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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1366 - (show annotations)
Sun Oct 6 18:34:33 2013 UTC (6 years, 2 months ago) by ph10
File size: 252712 byte(s)
Forgot to commit this for the previous commit.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.34 xx-xxxx-201x
5 --------------------------
6
7 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8 executable memory. Patch inspired by Carsten Klein.
9
10 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11 this macro is never tested and has no effect, because the work to support
12 coverage involves only compiling and linking options and special targets in
13 the Makefile. The comment in config.h implied that defining the macro would
14 enable coverage support, which is totally false. There was also support for
15 setting this macro in the CMake files (my fault, I just copied it from
16 configure). SUPPORT_GCOV has now been removed.
17
18 3. Make a small performance improvement in strlen16() and strlen32() in
19 pcretest.
20
21 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23 these cases). There was also one in pcretest.c.
24
25 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
26
27 6. In UTF mode, the code for checking whether a group could match an empty
28 string (which is used for indefinitely repeated groups to allow for
29 breaking an infinite loop) was broken when the group contained a repeated
30 negated single-character class with a character that occupied more than one
31 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
32 UTF-8 mode). The effect was undefined: the group might or might not be
33 deemed as matching an empty string, or the program might have crashed.
34
35 7. The code for checking whether a group could match an empty string was not
36 recognizing that \h, \H, \v, \V, and \R must match a character.
37
38 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
39 an empty string. If it can, pcretest shows this in its information output.
40
41 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
42 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
43 matched by pcre_exec() without using JIT:
44
45 (a) If the rest of the pattern did not match after a maximal run of
46 grapheme clusters, the code for backing up to try with fewer of them
47 did not always back up over a full grapheme when characters that do not
48 have the modifier quality were involved, e.g. Hangul syllables.
49
50 (b) If the match point in a subject started with modifier character, and
51 there was no match, the code could incorrectly back up beyond the match
52 point, and potentially beyond the first character in the subject,
53 leading to a segfault or an incorrect match result.
54
55 10. A conditional group with an assertion condition could lead to PCRE
56 recording an incorrect first data item for a match if no other first data
57 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58 first data item, and therefore matched "ca" after "c" instead of at the
59 start.
60
61 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
62 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
63
64 12. The source of pcregrep now includes z/OS-specific code so that it can be
65 compiled for z/OS as part of the special z/OS distribution.
66
67 13. Added the -T and -TM options to pcretest.
68
69 14. The code in pcre_compile.c for creating the table of named capturing groups
70 has been refactored. Instead of creating the table dynamically during the
71 actual compiling pass, the information is remembered during the pre-compile
72 pass (on the stack unless there are more than 20 named groups, in which
73 case malloc() is used) and the whole table is created before the actual
74 compile happens. This has simplified the code (it is now nearly 150 lines
75 shorter) and prepared the way for better handling of references to groups
76 with duplicate names.
77
78 15. A back reference to a named subpattern when there is more than one of the
79 same name now checks them in the order in which they appear in the pattern.
80 The first one that is set is used for the reference. Previously only the
81 first one was inspected. This change makes PCRE more compatible with Perl.
82
83 16. Unicode character properties were updated from Unicode 6.3.0.
84
85 17. The compile-time code for auto-possessification has been refactored, based
86 on a patch by Zoltan Herczeg. It now happens after instead of during
87 compilation. The code is cleaner, and more cases are handled. The option
88 PCRE_NO_AUTO_POSSESSIFY is added for testing purposes, and the -O and /O
89 options in pcretest are provided to set it.
90
91 18. The character VT has been added to the set of characters that match \s and
92 are generally treated as white space, following this same change in Perl
93 5.18. There is now no difference between "Perl space" and "POSIX space".
94
95 19. The code for checking named groups as conditions, either for being set or
96 for being recursed, has been refactored (this is related to 14 and 15
97 above). Processing unduplicated named groups should now be as fast at
98 numerical groups, and processing duplicated groups should be faster than
99 before.
100
101
102 Version 8.33 28-May-2013
103 ------------------------
104
105 1. Added 'U' to some constants that are compared to unsigned integers, to
106 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
107 variables that are added to signed variables, to ensure the result is
108 signed and can be negated.
109
110 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
111 CMake config files.
112
113 3. Revise the creation of config.h.generic so that all boolean macros are
114 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
115 overriding via -D on the command line possible.
116
117 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
118 to unsigned int is reported to make a quite noticeable speed difference in
119 a specific Windows environment. Testing on Linux did also appear to show
120 some benefit (and it is clearly not harmful). Also fixed the definition of
121 Xop which should be unsigned.
122
123 5. Related to (4), changing the definition of the intermediate variable cc
124 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
125 improvements.
126
127 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
128 unnecessary spaces.
129
130 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
131 and later.
132
133 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
134
135 9. Optimizing fast_forward_start_bits in JIT.
136
137 10. Adding support for callouts in JIT, and fixing some issues revealed
138 during this work. Namely:
139
140 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
141
142 (b) Minimum length was not checked before the matching is started.
143
144 11. The value of capture_last that is passed to callouts was incorrect in some
145 cases when there was a capture on one path that was subsequently abandoned
146 after a backtrack. Also, the capture_last value is now reset after a
147 recursion, since all captures are also reset in this case.
148
149 12. The interpreter no longer returns the "too many substrings" error in the
150 case when an overflowing capture is in a branch that is subsequently
151 abandoned after a backtrack.
152
153 13. In the pathological case when an offset vector of size 2 is used, pcretest
154 now prints out the matched string after a yield of 0 or 1.
155
156 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
157 Only supported by the JIT compiler at the moment.
158
159 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
160
161 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
162 offset of the starting point of the matching process, provided the offsets
163 vector is large enough.
164
165 17. The \A escape now records a lookbehind value of 1, though its execution
166 does not actually inspect the previous character. This is to ensure that,
167 in partial multi-segment matching, at least one character from the old
168 segment is retained when a new segment is processed. Otherwise, if there
169 are no lookbehinds in the pattern, \A might match incorrectly at the start
170 of a new segment.
171
172 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
173
174 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
175 gives some modest performance improvement in 8-bit mode.
176
177 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
178 be expressed in certain programming languages using Universal Character
179 Names.
180
181 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
182 which points out that "non characters" are not "characters that may not
183 appear in Unicode strings" but rather "characters that are reserved for
184 internal use and have only local meaning".
185
186 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
187 there was a conditional group that depended on an assertion, if the
188 assertion was false, the callout that immediately followed the alternation
189 in the condition was skipped when pcre_exec() was used for matching.
190
191 23. Allow an explicit callout to be inserted before an assertion that is the
192 condition for a conditional group, for compatibility with automatic
193 callouts, which always insert a callout at this point.
194
195 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
196 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
197
198 25. (*PRUNE) is now supported by the JIT compiler.
199
200 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
201
202 27. Fix the case where there are two or more SKIPs with arguments that may be
203 ignored.
204
205 28. (*SKIP) is now supported by the JIT compiler.
206
207 29. (*THEN) is now supported by the JIT compiler.
208
209 30. Update RunTest with additional test selector options.
210
211 31. The way PCRE handles backtracking verbs has been changed in two ways.
212
213 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
214 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
215 backtracking. In some cases this makes it more Perl-compatible, but Perl's
216 rather obscure rules do not always do the same thing.
217
218 (2) Previously, backtracking verbs were confined within assertions. This is
219 no longer the case for positive assertions, except for (*ACCEPT). Again,
220 this sometimes improves Perl compatibility, and sometimes does not.
221
222 32. A number of tests that were in test 2 because Perl did things differently
223 have been moved to test 1, because either Perl or PCRE has changed, and
224 these tests are now compatible.
225
226 32. Backtracking control verbs are now handled in the same way in JIT and
227 interpreter.
228
229 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
230 contained a forward subroutine reference caused a compile error.
231
232 34. Auto-detect and optimize limited repetitions in JIT.
233
234 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
235 blocking (*UTF) etc.
236
237 36. In the interpreter, maximizing pattern repetitions for characters and
238 character types now use tail recursion, which reduces stack usage.
239
240 37. The value of the max lookbehind was not correctly preserved if a compiled
241 and saved regex was reloaded on a host of different endianness.
242
243 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
244 of the compiled pattern block, expand the flags field from 16 to 32 bits
245 because it was almost full.
246
247 39. Try madvise first before posix_madvise.
248
249 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
250 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
251 original change locked out all matches of empty strings. This has been
252 changed so that one match of an empty string per line is recognized.
