 rid of the warnings.
3904
3905 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3906 both compile and run time, and gives an error if an invalid UTF-8 sequence
3907 is found. There is a option for disabling this check in cases where the
3908 string is known to be correct and/or the maximum performance is wanted.
3909
3910 11. In response to a bug report, I changed one line in Makefile.in from
3911
3912 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3913 to
3914 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3915
3916 to look similar to other lines, but I have no way of telling whether this
3917 is the right thing to do, as I do not use Windows. No doubt I'll get told
3918 if it's wrong...
3919
3920
3921 Version 4.3 21-May-03
3922 ---------------------
3923
3924 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3925 Makefile.
3926
3927 2. Some refactoring to improve the quality of the code:
3928
3929 (i) The utf8_table... variables are now declared "const".
3930
3931 (ii) The code for \cx, which used the "case flipping" table to upper case
3932 lower case letters, now just substracts 32. This is ASCII-specific,
3933 but the whole concept of \cx is ASCII-specific, so it seems
3934 reasonable.
3935
3936 (iii) PCRE was using its character types table to recognize decimal and
3937 hexadecimal digits in the pattern. This is silly, because it handles
3938 only 0-9, a-f, and A-F, but the character types table is locale-
3939 specific, which means strange things might happen. A private
3940 table is now used for this - though it costs 256 bytes, a table is
3941 much faster than multiple explicit tests. Of course, the standard
3942 character types table is still used for matching digits in subject
3943 strings against \d.
3944
3945 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3946 ending in _t are). So I've renamed it as ESC_tee.
3947
3948 3. The first argument for regexec() in the POSIX wrapper should have been
3949 defined as "const".
3950
3951 4. Changed pcretest to use malloc() for its buffers so that they can be
3952 Electric Fenced for debugging.
3953
3954 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3955 to read one or more bytes before the start of the subject string. Often this
3956 had no effect on PCRE's behaviour, but in some circumstances it could
3957 provoke a segmentation fault.
3958
3959 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3960 to try to read one or more bytes before the start of the subject string.
3961
3962 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3963 UTF-8 support could misbehave in various ways if the subject string
3964 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3965 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3966 back over UTF-8 characters.)
3967
3968
3969 Version 4.2 14-Apr-03
3970 ---------------------
3971
3972 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3973
3974 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3975 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3976 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3977 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3978 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3979 and BUILD_EXEEXT
3980 Note: automatic setting of the BUILD variables is not yet working
3981 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3982 compile-time but not at link-time
3983 [LINK]: use for linking executables only
3984 make different versions for Windows and non-Windows
3985 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3986 libraries
3987 [LINK_FOR_BUILD]: new variable
3988 [OBJEXT]: use throughout
3989 [EXEEXT]: use throughout
3990 <winshared>: new target
3991 <wininstall>: new target
3992 <dftables.o>: use native compiler
3993 <dftables>: use native linker
3994 <install>: handle Windows platform correctly
3995 <clean>: ditto
3996 <check>: ditto
3997 copy DLL to top builddir before testing
3998
3999 As part of these changes, -no-undefined was removed again. This was reported
4000 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4001 in any case.
4002
4003 3. Some tidies to get rid of compiler warnings:
4004
4005 . In the match_data structure, match_limit was an unsigned long int, whereas
4006 match_call_count was an int. I've made them both unsigned long ints.
4007
4008 . In pcretest the fact that a const uschar * doesn't automatically cast to
4009 a void * provoked a warning.
4010
4011 . Turning on some more compiler warnings threw up some "shadow" variables
4012 and a few more missing casts.
4013
4014 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4015 option, a class that contained a single character with a value between 128
4016 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4017
4018 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4019 option, a class that contained several characters, but with at least one
4020 whose value was between 128 and 255 caused PCRE to crash.
4021
4022
4023 Version 4.1 12-Mar-03
4024 ---------------------
4025
4026 1. Compiling with gcc -pedantic found a couple of places where casts were
4027 needed, and a string in dftables.c that was longer than standard compilers are
4028 required to support.
4029
4030 2. Compiling with Sun's compiler found a few more places where the code could
4031 be tidied up in order to avoid warnings.
4032
4033 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4034 first of these names is deprecated in the latest Autoconf in favour of the name
4035 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4036 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4037 analogy I have changed it to CFLAGS_FOR_BUILD.
4038
4039 4. Added -no-undefined to the linking command in the Makefile, because this is
4040 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4041 linking step for the pcreposix library.
4042
4043 5. PCRE was failing to diagnose the case of two named groups with the same
4044 name.
4045
4046 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4047 literal character that is needed in the subject for a match, and scans along to
4048 ensure that it is present before embarking on the full matching process. This
4049 saves time in cases of nested unlimited repeats that are never going to match.
4050 Problem: the scan can take a lot of time if the subject is very long (e.g.
4051 megabytes), thus penalizing straightforward matches. It is now done only if the
4052 amount of subject to be scanned is less than 1000 bytes.
4053
4054 7. A lesser problem with the same optimization is that it was recording the
4055 first character of an anchored pattern as "needed", thus provoking a search
4056 right along the subject, even when the first match of the pattern was going to
4057 fail. The "needed" character is now not set for anchored patterns, unless it
4058 follows something in the pattern that is of non-fixed length. Thus, it still
4059 fulfils its original purpose of finding quick non-matches in cases of nested
4060 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4061
4062
4063 Version 4.0 17-Feb-03
4064 ---------------------
4065
4066 1. If a comment in an extended regex that started immediately after a meta-item
4067 extended to the end of string, PCRE compiled incorrect data. This could lead to
4068 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4069
4070 2. Moved to autoconf 2.53 and libtool 1.4.2.