253 Subsequent searches on the same line (for colouring or for --only-matching,
254 for example) do not recognize empty strings.
255
256 41. Applied a user patch to fix a number of spelling mistakes in comments.
257
258 42. Data lines longer than 65536 caused pcretest to crash.
259
260 43. Clarified the data type for length and startoffset arguments for pcre_exec
261 and pcre_dfa_exec in the function-specific man pages, where they were
262 explicitly stated to be in bytes, never having been updated. I also added
263 some clarification to the pcreapi man page.
264
265 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
266 a segmentation fault.
267
268
269 Version 8.32 30-November-2012
270 -----------------------------
271
272 1. Improved JIT compiler optimizations for first character search and single
273 character iterators.
274
275 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
276 Patch by Daniel Richard G.
277
278 3. Single character iterator optimizations in the JIT compiler.
279
280 4. Improved JIT compiler optimizations for character ranges.
281
282 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
283 Reported by Giuseppe D'Angelo.
284
285 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
286 of a line, was being set incorrectly in cases where .* appeared inside
287 atomic brackets at the start of a pattern, or where there was a subsequent
288 *PRUNE or *SKIP.
289
290 7. Improved instruction cache flush for POWER/PowerPC.
291 Patch by Daniel Richard G.
292
293 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
294 grep:
295
296 (a) There is now no limit to the number of patterns to be matched.
297
298 (b) An error is given if a pattern is too long.
299
300 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
301 are now supported.
302
303 (d) --exclude-from and --include-from (multiple use) have been added.
304
305 (e) Exclusions and inclusions now apply to all files and directories, not
306 just to those obtained from scanning a directory recursively.
307
308 (f) Multiple uses of -f and --file-list are now supported.
309
310 (g) In a Windows environment, the default for -d has been changed from
311 "read" (the GNU grep default) to "skip", because otherwise the presence
312 of a directory in the file list provokes an error.
313
314 (h) The documentation has been revised and clarified in places.
315
316 9. Improve the matching speed of capturing brackets.
317
318 10. Changed the meaning of \X so that it now matches a Unicode extended
319 grapheme cluster.
320
321 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
322 out POSIX threads when JIT support is configured.
323
324 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
325
326 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
327 structure could go wrong in environments where size_t is not the same size
328 as int.
329
330 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
331 set.
332
333 15. The EBCDIC support had decayed; later updates to the code had included
334 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
335 general tidy up of EBCDIC-related issues, and the documentation was also
336 not quite right. There is now a test that can be run on ASCII systems to
337 check some of the EBCDIC-related things (but is it not a full test).
338
339 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
340 in a small tidy to the code.
341
342 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
343
344 18. If the --only-matching (-o) option in pcregrep is specified multiple
345 times, each one causes appropriate output. For example, -o1 -o2 outputs the
346 substrings matched by the 1st and 2nd capturing parentheses. A separating
347 string can be specified by --om-separator (default empty).
348
349 19. Improving the first n character searches.
350
351 20. Turn case lists for horizontal and vertical white space into macros so that
352 they are defined only once.
353
354 21. This set of changes together give more compatible Unicode case-folding
355 behaviour for characters that have more than one other case when UCP
356 support is available.
357
358 (a) The Unicode property table now has offsets into a new table of sets of
359 three or more characters that are case-equivalent. The MultiStage2.py
360 script that generates these tables (the pcre_ucd.c file) now scans
361 CaseFolding.txt instead of UnicodeData.txt for character case
362 information.
363
364 (b) The code for adding characters or ranges of characters to a character
365 class has been abstracted into a generalized function that also handles
366 case-independence. In UTF-mode with UCP support, this uses the new data
367 to handle characters with more than one other case.
368
369 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
370 whose other case is greater than 256 are now correctly matched
371 caselessly. Previously, the high codepoint matched the low one, but not
372 vice versa.
373
374 (d) The processing of \h, \H, \v, and \ in character classes now makes use
375 of the new class addition function, using character lists defined as
376 macros alongside the case definitions of 20 above.
377
378 (e) Caseless back references now work with characters that have more than
379 one other case.
380
381 (f) General caseless matching of characters with more than one other case
382 is supported.
383
384 22. Unicode character properties were updated from Unicode 6.2.0
385
386 23. Improved CMake support under Windows. Patch by Daniel Richard G.
387
388 24. Add support for 32-bit character strings, and UTF-32
389
390 25. Major JIT compiler update (code refactoring and bugfixing).
391 Experimental Sparc 32 support is added.
392
393 26. Applied a modified version of Daniel Richard G's patch to create
394 pcre.h.generic and config.h.generic by "make" instead of in the
395 PrepareRelease script.
396
397 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
398 pcre_compile.c when checking for a zero character.
399
400 28. Introducing a native interface for JIT. Through this interface, the compiled
401 machine code can be directly executed. The purpose of this interface is to
402 provide fast pattern matching, so several sanity checks are not performed.
403 However, feature tests are still performed. The new interface provides
404 1.4x speedup compared to the old one.
405
406 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
407 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
408 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
409
410 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
411 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
412 the "old" RFC 2279). Instead, it ended up passing a negative length to
413 pcre_exec().
414
415 31. Add support for GCC's visibility feature to hide internal functions.
416
417 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
418 "unknown -C option" after outputting 0 or 1.
419
420 33. There is now support for generating a code coverage report for the test
421 suite in environments where gcc is the compiler and lcov is installed. This
422 is mainly for the benefit of the developers.
423
424 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
425 unaddressable using valgrind annotations, allowing valgrind to detect
426 invalid memory accesses. This is mainly for the benefit of the developers.
427
428 25. (*UTF) can now be used to start a pattern in any of the three libraries.
429
430 26. Give configure error if --enable-cpp but no C++ compiler found.
431
432
433 Version 8.31 06-July-2012
434 -------------------------
435
436 1. Fixing a wrong JIT test case and some compiler warnings.
437
438 2. Removed a bashism from the RunTest script.
439
440 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
441 to unsigned type, result still unsigned" that was given by an MS compiler
442 on encountering the code "-sizeof(xxx)".
443
444 4. Partial matching support is added to the JIT compiler.
445
446 5. Fixed several bugs concerned with partial matching of items that consist
447 of more than one character:
448
449 (a) /^(..)\1/ did not partially match "aba" because checking references was
450 done on an "all or nothing" basis. This also applied to repeated
451 references.
452
453 (b) \R did not give a hard partial match if \r was found at the end of the
454 subject.
455
456 (c) \X did not give a hard partial match after matching one or more
457 characters at the end of the subject.
458
459 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
460 a partial match for the string "\r".
461
462 (e) When newline was set to CRLF, the metacharacter "." did not recognize
463 a partial match for a CR character at the end of the subject string.
464
465 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
466 running pcretest, the text "(JIT)" added to the output whenever JIT is
467 actually used to run the match.
468
469 7. Individual JIT compile options can be set in pcretest by following -s+[+]
470 or /S+[+] with a digit between 1 and 7.
471
472 8. OP_NOT now supports any UTF character not just single-byte ones.
473
474 9. (*MARK) control verb is now supported by the JIT compiler.
475
476 10. The command "./RunTest list" lists the available tests without actually
477 running any of them. (Because I keep forgetting what they all are.)
478
479 11. Add PCRE_INFO_MAXLOOKBEHIND.
480
481 12. Applied a (slightly modified) user-supplied patch that improves performance
482 when the heap is used for recursion (compiled with --disable-stack-for-
483 recursion). Instead of malloc and free for each heap frame each time a
484 logical recursion happens, frames are retained on a chain and re-used where
485 possible. This sometimes gives as much as 30% improvement.
486
487 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
488 call.
489
490 14. As documented, (*COMMIT) is now confined to within a positive assertion.
491
492 15. It is now possible to link pcretest with libedit as an alternative to
493 libreadline.
494
495 16. (*COMMIT) control verb is now supported by the JIT compiler.
496
497 17. The Unicode data tables have been updated to Unicode 6.1.0.
498
499 18. Added --file-list option to pcregrep.
500
501 19. Added binary file support to pcregrep, including the -a, --binary-files,
502 -I, and --text options.
503
504 20. The madvise function is renamed for posix_madvise for QNX compatibility
505 reasons. Fixed by Giuseppe D'Angelo.
506
507 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
508
509 22. Changed the option for creating symbolic links for 16-bit man pages from
510 -s to -sf so that re-installing does not cause issues.
511
512 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
513
514 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
515 matches in certain environments (the workspace was not being correctly
516 retained). Also added to pcre_dfa_exec() a simple plausibility check on
517 some of the workspace data at the beginning of a restart.
518
519 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
520 was not doing so when it should - probably a typo introduced by SVN 528
521 (change 8.10/14).
522
523 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
524 \w+ when the character tables indicated that \x{c4} was a word character.
525 There were several related cases, all because the tests for doing a table
526 lookup were testing for characters less than 127 instead of 255.
527
528 27. If a pattern contains capturing parentheses that are not used in a match,
529 their slots in the ovector are set to -1. For those that are higher than
530 any matched groups, this happens at the end of processing. In the case when
531 there were back references that the ovector was too small to contain
532 (causing temporary malloc'd memory to be used during matching), and the
533 highest capturing number was not used, memory off the end of the ovector
534 was incorrectly being set to -1. (It was using the size of the temporary
535 memory instead of the true size.)
536
537 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
538 ovector size, it uses memory at the end of the block that it has got.
539
540 29. Check for an overlong MARK name and give an error at compile time. The
541 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
542
543 30. JIT compiler update.
544
545 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
546 Rill for the patch.
547
548 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
549
550 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
551
552 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
553 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
554 was enabled.
555
556 35. Improve JIT code generation for greedy plus quantifier.
557
558 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
559 1 to "aa" instead of to an empty string. The bug affected repeated groups
560 that could potentially match an empty string.
561
562 37. Optimizing single character iterators in JIT.
563
564 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
565 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
566 codepoints that are too big for the mode are faulted, and in a UTF mode,
567 disallowed codepoints are also faulted.
568
569 39. If PCRE was compiled with UTF support, in three places in the DFA
570 matcher there was code that should only have been obeyed in UTF mode, but
571 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
572 processing when bytes with values greater than 127 were present. In 16-bit
573 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
574 both cases the values are those that cannot be the first data item in a UTF
575 character. The three items that might have provoked this were recursions,
576 possessively repeated groups, and atomic groups.
577
578 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
579 and pcregrep, because some OS require shared objects to be explicitly
580 passed to ld, causing the link step to fail if they are not.
581
582 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
583 mode, patterns that started with \h* or \R* might be incorrectly matched.
584
585
586 Version 8.30 04-February-2012
587 -----------------------------
588
589 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
590 name is defined in ctype.h.
591
592 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
593 only in quite long subpatterns.
594
595 3. Removed the function pcre_info(), which has been obsolete and deprecated
596 since it was replaced by pcre_fullinfo() in February 2000.
597
598 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
599 match a (*MARK), and the match failed at the start of the subject, a
600 reference to memory before the start of the subject could occur. This bug
601 was introduced by fix 17 of release 8.21.
602
603 5. A reference to an unset group with zero minimum repetition was giving
604 totally wrong answers (in non-JavaScript-compatibility mode). For example,
605 /(another)?(\1?)test/ matched against "hello world test". This bug was
606 introduced in release 8.13.
607
608 6. Add support for 16-bit character strings (a large amount of work involving
609 many changes and refactorings).
610
611 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
612 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
613 from a file.
614
615 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
616 rounding is not applied in this particular case).
617
618 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
619 if they appear, or are escaped, in patterns.
620
621 10. Get rid of a number of -Wunused-but-set-variable warnings.
622
623 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
624 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
625 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
626 also returns the mark "x". This bug applied to capturing parentheses,
627 non-capturing parentheses, and atomic parentheses. It also applied to some
628 assertions.
629
630 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
631 information out of configure.ac instead of relying on pcre.h.generic, which
632 is not stored in the repository.
633
634 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
635 -lreadline.
636
637 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
638
639 15. Applied Graycode's patch to put the top-level frame on the stack rather
640 than the heap when not using the stack for recursion. This gives a
641 performance improvement in many cases when recursion is not deep.
642
643 16. Experimental code added to "pcretest -C" to output the stack frame size.
644
645
646 Version 8.21 12-Dec-2011
647 ------------------------
648
649 1. Updating the JIT compiler.
650
651 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
652 are added as well.
653
654 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
655 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
656 calling _pcre_jit_exec. Some extra comments are added.
657
658 4. (*MARK) settings inside atomic groups that do not contain any capturing
659 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
660 was introduced by change 18 for 8.20.
661
662 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
663 ECMA-262 standard.
664
665 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
666 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
667 This bug was probably introduced by change 9 of 8.13.
668
669 7. While fixing 6 above, I noticed that a number of other items were being
670 incorrectly rejected as "not fixed length". This arose partly because newer
671 opcodes had not been added to the fixed-length checking code. I have (a)
672 corrected the bug and added tests for these items, and (b) arranged for an
673 error to occur if an unknown opcode is encountered while checking for fixed
674 length instead of just assuming "not fixed length". The items that were
675 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
676 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
677 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
678
679 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
680 being incorrectly compiled and would have given unpredicatble results.
681
682 9. A possessively repeated subpattern with minimum repeat count greater than
683 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
684 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
685 the first (A) could occur when it should not.
686
687 10. Add a cast and remove a redundant test from the code.
688
689 11. JIT should use pcre_malloc/pcre_free for allocation.
690
691 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
692 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
693 is anything other than /usr, -L is still shown).
694
695 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
696
697 14. Perl does not support \N without a following name in a [] class; PCRE now
698 also gives an error.
699
700 15. If a forward reference was repeated with an upper limit of around 2000,
701 it caused the error "internal error: overran compiling workspace". The
702 maximum number of forward references (including repeats) was limited by the
703 internal workspace, and dependent on the LINK_SIZE. The code has been
704 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
705 the default depends on LINK_SIZE. There is a new upper limit (for safety)
706 of around 200,000 forward references. While doing this, I also speeded up
707 the filling in of repeated forward references.
708
709 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
710 incorrectly expecting the subject to contain another "a" after the start.
711
712 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
713 in the match, the SKIP should be ignored. This was not happening; instead
714 the SKIP was being treated as NOMATCH. For patterns such as
715 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
716 tested.
717
718 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
719 now much more compatible with Perl, in particular in cases where the result
720 is a non-match for a non-anchored pattern. For example, if
721 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
722 "m", where previously it did not return a name. A side effect of this
723 change is that for partial matches, the last encountered mark name is
724 returned, as for non matches. A number of tests that were previously not
725 Perl-compatible have been moved into the Perl-compatible test files. The
726 refactoring has had the pleasing side effect of removing one argument from
727 the match() function, thus reducing its stack requirements.
728
729 19. If the /S+ option was used in pcretest to study a pattern using JIT,
730 subsequent uses of /S (without +) incorrectly behaved like /S+.
731
732 21. Retrieve executable code size support for the JIT compiler and fixing
733 some warnings.
734
735 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
736 not work when the shorter character appeared right at the end of the
737 subject string.
738
739 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
740 systems.
741
742 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
743 output it when the /M option is used in pcretest.
744
745 25. The CheckMan script was not being included in the distribution. Also, added
746 an explicit "perl" to run Perl scripts from the PrepareRelease script
747 because this is reportedly needed in Windows.
748
749 26. If study data was being save in a file and studying had not found a set of
750 "starts with" bytes for the pattern, the data written to the file (though
751 never used) was taken from uninitialized memory and so caused valgrind to
752 complain.
753
754 27. Updated RunTest.bat as provided by Sheri Pierce.
755
756 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
757
758 29. Computation of memory usage for the table of capturing group names was
759 giving an unnecessarily large value.
760
761
762 Version 8.20 21-Oct-2011
763 ------------------------
764
765 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
766 a POSIX class. After further experiments with Perl, which convinced me that
767 Perl has bugs and confusions, a closing square bracket is no longer allowed
768 in a POSIX name. This bug also affected patterns with classes that started
769 with full stops.
770
771 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
772 captured substring, but while checking the failing first alternative,
773 substring 1 is temporarily captured. If the output vector supplied to
774 pcre_exec() was not big enough for this capture, the yield of the function
775 was still zero ("insufficient space for captured substrings"). This cannot
776 be totally fixed without adding another stack variable, which seems a lot
777 of expense for a edge case. However, I have improved the situation in cases
778 such as /(a)(b)x|abc/ matched against "abc", where the return code
779 indicates that fewer than the maximum number of slots in the ovector have
780 been set.