4071
4072 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4073 the special perltest8 script is no longer needed - all the tests can be run
4074 from a single perltest script.
4075
4076 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4077 by \s. It has now been removed in PCRE. This means it isn't recognized as
4078 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4079 class [:space:] *does* include VT, thereby creating a mess.
4080
4081 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4082 space and tab.
4083
4084 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4085 its new features into the main test script, reducing the number of scripts.
4086
4087 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4088 were backward compatible, and made the (?i) apply to the whole pattern, as if
4089 /i were given. Now it behaves more logically, and applies the option setting
4090 only to what follows. PCRE has been changed to follow suit. However, if it
4091 finds options settings right at the start of the pattern, it extracts them into
4092 the global options, as before. Thus, they show up in the info data.
4093
4094 8. Added support for the \Q...\E escape sequence. Characters in between are
4095 treated as literals. This is slightly different from Perl in that $ and @ are
4096 also handled as literals inside the quotes. In Perl, they will cause variable
4097 interpolation. Note the following examples:
4098
4099 Pattern PCRE matches Perl matches
4100
4101 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4102 \Qabc\$xyz\E abc\$xyz abc\$xyz
4103 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4104
4105 For compatibility with Perl, \Q...\E sequences are recognized inside character
4106 classes as well as outside them.
4107
4108 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4109 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4110 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4111 signed/unsigned warnings.
4112
4113 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4114 option for pcretest, so I've replaced it by a simple function that does just
4115 that job.
4116
4117 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4118 "pcregrep -".
4119
4120 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4121 Java package. This provides some syntactic sugar for simple cases of what my
4122 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4123 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4124 item, you can use this simplified notation. Note that only makes sense with
4125 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4126 greediness, whatever the setting of the PCRE_UNGREEDY option.
4127
4128 13. A change of greediness default within a pattern was not taking effect at
4129 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4130 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4131 was abstracted outside.
4132
4133 14. PCRE now supports the \G assertion. It is true when the current matching
4134 position is at the start point of the match. This differs from \A when the
4135 starting offset is non-zero. Used with the /g option of pcretest (or similar
4136 code), it works in the same way as it does for Perl's /g option. If all
4137 alternatives of a regex begin with \G, the expression is anchored to the start
4138 match position, and the "anchored" flag is set in the compiled expression.
4139
4140 15. Some bugs concerning the handling of certain option changes within patterns
4141 have been fixed. These applied to options other than (?ims). For example,
4142 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4143 been the other way round. Some of this was related to change 7 above.
4144
4145 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4146 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4147 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4148 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4149
4150 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4151 mode. Unlike ".", it always matches newline, whatever the setting of
4152 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4153 assertions. Perl allows it, but it doesn't (in general) work because it can't
4154 calculate the length of the lookbehind. At least, that's the case for Perl
4155 5.8.0 - I've been told they are going to document that it doesn't work in
4156 future.
4157
4158 18. Added an error diagnosis for escapes that PCRE does not support: these are
4159 \L, \l, \N, \P, \p, \U, \u, and \X.
4160
4161 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4162 reading past the end of the pattern in cases such as /[abcd/.
4163
4164 20. PCRE was getting more memory than necessary for patterns with classes that
4165 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4166
4167 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4168 compiling PCRE for use with Virtual Pascal.
4169
4170 22. Small fix to the Makefile to make it work properly if the build is done
4171 outside the source tree.
4172
4173 23. Added a new extension: a condition to go with recursion. If a conditional
4174 subpattern starts with (?(R) the "true" branch is used if recursion has
4175 happened, whereas the "false" branch is used only at the top level.
4176
4177 24. When there was a very long string of literal characters (over 255 bytes
4178 without UTF support, over 250 bytes with UTF support), the computation of how
4179 much memory was required could be incorrect, leading to segfaults or other
4180 strange effects.
4181
4182 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4183 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4184 there was a subsequent back reference to those brackets. This meant that, for
4185 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4186 possible to check for precisely this case. All we can do is abandon the
4187 optimization if .* occurs inside capturing brackets when there are any back
4188 references whatsoever. (See below for a better fix that came later.)
4189
4190 26. The handling of the optimization for finding the first character of a
4191 non-anchored pattern, and for finding a character that is required later in the
4192 match were failing in some cases. This didn't break the matching; it just
4193 failed to optimize when it could. The way this is done has been re-implemented.
4194
4195 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4196
4197 28. Added a new feature that provides some of the functionality that Perl
4198 provides with (?{...}). The facility is termed a "callout". The way it is done
4199 in PCRE is for the caller to provide an optional function, by setting
4200 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4201 global variable. By default it is unset, which disables all calling out. To get
4202 the function called, the regex must include (?C) at appropriate points. This
4203 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4204 This provides a means of identifying different callout points. When PCRE
4205 reaches such a point in the regex, if pcre_callout has been set, the external
4206 function is called. It is provided with data in a structure called
4207 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4208 matching continues; if it returns a non-zero value, the match at the current
4209 point fails. However, backtracking will occur if possible. [This was changed
4210 later and other features added - see item 49 below.]