781
782 3. Related to (2) above: when there are more back references in a pattern than
783 slots in the output vector, pcre_exec() uses temporary memory during
784 matching, and copies in the captures as far as possible afterwards. It was
785 using the entire output vector, but this conflicts with the specification
786 that only 2/3 is used for passing back captured substrings. Now it uses
787 only the first 2/3, for compatibility. This is, of course, another edge
788 case.
789
790 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
791 main code base, and can be used by building with --enable-jit. When this is
792 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
793 runtime --no-jit option is given.
794
795 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
796 ovector, the return from the function was zero, implying that there were
797 other matches that did not fit. The correct "exactly full" value is now
798 returned.
799
800 6. If a subpattern that was called recursively or as a subroutine contained
801 (*PRUNE) or any other control that caused it to give a non-standard return,
802 invalid errors such as "Error -26 (nested recursion at the same subject
803 position)" or even infinite loops could occur.
804
805 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
806 computing the minimum length on reaching *ACCEPT, and so ended up with the
807 wrong value of 1 rather than 0. Further investigation indicates that
808 computing a minimum subject length in the presence of *ACCEPT is difficult
809 (think back references, subroutine calls), and so I have changed the code
810 so that no minimum is registered for a pattern that contains *ACCEPT.
811
812 8. If (*THEN) was present in the first (true) branch of a conditional group,
813 it was not handled as intended. [But see 16 below.]
814
815 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
816 Sheri Pierce.
817
818 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
819 the first byte in a match must be "a".
820
821 11. Change 17 for 8.13 increased the recursion depth for patterns like
822 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
823 pattern contains any instances of (*THEN). If it does not, the old
824 optimizations are restored. It would be nice to do this on a per-group
825 basis, but at the moment that is not feasible.
826
827 12. In some environments, the output of pcretest -C is CRLF terminated. This
828 broke RunTest's code that checks for the link size. A single white space
829 character after the value is now allowed for.
830
831 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
832 For "fr", it uses the Windows-specific input and output files.
833
834 14. If (*THEN) appeared in a group that was called recursively or as a
835 subroutine, it did not work as intended. [But see next item.]
836
837 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
838 pattern fragments (but not containing any | characters). If A and B are
839 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
840 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
841 D. In other words, Perl considers parentheses that do not contain any |
842 characters to be part of a surrounding alternative, whereas PCRE was
843 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
844 differently. PCRE now behaves in the same way as Perl, except in the case
845 of subroutine/recursion calls such as (?1) which have in any case always
846 been different (but PCRE had them first :-).
847
848 16. Related to 15 above: Perl does not treat the | in a conditional group as
849 creating alternatives. Such a group is treated in the same way as an
850 ordinary group without any | characters when processing (*THEN). PCRE has
851 been changed to match Perl's behaviour.
852
853 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
854 RunGrepTest script failed.
855
856 18. Change 22 for version 13 caused atomic groups to use more stack. This is
857 inevitable for groups that contain captures, but it can lead to a lot of
858 stack use in large patterns. The old behaviour has been restored for atomic
859 groups that do not contain any capturing parentheses.
860
861 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
862 suppress the check for a minimum subject length at run time. (If it was
863 given to pcre_exec() or pcre_dfa_exec() it did work.)
864
865 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
866 fail to work when decoding hex characters in data strings in EBCDIC
867 environments.
868
869 21. It appears that in at least one Mac OS environment, the isxdigit() function
870 is implemented as a macro that evaluates to its argument more than once,
871 contravening the C 90 Standard (I haven't checked a later standard). There
872 was an instance in pcretest which caused it to go wrong when processing
873 \x{...} escapes in subject strings. The has been rewritten to avoid using
874 things like p++ in the argument of isxdigit().
875
876
877 Version 8.13 16-Aug-2011
878 ------------------------
879
880 1. The Unicode data tables have been updated to Unicode 6.0.0.
881
882 2. Two minor typos in pcre_internal.h have been fixed.
883
884 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
885 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
886 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
887
888 4. There were a number of related bugs in the code for matching backrefences
889 caselessly in UTF-8 mode when codes for the characters concerned were
890 different numbers of bytes. For example, U+023A and U+2C65 are an upper
891 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
892 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
893 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
894 2-byte code at the end of the subject (it thought there wasn't enough data
895 left).
896
897 5. Comprehensive information about what went wrong is now returned by
898 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
899 as the output vector has at least 2 elements. The offset of the start of
900 the failing character and a reason code are placed in the vector.
901
902 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
903 now returned is for the first byte of the failing character, instead of the
904 last byte inspected. This is an incompatible change, but I hope it is small
905 enough not to be a problem. It makes the returned offset consistent with
906 pcre_exec() and pcre_dfa_exec().
907
908 7. pcretest now gives a text phrase as well as the error number when
909 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
910 failure, the offset and reason code are output.
911
912 8. When \R was used with a maximizing quantifier it failed to skip backwards
913 over a \r\n pair if the subsequent match failed. Instead, it just skipped
914 back over a single character (\n). This seems wrong (because it treated the
915 two characters as a single entity when going forwards), conflicts with the
916 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
917 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
918 has been changed.
919
920 9. Some internal refactoring has changed the processing so that the handling
921 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
922 time (the PCRE_DOTALL option was changed this way some time ago: version
923 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
924 which was always a bit of a fudge. It also means that there is one less
925 argument for the match() function, which reduces its stack requirements
926 slightly. This change also fixes an incompatibility with Perl: the pattern
927 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
928
929 10. More internal refactoring has drastically reduced the number of recursive
930 calls to match() for possessively repeated groups such as (abc)++ when
931 using pcre_exec().
932
933 11. While implementing 10, a number of bugs in the handling of groups were
934 discovered and fixed:
935
936 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
937 (a|)*(?1) gave a compile-time internal error.
938 ((a|)+)+ did not notice that the outer group could match an empty string.
939 (^a|^)+ was not marked as anchored.
940 (.*a|.*)+ was not marked as matching at start or after a newline.
941
942 12. Yet more internal refactoring has removed another argument from the match()
943 function. Special calls to this function are now indicated by setting a
944 value in a variable in the "match data" data block.
945
946 13. Be more explicit in pcre_study() instead of relying on "default" for
947 opcodes that mean there is no starting character; this means that when new
948 ones are added and accidentally left out of pcre_study(), testing should
949 pick them up.
950
951 14. The -s option of pcretest has been documented for ages as being an old
952 synonym of -m (show memory usage). I have changed it to mean "force study
953 for every regex", that is, assume /S for every regex. This is similar to -i
954 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
955 using it. It makes it easier to run collections of tests with and without
956 study enabled, and thereby test pcre_study() more easily. All the standard
957 tests are now run with and without -s (but some patterns can be marked as
958 "never study" - see 20 below).
959
960 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
961 restoration of the capturing data to the outer values was not happening
962 correctly.
963
964 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
965 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
966 pattern had matched an empty string, and so incorrectly returned a no
967 match.
968
969 17. There was optimizing code for the last branch of non-capturing parentheses,
970 and also for the obeyed branch of a conditional subexpression, which used
971 tail recursion to cut down on stack usage. Unfortunately, now that there is
972 the possibility of (*THEN) occurring in these branches, tail recursion is
973 no longer possible because the return has to be checked for (*THEN). These
974 two optimizations have therefore been removed. [But see 8.20/11 above.]
975
976 18. If a pattern containing \R was studied, it was assumed that \R always
977 matched two bytes, thus causing the minimum subject length to be
978 incorrectly computed because \R can also match just one byte.
979
980 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
981 was incorrectly computed.
982
983 20. If /S is present twice on a test pattern in pcretest input, it now
984 *disables* studying, thereby overriding the use of -s on the command line
985 (see 14 above). This is necessary for one or two tests to keep the output
986 identical in both cases.
987
988 21. When (*ACCEPT) was used in an assertion that matched an empty string and
989 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
990
991 22. When an atomic group that contained a capturing parenthesis was
992 successfully matched, but the branch in which it appeared failed, the
993 capturing was not being forgotten if a higher numbered group was later
994 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
995 group 1 to "a", when in fact it should be unset. This applied to multi-
996 branched capturing and non-capturing groups, repeated or not, and also to
997 positive assertions (capturing in negative assertions does not happen
998 in PCRE) and also to nested atomic groups.
999
1000 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1001 subject after a captured substring, to make it easier to tell which of a
1002 number of identical substrings has been captured.
1003
1004 24. The way atomic groups are processed by pcre_exec() has been changed so that
1005 if they are repeated, backtracking one repetition now resets captured
1006 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1007 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1008 "aaa". Previously, it would have been "a". As part of this code
1009 refactoring, the way recursive calls are handled has also been changed.
1010
1011 25. If an assertion condition captured any substrings, they were not passed
1012 back unless some other capturing happened later. For example, if
1013 (?(?=(a))a) was matched against "a", no capturing was returned.
1014
1015 26. When studying a pattern that contained subroutine calls or assertions,
1016 the code for finding the minimum length of a possible match was handling
1017 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1018 group 1 called group 2 while simultaneously a separate group 2 called group
1019 1). A stack overflow occurred in this case. I have fixed this by limiting
1020 the recursion depth to 10.
1021
1022 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1023 Fortmann. This supports explicit test numbers on the command line, and has
1024 argument validation and error reporting.
1025
1026 28. An instance of \X with an unlimited repeat could fail if at any point the
1027 first character it looked at was a mark character.
1028
1029 29. Some minor code refactoring concerning Unicode properties and scripts
1030 should reduce the stack requirement of match() slightly.
1031
1032 30. Added the '=' option to pcretest to check the setting of unused capturing
1033 slots at the end of the pattern, which are documented as being -1, but are
1034 not included in the return count.
1035
1036 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1037 compiled something random. Now it gives a compile-time error (as does
1038 Perl).
1039
1040 32. A *MARK encountered during the processing of a positive assertion is now
1041 recorded and passed back (compatible with Perl).
1042
1043 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1044 had alternative anchored branches, the search for a second match in a line
1045 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1046 matched the line "0102" twice. The same bug affected patterns that started
1047 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1048 twice.
1049
1050 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1051 does, and because of capturing effects, quantifying parenthesized
1052 assertions may at times be useful. Quantifiers are now allowed for
1053 parenthesized assertions.
1054
1055 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1056
1057 36. \g was being checked for fancy things in a character class, when it should
1058 just be a literal "g".
1059
1060 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1061 appearance of a nested POSIX class supersedes an apparent external class.
1062 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1063 unescaped square brackets may also appear as part of class names. For
1064 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1065 more like Perl. (But see 8.20/1 above.)
1066
1067 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1068 was because it thought it was \N{name}, which is not supported).
1069
1070 39. Add minix to OS list not supporting the -S option in pcretest.
1071
1072 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1073 cannot analyze patterns in sufficient detail to catch mutual recursions
1074 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1075 subgroup is called recursively as a subpattern for a second time at the
1076 same position in the subject string. In previous releases this might have
1077 been caught by the recursion limit, or it might have run out of stack.
1078
1079 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1080 happen only once. PCRE was, however incorrectly giving a compile time error
1081 "recursive call could loop indefinitely" because it cannot analyze the
1082 pattern in sufficient detail. The compile time test no longer happens when
1083 PCRE is compiling a conditional subpattern, but actual runaway loops are
1084 now caught at runtime (see 40 above).
1085
1086 42. It seems that Perl allows any characters other than a closing parenthesis
1087 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1088 has been changed to be the same.
1089
1090 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1091 as not to get warnings when autogen.sh is called. Also changed
1092 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1093
1094 44. To help people who use pcregrep to scan files containing exceedingly long
1095 lines, the following changes have been made:
1096
1097 (a) The default value of the buffer size parameter has been increased from
1098 8K to 20K. (The actual buffer used is three times this size.)
1099
1100 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1101 PCRE is built.
1102
1103 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1104 to be set at run time.
1105
1106 (d) Numerical values in pcregrep options can be followed by K or M, for
1107 example --buffer-size=50K.
1108
1109 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1110 given and the return code is set to 2.
1111
1112 45. Add a pointer to the latest mark to the callout data block.
1113
1114 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1115 partial match of an empty string instead of no match. This was specific to
1116 the use of ".".
1117
1118 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1119 complete match instead of a partial match. This bug was dependent on both
1120 the PCRE_UTF8 and PCRE_DOTALL options being set.
1121
1122 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1123 starting byte set, because \b was not being ignored.
1124
1125
1126 Version 8.12 15-Jan-2011
1127 ------------------------
1128
1129 1. Fixed some typos in the markup of the man pages, and wrote a script that
1130 checks for such things as part of the documentation building process.
1131
1132 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1133 --match-limit and --recursion-limit options (added for 8.11). In
1134 particular, this made one of the standard tests fail. (The integer value
1135 went into the wrong half of a long int.)
1136
1137 3. If the --colour option was given to pcregrep with -v (invert match), it
1138 did strange things, either producing crazy output, or crashing. It should,
1139 of course, ignore a request for colour when reporting lines that do not
1140 match.
1141
1142 4. Another pcregrep bug caused similar problems if --colour was specified with
1143 -M (multiline) and the pattern match finished with a line ending.
1144
1145 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1146 matched in multiline mode, the following line was shown as part of the
1147 match. This seems wrong, so I have changed it.
1148
1149 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1150 the check for further matches in the same line (so they could be coloured)
1151 to overrun the end of the current line. If another match was found, it was
1152 incorrectly shown (and then shown again when found in the next line).
1153
1154 7. If pcregrep was compiled under Windows, there was a reference to the
1155 function pcregrep_exit() before it was defined. I am assuming this was
1156 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1157 reported by a user. I've moved the definition above the reference.
1158
1159
1160 Version 8.11 10-Dec-2010
1161 ------------------------
1162
1163 1. (*THEN) was not working properly if there were untried alternatives prior
1164 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1165 backtracked to try for "b" instead of moving to the next alternative branch
1166 at the same level (in this case, to look for "c"). The Perl documentation
1167 is clear that when (*THEN) is backtracked onto, it goes to the "next
1168 alternative in the innermost enclosing group".
1169
1170 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1171 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1172 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1173 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1174 (*THEN).
1175
1176 3. If \s appeared in a character class, it removed the VT character from
1177 the class, even if it had been included by some previous item, for example
1178 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1179 of \s, but is part of the POSIX "space" class.)
1180
1181 4. A partial match never returns an empty string (because you can always
1182 match an empty string at the end of the subject); however the checking for
1183 an empty string was starting at the "start of match" point. This has been
1184 changed to the "earliest inspected character" point, because the returned
1185 data for a partial match starts at this character. This means that, for
1186 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1187 (previously it gave "no match").
1188
1189 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1190 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1191 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1192 has an implication that the given string is incomplete (because a partial
1193 match is preferred over a full match). For this reason, these items now
1194 give a partial match in this situation. [Aside: previously, the one case
1195 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1196 match rather than a full match, which was wrong by the old rules, but is
1197 now correct.]
1198
1199 6. There was a bug in the handling of #-introduced comments, recognized when
1200 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1201 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1202 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1203 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1204 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1205 places in pcre_compile().
1206
1207 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1208 comments when looking ahead for named forward references to subpatterns,
1209 the only newline sequence it recognized was NL. It now handles newlines
1210 according to the set newline convention.
1211
1212 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1213 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1214 cater for a lack of strerror(). These oversights have been fixed.
1215
1216 9. Added --match-limit and --recursion-limit to pcregrep.
1217
1218 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1219
1220 11. When the -o option was used, pcregrep was setting a return code of 1, even
1221 when matches were found, and --line-buffered was not being honoured.
1222
1223 12. Added an optional parentheses number to the -o and --only-matching options
1224 of pcregrep.
1225
1226 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1227 can match an empty string. The code to do it in pcretest and pcredemo
1228 needed fixing:
1229
1230 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1231 only one byte after an empty string match just before CRLF (this case
1232 just got forgotten; "any" and "anycrlf" were OK).
1233
1234 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1235 mode when an empty string match preceded an ASCII character followed by
1236 a non-ASCII character. (The code for advancing by one character rather
1237 than one byte was nonsense.)
1238
1239 (c) The pcredemo.c sample program did not have any code at all to handle
1240 the cases when CRLF is a valid newline sequence.
1241
1242 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1243 as a starting offset was within the subject string. There is now a new
1244 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1245 negative or greater than the length of the string. In order to test this,
1246 pcretest is extended to allow the setting of negative starting offsets.
1247
1248 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1249 starting offset points to the beginning of a UTF-8 character was
1250 unnecessarily clumsy. I tidied it up.