4211
4212 29. pcretest is upgraded to test the callout functionality. It provides a
4213 callout function that displays information. By default, it shows the start of
4214 the match and the current position in the text. There are some new data escapes
4215 to vary what happens:
4216
4217 \C+ in addition, show current contents of captured substrings
4218 \C- do not supply a callout function
4219 \C!n return 1 when callout number n is reached
4220 \C!n!m return 1 when callout number n is reached for the mth time
4221
4222 30. If pcregrep was called with the -l option and just a single file name, it
4223 output "<stdin>" if a match was found, instead of the file name.
4224
4225 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4226 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4227 pcre_exec(). This saves a malloc/free per call. The default value of
4228 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4229 when configuring.
4230
4231 32. The default maximum size of a compiled pattern is 64K. There have been a
4232 few cases of people hitting this limit. The code now uses macros to handle the
4233 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4234 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4235 configuring. Tests 2 and 5 work only with 2-byte links because they output
4236 debugging information about compiled patterns.
4237
4238 33. Internal code re-arrangements:
4239
4240 (a) Moved the debugging function for printing out a compiled regex into
4241 its own source file (printint.c) and used #include to pull it into
4242 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4243 separate copies.
4244
4245 (b) Defined the list of op-code names for debugging as a macro in
4246 internal.h so that it is next to the definition of the opcodes.
4247
4248 (c) Defined a table of op-code lengths for simpler skipping along compiled
4249 code. This is again a macro in internal.h so that it is next to the
4250 definition of the opcodes.
4251
4252 34. Added support for recursive calls to individual subpatterns, along the
4253 lines of Robin Houston's patch (but implemented somewhat differently).
4254
4255 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4256 allow it to read and process whole directories in Win32. This code was
4257 contributed by Lionel Fourquaux; it has not been tested by me.
4258
4259 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4260 used to name a group. Names consist of alphanumerics and underscores, and must
4261 be unique. Back references use the syntax (?P=name) and recursive calls use
4262 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4263 numbers. The function pcre_fullinfo() can be used after compilation to extract
4264 a name/number map. There are three relevant calls:
4265
4266 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4267 PCRE_INFO_NAMECOUNT yields the number of entries
4268 PCRE_INFO_NAMETABLE yields a pointer to the map.
4269
4270 The map is a vector of fixed-size entries. The size of each entry depends on
4271 the length of the longest name used. The first two bytes of each entry are the
4272 group number, most significant byte first. There follows the corresponding
4273 name, zero terminated. The names are in alphabetical order.
4274
4275 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4276 case instead of 255. Making it the same both with and without UTF-8 support
4277 means that the same test output works with both.
4278
4279 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4280 calling malloc() with a zero argument.
4281
4282 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4283 optimization. I've improved things by keeping a bitmap of backreferences with
4284 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4285 fact referenced, the optimization can be applied. It is unlikely that a
4286 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4287 the match to follow \n) will appear inside brackets with a number greater than
4288 31, but if it does, any back reference > 31 suppresses the optimization.
4289
4290 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4291 of disabling numbered capturing parentheses. Any opening parenthesis that is
4292 not followed by ? behaves as if it were followed by ?: but named parentheses
4293 can still be used for capturing (and they will acquire numbers in the usual
4294 way).
4295
4296 41. Redesigned the return codes from the match() function into yes/no/error so
4297 that errors can be passed back from deep inside the nested calls. A malloc
4298 failure while inside a recursive subpattern call now causes the
4299 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4300
4301 42. It is now possible to set a limit on the number of times the match()
4302 function is called in a call to pcre_exec(). This facility makes it possible to
4303 limit the amount of recursion and backtracking, though not in a directly
4304 obvious way, because the match() function is used in a number of different
4305 circumstances. The count starts from zero for each position in the subject
4306 string (for non-anchored patterns). The default limit is, for compatibility, a
4307 large number, namely 10 000 000. You can change this in two ways:
4308
4309 (a) When configuring PCRE before making, you can use --with-match-limit=n
4310 to set a default value for the compiled library.
4311
4312 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4313 a different value is set. See 45 below.
4314
4315 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4316
4317 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4318 of things that can be changed at compile time. The first argument specifies
4319 what is wanted and the second points to where the information is to be placed.
4320 The current list of available information is:
4321
4322 PCRE_CONFIG_UTF8
4323
4324 The output is an integer that is set to one if UTF-8 support is available;
4325 otherwise it is set to zero.
4326
4327 PCRE_CONFIG_NEWLINE
4328
4329 The output is an integer that it set to the value of the code that is used for
4330 newline. It is either LF (10) or CR (13).
4331
4332 PCRE_CONFIG_LINK_SIZE
4333
4334 The output is an integer that contains the number of bytes used for internal
4335 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4336
4337 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4338
4339 The output is an integer that contains the threshold above which the POSIX
4340 interface uses malloc() for output vectors. See item 31 above.
4341
4342 PCRE_CONFIG_MATCH_LIMIT
4343
4344 The output is an unsigned integer that contains the default limit of the number
4345 of match() calls in a pcre_exec() execution. See 42 above.
4346
4347 44. pcretest has been upgraded by the addition of the -C option. This causes it
4348 to extract all the available output from the new pcre_config() function, and to
4349 output it. The program then exits immediately.
4350
4351 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4352 order to support additional features. One way would have been to define
4353 pcre_exec2() (for example) with extra arguments, but this would not have been
4354 extensible, and would also have required all calls to the original function to
4355 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4356 is used for passing in "extra" data from pcre_study().
4357
4358 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4359 contains the following fields:
4360
4361 flags a bitmap indicating which of the following fields are set
4362 study_data opaque data from pcre_study()
4363 match_limit a way of specifying a limit on match() calls for a specific
4364 call to pcre_exec()
4365 callout_data data for callouts (see 49 below)
4366
4367 The flag bits are also defined in pcre.h, and are
4368
4369 PCRE_EXTRA_STUDY_DATA
4370 PCRE_EXTRA_MATCH_LIMIT
4371 PCRE_EXTRA_CALLOUT_DATA
4372
4373 The pcre_study() function now returns one of these new pcre_extra blocks, with
4374 the actual study data pointed to by the study_data field, and the
4375 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4376 before. That is, this change is entirely upwards-compatible and requires no
4377 change to existing code.
4378
4379 If you want to pass in additional data to pcre_exec(), you can either place it
4380 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4381 block.
4382
4383 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4384 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4385 times with different match limits, until it finds the minimum value needed for
4386 pcre_exec() to complete. The value is then output. This can be instructive; for
4387 most simple matches the number is quite small, but for pathological cases it
4388 gets very large very quickly.
4389
4390 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4391 returns the size of the data block pointed to by the study_data field in a
4392 pcre_extra block, that is, the value that was passed as the argument to
4393 pcre_malloc() when PCRE was getting memory in which to place the information
4394 created by pcre_study(). The fourth argument should point to a size_t variable.
4395 pcretest has been extended so that this information is shown after a successful
4396 pcre_study() call when information about the compiled regex is being displayed.
4397
4398 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4399 because what follows is always an absolute path. (Later: it turns out that this
4400 is more than cosmetic for MinGW, because it doesn't like empty path
4401 components.)
4402
4403 49. Some changes have been made to the callout feature (see 28 above):
4404
4405 (i) A callout function now has three choices for what it returns:
4406
4407 0 => success, carry on matching
4408 > 0 => failure at this point, but backtrack if possible
4409 < 0 => serious error, return this value from pcre_exec()
4410
4411 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4412 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4413 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4414 use by callout functions. It will never be used by PCRE itself.
4415
4416 (ii) The pcre_extra structure (see 45 above) has a void * field called
4417 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4418 pcre_callout_block structure has a field of the same name. The contents of
4419 the field passed in the pcre_extra structure are passed to the callout
4420 function in the corresponding field in the callout block. This makes it
4421 easier to use the same callout-containing regex from multiple threads. For
4422 testing, the pcretest program has a new data escape
4423
4424 \C*n pass the number n (may be negative) as callout_data
4425
4426 If the callout function in pcretest receives a non-zero value as
4427 callout_data, it returns that value.
4428
4429 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4430 there were some redundant $(CFLAGS) in commands that are now specified as
4431 $(LINK), which already includes $(CFLAGS).
4432
4433 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4434 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4435 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4436 one-byte characters throughout. Note that case-insensitive matching applies
4437 only to characters whose values are less than 256. PCRE doesn't support the
4438 notion of cases for higher-valued characters.
4439
4440 (i) A character class whose characters are all within 0-255 is handled as
4441 a bit map, and the map is inverted for negative classes. Previously, a
4442 character > 255 always failed to match such a class; however it should
4443 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4444
4445 (ii) A negated character class with a single character < 255 is coded as
4446 "not this character" (OP_NOT). This wasn't working properly when the test
4447 character was multibyte, either singly or repeated.
4448
4449 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4450 mode, for example: \x{100}{2,3}.
4451
4452 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4453 singly or repeated) now correctly test multibyte characters. However,
4454 PCRE doesn't recognize any characters with values greater than 255 as
4455 digits, spaces, or word characters. Such characters always match \D, \S,
4456 and \W, and never match \d, \s, or \w.
4457
4458 (v) Classes may now contain characters and character ranges with values
4459 greater than 255. For example: [ab\x{100}-\x{400}].
4460
4461 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4462 PCRE in UTF-8 mode.
4463
4464 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4465 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4466 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4467 value.)
4468
4469 53. The single man page has become too large. I have therefore split it up into
4470 a number of separate man pages. These also give rise to individual HTML pages;
4471 these are now put in a separate directory, and there is an index.html page that
4472 lists them all. Some hyperlinking between the pages has been installed.
4473
4474 54. Added convenience functions for handling named capturing parentheses.
4475
4476 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4477 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4478 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4479 are faulted.
4480
4481 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4482 calling configure. These values are used when compiling the dftables.c program
4483 which is run to generate the source of the default character tables. They
4484 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4485 you will need to set these values.
4486
4487 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4488
4489
4490 Version 3.9 02-Jan-02
4491 ---------------------
4492
4493 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4494
4495 2. If --disable-static was given, the building process failed when trying to
4496 build pcretest and pcregrep. (For some reason it was using libtool to compile
4497 them, which is not right, as they aren't part of the library.)
4498
4499
4500 Version 3.8 18-Dec-01
4501 ---------------------
4502
4503 1. The experimental UTF-8 code was completely screwed up. It was packing the
4504 bytes in the wrong order. How dumb can you get?
4505
4506
4507 Version 3.7 29-Oct-01
4508 ---------------------
4509
4510 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4511 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4512 this didn't happen under Solaris 8, where I normally test things.
4513
4514 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4515 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4516 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4517 this fix an hour or so after the initial 3.7 release.)