1251
1252 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1253 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1254
1255 17. Nobody had reported that the --include_dir option, which was added in
1256 release 7.7 should have been called --include-dir (hyphen, not underscore)
1257 for compatibility with GNU grep. I have changed it to --include-dir, but
1258 left --include_dir as an undocumented synonym, and the same for
1259 --exclude-dir, though that is not available in GNU grep, at least as of
1260 release 2.5.4.
1261
1262 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1263 characters from a string of bytes) have been redefined so as not to use
1264 loops, in order to improve performance in some environments. At the same
1265 time, I abstracted some of the common code into auxiliary macros to save
1266 repetition (this should not affect the compiled code).
1267
1268 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1269 compile-time error is now given if \c is not followed by an ASCII
1270 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1271 different, and any byte value is allowed.)
1272
1273 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1274 START_OPTIMIZE option, which is now allowed at compile time - but just
1275 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1276 to pcregrep and other applications that have no direct access to PCRE
1277 options. The new /Y option in pcretest sets this option when calling
1278 pcre_compile().
1279
1280 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1281 back references. Groups containing recursive back references were forced to
1282 be atomic by that change, but in the case of named groups, the amount of
1283 memory required was incorrectly computed, leading to "Failed: internal
1284 error: code overflow". This has been fixed.
1285
1286 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1287 pcretest.c, to avoid build problems in some Borland environments.
1288
1289
1290 Version 8.10 25-Jun-2010
1291 ------------------------
1292
1293 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1294 THEN.
1295
1296 2. (*ACCEPT) was not working when inside an atomic group.
1297
1298 3. Inside a character class, \B is treated as a literal by default, but
1299 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1300 causes the error). The code is unchanged, but I tidied the documentation.
1301
1302 4. Inside a character class, PCRE always treated \R and \X as literals,
1303 whereas Perl faults them if its -w option is set. I have changed PCRE so
1304 that it faults them when PCRE_EXTRA is set.
1305
1306 5. Added support for \N, which always matches any character other than
1307 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1308
1309 6. When compiling pcregrep with newer versions of gcc which may have
1310 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1311 declared with attribute warn_unused_result" were given. Just casting the
1312 result to (void) does not stop the warnings; a more elaborate fudge is
1313 needed. I've used a macro to implement this.
1314
1315 7. Minor change to pcretest.c to avoid a compiler warning.
1316
1317 8. Added four artifical Unicode properties to help with an option to make
1318 \s etc use properties (see next item). The new properties are: Xan
1319 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1320
1321 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1322 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1323 this option. Modified pcretest to add /W to test this facility. Added
1324 REG_UCP to make it available via the POSIX interface.
1325
1326 10. Added --line-buffered to pcregrep.
1327
1328 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1329 studied, and the match started with a letter with a code point greater than
1330 127 whose first byte was different to the first byte of the other case of
1331 the letter, the other case of this starting letter was not recognized
1332 (#976).
1333
1334 12. If a pattern that was studied started with a repeated Unicode property
1335 test, for example, \p{Nd}+, there was the theoretical possibility of
1336 setting up an incorrect bitmap of starting bytes, but fortunately it could
1337 not have actually happened in practice until change 8 above was made (it
1338 added property types that matched character-matching opcodes).
1339
1340 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1341 possible starting bytes for non-anchored patterns.
1342
1343 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1344 \R, and also a number of cases that involve Unicode properties, both
1345 explicit and implicit when PCRE_UCP is set.
1346
1347 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1348 input, it could crash or give wrong results if characters with values
1349 greater than 0xc0 were present in the subject string. (Detail: it assumed
1350 UTF-8 input when processing these items.)
1351
1352 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1353 size_t is 64-bit (#991).
1354
1355 17. Added a check for running out of memory when PCRE is compiled with
1356 --disable-stack-for-recursion (#990).
1357
1358 18. If the last data line in a file for pcretest does not have a newline on
1359 the end, a newline was missing in the output.
1360
1361 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1362 less than 128) in its various bitmaps. However, there is a facility for
1363 generating tables according to the current locale when PCRE is compiled. It
1364 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1365 characters, are recognized by isspace() and therefore were getting set in
1366 these tables, and indeed these tables seem to approximate to ISO 8859. This
1367 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1368 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1369 which of course cannot start UTF-8 characters. I have changed the code so
1370 that only real ASCII characters (less than 128) and the correct starting
1371 bytes for UTF-8 encodings are set for characters greater than 127 when in
1372 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1373 altogether.)
1374
1375 20. Added the /T option to pcretest so as to be able to run tests with non-
1376 standard character tables, thus making it possible to include the tests
1377 used for 19 above in the standard set of tests.
1378
1379 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1380 reference to a subpattern the other side of a comment that contains an
1381 opening parenthesis caused either an internal compiling error, or a
1382 reference to the wrong subpattern.
1383
1384
1385 Version 8.02 19-Mar-2010
1386 ------------------------
1387
1388 1. The Unicode data tables have been updated to Unicode 5.2.0.
1389
1390 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1391 configured.
1392
1393 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1394 original author of that file, following a query about its status.
1395
1396 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1397 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1398
1399 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1400 quantifier applied to a forward-referencing subroutine call, could compile
1401 incorrect code or give the error "internal error: previously-checked
1402 referenced subpattern not found".
1403
1404 6. Both MS Visual Studio and Symbian OS have problems with initializing
1405 variables to point to external functions. For these systems, therefore,
1406 pcre_malloc etc. are now initialized to local functions that call the
1407 relevant global functions.
1408
1409 7. There were two entries missing in the vectors called coptable and poptable
1410 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1411 I've fixed the data, and added a kludgy way of testing at compile time that
1412 the lengths are correct (equal to the number of opcodes).
1413
1414 8. Following on from 7, I added a similar kludge to check the length of the
1415 eint vector in pcreposix.c.
1416
1417 9. Error texts for pcre_compile() are held as one long string to avoid too
1418 much relocation at load time. To find a text, the string is searched,
1419 counting zeros. There was no check for running off the end of the string,
1420 which could happen if a new error number was added without updating the
1421 string.
1422
1423 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1424
1425 11. \K was not working if it appeared in an atomic group or in a group that
1426 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1427 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1428 the assertion is positive, but not if it is negative.
1429
1430 12. Change 11 fortuitously reduced the size of the stack frame used in the
1431 "match()" function of pcre_exec.c by one pointer. Forthcoming
1432 implementation of support for (*MARK) will need an extra pointer on the
1433 stack; I have reserved it now, so that the stack frame size does not
1434 decrease.
1435
1436 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1437 item in branch that calls a recursion is a subroutine call - as in the
1438 second branch in the above example - was incorrectly given the compile-
1439 time error "recursive call could loop indefinitely" because pcre_compile()
1440 was not correctly checking the subroutine for matching a non-empty string.
1441
1442 14. The checks for overrunning compiling workspace could trigger after an
1443 overrun had occurred. This is a "should never occur" error, but it can be
1444 triggered by pathological patterns such as hundreds of nested parentheses.
1445 The checks now trigger 100 bytes before the end of the workspace.
1446
1447 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1448
1449
1450 Version 8.01 19-Jan-2010
1451 ------------------------
1452
1453 1. If a pattern contained a conditional subpattern with only one branch (in
1454 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1455 computed the wrong minimum data length (which is of course zero for such
1456 subpatterns). This could cause incorrect "no match" results.
1457
1458 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1459 the pattern is reset in the first branch, pcre_compile() failed with
1460 "internal error: code overflow at offset...". This happened only when
1461 the reset was to the original external option setting. (An optimization
1462 abstracts leading options settings into an external setting, which was the
1463 cause of this.)
1464
1465 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1466 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1467 assertion pattern did not match (meaning that the assertion was true), it
1468 was incorrectly treated as false if the SKIP had been reached during the
1469 matching. This also applied to assertions used as conditions.
1470
1471 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1472 assertion subpattern, including such a pattern used as a condition,
1473 unpredictable results occurred, instead of the error return
1474 PCRE_ERROR_DFA_UITEM.
1475
1476 5. The C++ GlobalReplace function was not working like Perl for the special
1477 situation when an empty string is matched. It now does the fancy magic
1478 stuff that is necessary.
1479
1480 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1481 removed. (These were left over from very, very early versions of PCRE.)
1482
1483 7. Some cosmetic changes to the code to make life easier when compiling it
1484 as part of something else:
1485
1486 (a) Change DEBUG to PCRE_DEBUG.
1487
1488 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1489 called "current" as "current_branch", to prevent a collision with the
1490 Linux macro when compiled as a kernel module.
1491
1492 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1493 prevent a collision with the Linux macro when compiled as a kernel
1494 module.
1495
1496 8. In pcre_compile() there are some checks for integer overflows that used to
1497 cast potentially large values to (double). This has been changed to that
1498 when building, a check for int64_t is made, and if it is found, it is used
1499 instead, thus avoiding the use of floating point arithmetic. (There is no
1500 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1501 double.
1502
1503 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1504 2005 (difference between two addresses compared to an unsigned value).
1505
1506 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1507 custom one, because of the following reported problem in Windows:
1508
1509 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1510 under Win32.
1511 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1512 therefore missing the function definition.
1513 - The compiler thus generates a "C" signature for the test function.
1514 - The linker fails to find the "C" function.
1515 - PCRE fails to configure if asked to do so against libbz2.
1516
1517 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1518 messages were output:
1519
1520 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1521 rerunning libtoolize, to keep the correct libtool macros in-tree.
1522 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1523
1524 I have done both of these things.
1525
1526 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1527 most of the time, it *can* run out if it is given a pattern that contains a
1528 runaway infinite recursion. I updated the discussion in the pcrestack man
1529 page.
1530
1531 13. Now that we have gone to the x.xx style of version numbers, the minor
1532 version may start with zero. Using 08 or 09 is a bad idea because users
1533 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1534 interpreted as invalid octal numbers. I've updated the previous comment in
1535 configure.ac, and also added a check that gives an error if 08 or 09 are
1536 used.
1537
1538 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1539 causing partial matching to fail when the end of the subject matched \W
1540 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1541
1542 15. There were some discrepancies between the declarations in pcre_internal.h
1543 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1544 their definitions. The declarations used "const uschar *" and the
1545 definitions used USPTR. Even though USPTR is normally defined as "const
1546 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1547 reported that: "This difference in casting confuses some C++ compilers, for
1548 example, SunCC recognizes above declarations as different functions and
1549 generates broken code for hbpcre." I have changed the declarations to use
1550 USPTR.
1551
1552 16. GNU libtool is named differently on some systems. The autogen.sh script now
1553 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1554 (FreeBSD).
1555
1556 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1557 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1558 comment: "Figure out how to create a longlong from a string: strtoll and
1559 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1560 instance, but it only takes 2 args instead of 3!"
1561
1562 18. A subtle bug concerned with back references has been fixed by a change of
1563 specification, with a corresponding code fix. A pattern such as
1564 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1565 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1566 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1567 same bug. Such groups have to be quantified to be useful, or contained
1568 inside another quantified group. (If there's no repetition, the reference
1569 can never match.) The problem arises because, having left the group and
1570 moved on to the rest of the pattern, a later failure that backtracks into
1571 the group uses the captured value from the final iteration of the group
1572 rather than the correct earlier one. I have fixed this in PCRE by forcing
1573 any group that contains a reference to itself to be an atomic group; that
1574 is, there cannot be any backtracking into it once it has completed. This is
1575 similar to recursive and subroutine calls.
1576
1577
1578 Version 8.00 19-Oct-09
1579 ----------------------
1580
1581 1. The table for translating pcre_compile() error codes into POSIX error codes
1582 was out-of-date, and there was no check on the pcre_compile() error code
1583 being within the table. This could lead to an OK return being given in
1584 error.
1585
1586 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1587 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1588 in a Windows environment.
1589
1590 3. The pcregrep --count option prints the count for each file even when it is
1591 zero, as does GNU grep. However, pcregrep was also printing all files when
1592 --files-with-matches was added. Now, when both options are given, it prints
1593 counts only for those files that have at least one match. (GNU grep just
1594 prints the file name in this circumstance, but including the count seems
1595 more useful - otherwise, why use --count?) Also ensured that the
1596 combination -clh just lists non-zero counts, with no names.
1597
1598 4. The long form of the pcregrep -F option was incorrectly implemented as
1599 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1600 but it seems right to fix it, and I didn't think it was worth preserving
1601 the old behaviour.
1602
1603 5. The command line items --regex=pattern and --regexp=pattern were not
1604 recognized by pcregrep, which required --regex pattern or --regexp pattern
1605 (with a space rather than an '='). The man page documented the '=' forms,
1606 which are compatible with GNU grep; these now work.
1607
1608 6. No libpcreposix.pc file was created for pkg-config; there was just
1609 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1610
1611 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1612 when UCP support is not needed, by modifying the Python script that
1613 generates it from Unicode data files. This should not matter if the module
1614 is correctly used as a library, but I received one complaint about 50K of
1615 unwanted data. My guess is that the person linked everything into his
1616 program rather than using a library. Anyway, it does no harm.
1617
1618 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1619 was a minimum greater than 1 for a wide character in a possessive
1620 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1621 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1622 character. Chaos in the form of incorrect output or a compiling loop could
1623 result.
1624
1625 9. The restrictions on what a pattern can contain when partial matching is
1626 requested for pcre_exec() have been removed. All patterns can now be
1627 partially matched by this function. In addition, if there are at least two
1628 slots in the offset vector, the offset of the earliest inspected character
1629 for the match and the offset of the end of the subject are set in them when
1630 PCRE_ERROR_PARTIAL is returned.
1631
1632 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1633 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1634 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1635 and may be more useful for multi-segment matching.
1636
1637 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1638 used to be given if ever the end of the subject was reached; now it is
1639 given only if matching could not proceed because another character was
1640 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1641 string "Z", which now yields "no match" instead of "partial match". In the
1642 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1643 final character ended with (*FAIL).
1644
1645 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1646 if the pattern had a "must contain" character that was already found in the
1647 earlier partial match, unless partial matching was again requested. For
1648 example, with the pattern /dog.(body)?/, the "must contain" character is
1649 "g". If the first part-match was for the string "dog", restarting with
1650 "sbody" failed. This bug has been fixed.
1651
1652 13. The string returned by pcre_dfa_exec() after a partial match has been
1653 changed so that it starts at the first inspected character rather than the
1654 first character of the match. This makes a difference only if the pattern
1655 starts with a lookbehind assertion or \b or \B (\K is not supported by
1656 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1657 matching functions compatible, and I think it's the right thing to do.
1658
1659 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1660 so that the demonstration program is easily available in environments where
1661 PCRE has not been installed from source.
1662
1663 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1664 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1665 library.
1666
1667 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1668 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1669 is not the first non-POSIX option to be added. Clearly some people find
1670 these options useful.
1671
1672 17. If a caller to the POSIX matching function regexec() passes a non-zero
1673 value for nmatch with a NULL value for pmatch, the value of
1674 nmatch is forced to zero.
1675
1676 18. RunGrepTest did not have a test for the availability of the -u option of
1677 the diff command, as RunTest does. It now checks in the same way as
1678 RunTest, and also checks for the -b option.
1679
1680 19. If an odd number of negated classes containing just a single character
1681 interposed, within parentheses, between a forward reference to a named
1682 subpattern and the definition of the subpattern, compilation crashed with
1683 an internal error, complaining that it could not find the referenced
1684 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1685 [The bug was that it was starting one character too far in when skipping
1686 over the character class, thus treating the ] as data rather than
1687 terminating the class. This meant it could skip too much.]
1688
1689 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1690 /g option in pcretest when the pattern contains \K, which makes it possible
1691 to have an empty string match not at the start, even when the pattern is
1692 anchored. Updated pcretest and pcredemo to use this option.
1693
1694 21. If the maximum number of capturing subpatterns in a recursion was greater
1695 than the maximum at the outer level, the higher number was returned, but
1696 with unset values at the outer level. The correct (outer level) value is
1697 now given.
1698
1699 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1700 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1701 make it do so. The string so far is captured, making this feature
1702 compatible with Perl.
1703
1704 23. The tests have been re-organized, adding tests 11 and 12, to make it
1705 possible to check the Perl 5.10 features against Perl 5.10.
1706
1707 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1708 pattern matches a fixed length string. PCRE did not allow this; now it
1709 does. Neither allows recursion.
1710
1711 25. I finally figured out how to implement a request to provide the minimum
1712 length of subject string that was needed in order to match a given pattern.
1713 (It was back references and recursion that I had previously got hung up
1714 on.) This code has now been added to pcre_study(); it finds a lower bound
1715 to the length of subject needed. It is not necessarily the greatest lower
1716 bound, but using it to avoid searching strings that are too short does give
1717 some useful speed-ups. The value is available to calling programs via
1718 pcre_fullinfo().