4518
4519
4520 Version 3.6 23-Oct-01
4521 ---------------------
4522
4523 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4524 offsets passed as NULL with zero offset count.
4525
4526 2. The config.guess and config.sub files had not been updated when I moved to
4527 the latest autoconf.
4528
4529
4530 Version 3.5 15-Aug-01
4531 ---------------------
4532
4533 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4534 had been forgotten.
4535
4536 2. By using declared but undefined structures, we can avoid using "void"
4537 definitions in pcre.h while keeping the internal definitions of the structures
4538 private.
4539
4540 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4541 user point of view, this means that both static and shared libraries are built
4542 by default, but this can be individually controlled. More of the work of
4543 handling this static/shared cases is now inside libtool instead of PCRE's make
4544 file.
4545
4546 4. The pcretest utility is now installed along with pcregrep because it is
4547 useful for users (to test regexs) and by doing this, it automatically gets
4548 relinked by libtool. The documentation has been turned into a man page, so
4549 there are now .1, .txt, and .html versions in /doc.
4550
4551 5. Upgrades to pcregrep:
4552 (i) Added long-form option names like gnu grep.
4553 (ii) Added --help to list all options with an explanatory phrase.
4554 (iii) Added -r, --recursive to recurse into sub-directories.
4555 (iv) Added -f, --file to read patterns from a file.
4556
4557 6. pcre_exec() was referring to its "code" argument before testing that
4558 argument for NULL (and giving an error if it was NULL).
4559
4560 7. Upgraded Makefile.in to allow for compiling in a different directory from
4561 the source directory.
4562
4563 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4564 options bits, the pointer it was passed was to an int instead of to an unsigned
4565 long int. This mattered only on 64-bit systems.
4566
4567 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4568 generated) instead of pcre.in, which it its source. Also made the same change
4569 in several of the .c files.
4570
4571 10. A new release of gcc defines printf() as a macro, which broke pcretest
4572 because it had an ifdef in the middle of a string argument for printf(). Fixed
4573 by using separate calls to printf().
4574
4575 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4576 script, to force use of CR or LF instead of \n in the source. On non-Unix
4577 systems, the value can be set in config.h.
4578
4579 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4580 absolute limit. Changed the text of the error message to make this clear, and
4581 likewise updated the man page.
4582
4583 13. The limit of 99 on the number of capturing subpatterns has been removed.
4584 The new limit is 65535, which I hope will not be a "real" limit.
4585
4586
4587 Version 3.4 22-Aug-00
4588 ---------------------
4589
4590 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4591
4592 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4593
4594
4595 Version 3.3 01-Aug-00
4596 ---------------------
4597
4598 1. If an octal character was given, but the value was greater than \377, it
4599 was not getting masked to the least significant bits, as documented. This could
4600 lead to crashes in some systems.
4601
4602 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4603 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4604
4605 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4606 These just pass their arguments on to (pcre_free)(), but they are provided
4607 because some uses of PCRE bind it to non-C systems that can call its functions,
4608 but cannot call free() or pcre_free() directly.
4609
4610 4. Add "make test" as a synonym for "make check". Corrected some comments in
4611 the Makefile.
4612
4613 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4614 Makefile.
4615
4616 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4617 command called pgrep for grepping around the active processes.
4618
4619 7. Added the beginnings of support for UTF-8 character strings.
4620
4621 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4622 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4623 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4624 out for the ar command.)
4625
4626
4627 Version 3.2 12-May-00
4628 ---------------------
4629
4630 This is purely a bug fixing release.
4631
4632 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4633 of ZA. This was just one example of several cases that could provoke this bug,
4634 which was introduced by change 9 of version 2.00. The code for breaking
4635 infinite loops after an iteration that matches an empty string was't working
4636 correctly.
4637
4638 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4639 when matched against abbab (for example). After matching an empty string, it
4640 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4641 caused it to match further down the string than it should.
4642
4643 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4644 was there because it doesn't seem to be needed, and it causes trouble on some
4645 systems, as it is not a Standard C header. It has been removed.
4646
4647 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4648 were reported on the Macintosh. The changes were from
4649
4650 while ((c = *(++ptr)) != 0 && c != '\n');
4651 to
4652 while ((c = *(++ptr)) != 0 && c != '\n') ;
4653
4654 Totally extraordinary, but if that's what it takes...
4655
4656 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4657 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4658 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4659 assumes the way PCRE uses memmove() (always moving upwards).
4660
4661 6. PCRE is being used in one environment where strchr() is not available. There
4662 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4663 faster code anyway.
4664
4665
4666 Version 3.1 09-Feb-00
4667 ---------------------
4668
4669 The only change in this release is the fixing of some bugs in Makefile.in for
4670 the "install" target:
4671
4672 (1) It was failing to install pcreposix.h.
4673
4674 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4675
4676
4677 Version 3.0 01-Feb-00
4678 ---------------------
4679
4680 1. Add support for the /+ modifier to perltest (to output $` like it does in
4681 pcretest).
4682
4683 2. Add support for the /g modifier to perltest.
4684
4685 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4686 matches null strings.
4687
4688 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4689 pattern. Perl treats empty patterns specially - it reuses the most recent
4690 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4691 effect.
4692
4693 5. The POSIX interface was broken in that it was just handing over the POSIX
4694 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4695 required a bigger vector, with some working space on the end. This means that
4696 the POSIX wrapper now has to get and free some memory, and copy the results.
4697
4698 6. Added some simple autoconf support, placing the test data and the
4699 documentation in separate directories, re-organizing some of the
4700 information files, and making it build pcre-config (a GNU standard). Also added
4701 libtool support for building PCRE as a shared library, which is now the
4702 default.
4703
4704 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4705 09 are not valid octal constants. Single digits will be used for minor values
4706 less than 10.
4707
4708 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4709 existing programs that set these in the POSIX interface can use PCRE without
4710 modification.
4711
4712 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4713 return all that pcre_info() returns, plus additional data. The pcre_info()
4714 function is retained for compatibility, but is considered to be obsolete.
4715
4716 10. Added experimental recursion feature (?R) to handle one common case that
4717 Perl 5.6 will be able to do with (?p{...}).
4718
4719 11. Added support for POSIX character classes like [:alpha:], which Perl is
4720 adopting.
4721
4722
4723 Version 2.08 31-Aug-99
4724 ----------------------
4725
4726 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4727 trying to match at the startoffset position, but instead was moving forward to
4728 the next newline as if a previous match had failed.
4729
4730 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4731 and could get into a loop if a null string was matched other than at the start
4732 of the subject.
4733
4734 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4735 be distinguished at compile time, and for completeness also added PCRE_DATE.
4736
4737 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4738 in GnuWin32 environments.
4739
4740
4741 Version 2.07 29-Jul-99
4742 ----------------------
4743
4744 1. The documentation is now supplied in plain text form and HTML as well as in
4745 the form of man page sources.
4746
4747 2. C++ compilers don't like assigning (void *) values to other pointer types.
4748 In particular this affects malloc(). Although there is no problem in Standard
4749 C, I've put in casts to keep C++ compilers happy.
4750
4751 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4752 should be (const char *).
4753
4754 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4755 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4756 However, I haven't made this a standard facility. The documentation doesn't
4757 mention it, and the Makefile doesn't support it.
4758
4759 5. The Makefile now contains an "install" target, with editable destinations at
4760 the top of the file. The pcretest program is not installed.
4761
4762 6. pgrep -V now gives the PCRE version number and date.
4763
4764 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4765 causing the entire string to be ignored, instead of just the last character.
4766
4767 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4768 non-matching string, it can take a very, very long time, even for strings of
4769 quite modest length, because of the nested recursion. PCRE now does better in
4770 some of these cases. It does this by remembering the last required literal
4771 character in the pattern, and pre-searching the subject to ensure it is present
4772 before running the real match. In other words, it applies a heuristic to detect
4773 some types of certain failure quickly, and in the above example, if presented
4774 with a string that has no trailing " it gives "no match" very quickly.
4775
4776 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4777 other alternatives are tried instead.
4778
4779
4780 Version 2.06 09-Jun-99
4781 ----------------------
4782
4783 1. Change pcretest's output for amount of store used to show just the code
4784 space, because the remainder (the data block) varies in size between 32-bit and
4785 64-bit systems.
4786
4787 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4788 start matching at. This allows lookbehinds to work when searching for multiple
4789 occurrences in a string.
4790
4791 3. Added additional options to pcretest for testing multiple occurrences:
4792
4793 /+ outputs the rest of the string that follows a match
4794 /g loops for multiple occurrences, using the new startoffset argument
4795 /G loops for multiple occurrences by passing an incremented pointer
4796
4797 4. PCRE wasn't doing the "first character" optimization for patterns starting
4798 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4799 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4800 the letter 'x'. On long subject strings, this gives a significant speed-up.
4801
4802
4803 Version 2.05 21-Apr-99
4804 ----------------------
4805
4806 1. Changed the type of magic_number from int to long int so that it works
4807 properly on 16-bit systems.
4808
4809 2. Fixed a bug which caused patterns starting with .* not to work correctly
4810 when the subject string contained newline characters. PCRE was assuming
4811 anchoring for such patterns in all cases, which is not correct because .* will
4812 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4813 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4814 must be retried after every newline in the subject.
4815
4816
4817 Version 2.04 18-Feb-99
4818 ----------------------
4819
4820 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4821 computation of the store needed to hold the pattern was incorrect (too large).
4822 If such patterns were nested a few deep, this could multiply and become a real
4823 problem.
4824
4825 2. Added /M option to pcretest to show the memory requirement of a specific
4826 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4827
4828 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4829 compiled in such a way that the backtracking after subsequent failure was
4830 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4831 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4832
4833
4834 Version 2.03 02-Feb-99
4835 ----------------------
4836
4837 1. Fixed typo and small mistake in man page.
4838
4839 2. Added 4th condition (GPL supersedes if conflict) and created separate
4840 LICENCE file containing the conditions.
4841
4842 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4843 Perl, that is the internal \ allows the delimiter to be included in the
4844 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4845 the final delimiter, add \ to the end of the pattern (to test the error).
4846
4847 4. Added the convenience functions for extracting substrings after a successful
4848 match. Updated pcretest to make it able to test these functions.
4849
4850
4851 Version 2.02 14-Jan-99
4852 ----------------------
4853
4854 1. Initialized the working variables associated with each extraction so that
4855 their saving and restoring doesn't refer to uninitialized store.
4856
4857 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4858 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4859 fix the problem.
4860
4861 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4862 calls, and wasn't printing the correct value for compiling calls. Increased the
4863 default value of LOOPREPEAT, and the number of significant figures in the
4864 times.
4865
4866 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4867
4868 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4869 a building problem on Windows NT with a FAT file system.
4870
4871
4872 Version 2.01 21-Oct-98
4873 ----------------------
4874
4875 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4876 to character tables built by pcre_maketables() in the current locale. If NULL
4877 is passed, the default tables are used.
4878
4879
4880 Version 2.00 24-Sep-98
4881 ----------------------
4882
4883 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4884 it any more.
4885
4886 2. Allow quantification of (?>) groups, and make it work correctly.
4887
4888 3. The first character computation wasn't working for (?>) groups.
4889
4890 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4891 end of the subject) and add 5.005's \z, which really does match only at the
4892 very end of the subject.
4893
4894 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4895
4896 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4897 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4898 localized options. All options to pcre_study() were also removed.
4899
4900 7. Add other new features from 5.005:
4901
4902 $(?<= positive lookbehind
4903 $(?<! negative lookbehind
4904 (?imsx-imsx) added the unsetting capability
4905 such a setting is global if at outer level; local otherwise
4906 (?imsx-imsx:) non-capturing groups with option setting
4907 (?(cond)re|re) conditional pattern matching
4908
4909 A backreference to itself in a repeated group matches the previous
4910 captured string.
4911
4912 8. General tidying up of studying (both automatic and via "study")
4913 consequential on the addition of new assertions.
4914
4915 9. As in 5.005, unlimited repeated groups that could match an empty substring
4916 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4917 runtime if any iteration does actually match an empty substring.
4918
4919 10. Include the RunTest script in the distribution.
4920
4921 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4922 discrepancies, some of which were old and were also with respect to 5.004. They
4923 have now been fixed.
4924
4925
4926 Version 1.09 28-Apr-98
4927 ----------------------
4928
4929 1. A negated single character class followed by a quantifier with a minimum
4930 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4931 program crashes, or just wrong answers. This did not apply to negated classes
4932 containing more than one character, or to minima other than one.
4933
4934
4935 Version 1.08 27-Mar-98
4936 ----------------------
4937
4938 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4939
4940 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4941 latter must appear before anything that relies on it in the pattern.
4942
4943
4944 Version 1.07 16-Feb-98
4945 ----------------------
4946
4947 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4948 repeat of a potentially empty string).
4949
4950
4951 Version 1.06 23-Jan-98
4952 ----------------------
4953
4954 1. Added Markus Oberhumer's little patches for C++.
4955
4956 2. Literal strings longer than 255 characters were broken.
4957
4958
4959 Version 1.05 23-Dec-97
4960 ----------------------
4961
4962 1. Negated character classes containing more than one character were failing if
4963 PCRE_CASELESS was set at run time.
4964
4965
4966 Version 1.04 19-Dec-97
4967 ----------------------
4968
4969 1. Corrected the man page, where some "const" qualifiers had been omitted.
4970
4971 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4972 input syntax.
4973
4974 3. Fixed memory leak which occurred when a regex with back references was
4975 matched with an offsets vector that wasn't big enough. The temporary memory
4976 that is used in this case wasn't being freed if the match failed.
4977
4978 4. Tidied pcretest to ensure it frees memory that it gets.
4979
4980 5. Temporary memory was being obtained in the case where the passed offsets
4981 vector was exactly big enough.
4982
4983 6. Corrected definition of offsetof() from change 5 below.
4984
4985 7. I had screwed up change 6 below and broken the rules for the use of
4986 setjmp(). Now fixed.
4987
4988
4989 Version 1.03 18-Dec-97
4990 ----------------------
4991
4992 1. A erroneous regex with a missing opening parenthesis was correctly
4993 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4994 on some systems.
4995
4996 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4997 it was reported that one broken compiler failed on the former because "code" is
4998 also an independent variable.
4999
5000 3. The erroneous regex a[]b caused an array overrun reference.
5001
5002 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5003 fail on data ending with that character. (It was going on too far, and checking
5004 the next character, typically a binary zero.) This was specific to the
5005 optimized code for single-character negative classes.
5006
5007 5. Added a contributed patch from the TIN world which does the following:
5008
5009 + Add an undef for memmove, in case the the system defines a macro for it.
5010
5011 + Add a definition of offsetof(), in case there isn't one. (I don't know
5012 the reason behind this - offsetof() is part of the ANSI standard - but
5013 it does no harm).
5014
5015 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5016 most of the places where whitespace preceded '#'. I have given up and
5017 allowed the remaining 2 cases to be at the margin.
5018
5019 + Rename some variables in pcre to eliminate shadowing. This seems very
5020 pedantic, but does no harm, of course.
5021
5022 6. Moved the call to setjmp() into its own function, to get rid of warnings
5023 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5024
5025 7. Constructs such as \d{8,} were compiling into the equivalent of
5026 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5027 outcome, but in this particular case used more store than had been allocated,
5028 which caused the bug to be discovered because it threw up an internal error.
5029
5030 8. The debugging code in both pcre and pcretest for outputting the compiled
5031 form of a regex was going wrong in the case of back references followed by
5032 curly-bracketed repeats.
5033
5034
5035 Version 1.02 12-Dec-97
5036 ----------------------
5037
5038 1. Typos in pcre.3 and comments in the source fixed.
5039
5040 2. Applied a contributed patch to get rid of places where it used to remove
5041 'const' from variables, and fixed some signed/unsigned and uninitialized
5042 variable warnings.
5043
5044 3. Added the "runtest" target to Makefile.
5045
5046 4. Set default compiler flag to -O2 rather than just -O.
5047
5048
5049 Version 1.01 19-Nov-97
5050 ----------------------
5051
5052 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5053 like /([ab]*)*/, that is, for classes with more than one character in them.
5054
5055 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5056 as /((?>a*))*/ (a PCRE_EXTRA facility).
5057
5058
5059 Version 1.00 18-Nov-97
5060 ----------------------
5061
5062 1. Added compile-time macros to support systems such as SunOS4 which don't have
5063 memmove() or strerror() but have other things that can be used instead.
5064
5065 2. Arranged that "make clean" removes the executables.
5066
5067
5068 Version 0.99 27-Oct-97
5069 ----------------------
5070
5071 1. Fixed bug in code for optimizing classes with only one character. It was
5072 initializing a 32-byte map regardless, which could cause it to run off the end
5073 of the memory it had got.
5074
5075 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5076
5077
5078 Version 0.98 22-Oct-97
5079 ----------------------
5080
5081 1. Fixed bug in code for handling temporary memory usage when there are more
5082 back references than supplied space in the ovector. This could cause segfaults.
5083
5084
5085 Version 0.97 21-Oct-97
5086 ----------------------
5087
5088 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5089
5090 2. Optimized negated single characters not to use a bit map.
5091
5092 3. Brought error texts together as macro definitions; clarified some of them;
5093 fixed one that was wrong - it said "range out of order" when it meant "invalid
5094 escape sequence".
5095
5096 4. Changed some char * arguments to const char *.
5097
5098 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5099
5100 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5101 pcretest.
5102
5103
5104 Version 0.96 16-Oct-97
5105 ----------------------
5106
5107 1. Added a simple "pgrep" utility to the distribution.
5108
5109 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5110 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5111 where "ddd" means "one or more decimal digits".
5112
5113 3. Fixed serious bug. If a pattern had a back reference, but the call to
5114 pcre_exec() didn't supply a large enough ovector to record the related
5115 identifying subpattern, the match always failed. PCRE now remembers the number
5116 of the largest back reference, and gets some temporary memory in which to save
5117 the offsets during matching if necessary, in order to ensure that
5118 backreferences always work.
5119
5120 4. Increased the compatibility with Perl in a number of ways:
5121
5122 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5123 to request this handling. The option can be set at compile or exec time.
5124
5125 (b) $ matches before a terminating newline by default; an option
5126 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5127 mode). The option can be set at compile or exec time.
5128
5129 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5130 the same as Perl's. If the decimal number it represents is less than 10
5131 or there aren't that many previous left capturing parentheses, an octal
5132 escape is read. Inside a character class, it's always an octal escape,
5133 even if it is a single digit.
5134
5135 (d) An escaped but undefined alphabetic character is taken as a literal,
5136 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5137 escapes.
5138
5139 (e) {0} is now permitted. (The previous item is removed from the compiled
5140 pattern).
5141
5142 5. Changed all the names of code files so that the basic parts are no longer
5143 than 10 characters, and abolished the teeny "globals.c" file.
5144
5145 6. Changed the handling of character classes; they are now done with a 32-byte
5146 bit map always.
5147
5148 7. Added the -d and /D options to pcretest to make it possible to look at the
5149 internals of compilation without having to recompile pcre.
5150
5151
5152 Version 0.95 23-Sep-97
5153 ----------------------
5154
5155 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5156 \x20 at the start of a run of normal characters. These were being treated as
5157 real characters, instead of the source characters being re-checked.
5158
5159
5160 Version 0.94 18-Sep-97
5161 ----------------------
5162
5163 1. The functions are now thread-safe, with the caveat that the global variables
5164 containing pointers to malloc() and free() or alternative functions are the
5165 same for all threads.
5166
5167 2. Get pcre_study() to generate a bitmap of initial characters for non-
5168 anchored patterns when this is possible, and use it if passed to pcre_exec().
5169
5170
5171 Version 0.93 15-Sep-97
5172 ----------------------
5173
5174 1. /(b)|(:+)/ was computing an incorrect first character.
5175
5176 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5177 but not actually doing anything yet.
5178
5179 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5180 as Perl does (e.g. [-az] or [az-]).
5181
5182 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5183 all possible positions.
5184
5185 5. Split up into different modules to avoid including unneeded functions in a
5186 compiled binary. However, compile and exec are still in one module. The "study"
5187 function is split off.
5188
5189 6. The character tables are now in a separate module whose source is generated
5190 by an auxiliary program - but can then be edited by hand if required. There are
5191 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5192 toupper() in the code.
5193
5194 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5195 make them global. Abolish the function for setting them, as the caller can now
5196 set them directly.
5197
5198
5199 Version 0.92 11-Sep-97
5200 ----------------------
5201
5202 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5203 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5204
5205 2. Caseless matching was not working in character classes if the characters in
5206 the pattern were in upper case.
5207
5208 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5209
5210 4. Make PCRE_ANCHORED public and accept as a compile option.
5211
5212 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5213 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5214 pass them.
5215
5216 6. Give an error if bad option bits passed at compile or run time.
5217
5218 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5219 pcretest to cause it to pass that flag.
5220
5221 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5222 options, and the first character, if set.
5223
5224 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5225
5226
5227 Version 0.91 10-Sep-97
5228 ----------------------
5229
5230 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5231 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5232
5233 2. PCRE was looping on encountering an indefinitely repeated back reference to
5234 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5235 Perl does - treats the match as successful.
5236
5237 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5