1719
1720 26. While implementing 25, I discovered to my embarrassment that pcretest had
1721 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1722 study optimizations had never been tested with that matching function.
1723 Oops. What is worse, even when it was passed study data, there was a bug in
1724 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1725 were also very few tests of studied patterns with pcre_dfa_exec().
1726
1727 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1728 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1729 on the other side of the coin, they are no longer allowed to have different
1730 names, because these cannot be distinguished in PCRE, and this has caused
1731 confusion. (This is a difference from Perl.)
1732
1733 28. When duplicate subpattern names are present (necessarily with different
1734 numbers, as required by 27 above), and a test is made by name in a
1735 conditional pattern, either for a subpattern having been matched, or for
1736 recursion in such a pattern, all the associated numbered subpatterns are
1737 tested, and the overall condition is true if the condition is true for any
1738 one of them. This is the way Perl works, and is also more like the way
1739 testing by number works.
1740
1741
1742 Version 7.9 11-Apr-09
1743 ---------------------
1744
1745 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1746 (pcretest), all targets were linked against these libraries. This included
1747 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1748 libraries. This caused unwanted dependencies to be created. This problem
1749 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1750 pcretest is linked with readline.
1751
1752 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1753 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1754 moved outside it again, because FALSE and TRUE are already defined in AIX,
1755 but BOOL is not.
1756
1757 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1758 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1759
1760 4. The pcregrep documentation said spaces were inserted as well as colons (or
1761 hyphens) following file names and line numbers when outputting matching
1762 lines. This is not true; no spaces are inserted. I have also clarified the
1763 wording for the --colour (or --color) option.
1764
1765 5. In pcregrep, when --colour was used with -o, the list of matching strings
1766 was not coloured; this is different to GNU grep, so I have changed it to be
1767 the same.
1768
1769 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1770 each matching line was coloured. Now it goes on to look for further matches
1771 of any of the test patterns, which is the same behaviour as GNU grep.
1772
1773 7. A pattern that could match an empty string could cause pcregrep to loop; it
1774 doesn't make sense to accept an empty string match in pcregrep, so I have
1775 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1776 seems to be how GNU grep behaves. [But see later change 40 for release
1777 8.33.]
1778
1779 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1780 start or after a newline", because the conditional assertion was not being
1781 correctly handled. The rule now is that both the assertion and what follows
1782 in the first alternative must satisfy the test.
1783
1784 9. If auto-callout was enabled in a pattern with a conditional group whose
1785 condition was an assertion, PCRE could crash during matching, both with
1786 pcre_exec() and pcre_dfa_exec().
1787
1788 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1789 used for matching.
1790
1791 11. Unicode property support in character classes was not working for
1792 characters (bytes) greater than 127 when not in UTF-8 mode.
1793
1794 12. Added the -M command line option to pcretest.
1795
1796 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1797
1798 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1799
1800 16. Added comments and documentation about mis-use of no_arg in the C++
1801 wrapper.
1802
1803 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1804 from Martin Jerabek that uses macro names for all relevant character and
1805 string constants.
1806
1807 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1808 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1809 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1810 these, but not everybody uses configure.
1811
1812 19. A conditional group that had only one branch was not being correctly
1813 recognized as an item that could match an empty string. This meant that an
1814 enclosing group might also not be so recognized, causing infinite looping
1815 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1816 with the subject "ab", where knowledge that the repeated group can match
1817 nothing is needed in order to break the loop.
1818
1819 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1820 exec(), but without supplying a callout function, matching went wrong.
1821
1822 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1823 leak if the size of the offset vector was greater than 30. When the vector
1824 is smaller, the saved offsets during recursion go onto a local stack
1825 vector, but for larger vectors malloc() is used. It was failing to free
1826 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1827 error, in fact).
1828
1829 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1830 heapframe that is used only when UTF-8 support is enabled. This caused no
1831 problem, but was untidy.
1832
1833 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1834 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1835 included within another project.
1836
1837 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1838 slightly modified by me:
1839
1840 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1841 not building pcregrep.
1842
1843 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1844 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1845
1846 25. Forward references, both numeric and by name, in patterns that made use of
1847 duplicate group numbers, could behave incorrectly or give incorrect errors,
1848 because when scanning forward to find the reference group, PCRE was not
1849 taking into account the duplicate group numbers. A pattern such as
1850 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1851
1852 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1853 the feature of a custom pointer more persuasive (as requested by a user).
1854
1855 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1856 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1857 pre-defined.
1858
1859 28. Added support for (*UTF8) at the start of a pattern.
1860
1861 29. Arrange for flags added by the "release type" setting in CMake to be shown
1862 in the configuration summary.
1863
1864
1865 Version 7.8 05-Sep-08
1866 ---------------------
1867
1868 1. Replaced UCP searching code with optimized version as implemented for Ad
1869 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1870 stage table and inline lookup instead of a function, giving speed ups of 2
1871 to 5 times on some simple patterns that I tested. Permission was given to
1872 distribute the MultiStage2.py script that generates the tables (it's not in
1873 the tarball, but is in the Subversion repository).
1874
1875 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1876 scripts.
1877
1878 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1879 a group with a zero qualifier. The result of the study could be incorrect,
1880 or the function might crash, depending on the pattern.
1881
1882 4. Caseless matching was not working for non-ASCII characters in back
1883 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1884 It now works when Unicode Property Support is available.
1885
1886 5. In pcretest, an escape such as \x{de} in the data was always generating
1887 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1888 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1889 truncation.
1890
1891 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1892
1893 7. Added two (int) casts to pcregrep when printing the difference of two
1894 pointers, in case they are 64-bit values.
1895
1896 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1897 test 2 if it fails.
1898
1899 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1900 and a #define of that name to empty if it is not externally set. This is to
1901 allow users of MSVC to set it if necessary.
1902
1903 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1904 the convenience functions in the pcre_get.c source file.
1905
1906 11. An option change at the start of a pattern that had top-level alternatives
1907 could cause overwriting and/or a crash. This command provoked a crash in
1908 some environments:
1909
1910 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1911
1912 This potential security problem was recorded as CVE-2008-2371.
1913
1914 12. For a pattern where the match had to start at the beginning or immediately
1915 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1916 pcre_dfa_exec() could read past the end of the passed subject if there was
1917 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1918 pcretest so that it places the subject at the end of its malloc-ed buffer.
1919
1920 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1921 exec() might read past the end of the data buffer in UTF-8 mode.
1922
1923 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1924 the data contained the byte 0x85 as part of a UTF-8 character within its
1925 first line. This applied both to normal and DFA matching.
1926
1927 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1928 /^[^d]*?$/8 failed to match "abc".
1929
1930 16. Added a missing copyright notice to pcrecpp_internal.h.
1931
1932 17. Make it more clear in the documentation that values returned from
1933 pcre_exec() in ovector are byte offsets, not character counts.
1934
1935 18. Tidied a few places to stop certain compilers from issuing warnings.
1936
1937 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1938 supplied by Stefan Weber. I made a further small update for 7.8 because
1939 there is a change of source arrangements: the pcre_searchfuncs.c module is
1940 replaced by pcre_ucd.c.
1941
1942
1943 Version 7.7 07-May-08
1944 ---------------------
1945
1946 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1947 a string to a long long, pretend we don't even have a long long." This is
1948 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1949
1950 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1951 pre-7.6 versions, which defined a global no_arg variable instead of putting
1952 it in the RE class. (See also #8 below.)
1953
1954 3. Remove a line of dead code, identified by coverity and reported by Nuno
1955 Lopes.
1956
1957 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1958
1959 (1) The include/exclude patterns were being applied to the whole pathnames
1960 of files, instead of just to the final components.
1961
1962 (2) If there was more than one level of directory, the subdirectories were
1963 skipped unless they satisfied the include/exclude conditions. This is
1964 inconsistent with GNU grep (and could even be seen as contrary to the
1965 pcregrep specification - which I improved to make it absolutely clear).
1966 The action now is always to scan all levels of directory, and just
1967 apply the include/exclude patterns to regular files.
1968
1969 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1970 --exclude_dir in the tests to avoid scanning .svn directories.
1971
1972 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1973 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1974 doesn't support NULs in patterns.
1975
1976 7. Added some missing "const"s to declarations of static tables in
1977 pcre_compile.c and pcre_dfa_exec.c.
1978
1979 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1980 caused by fix #2 above. (Subsequently also a second patch to fix the
1981 first patch. And a third patch - this was a messy problem.)
1982
1983 9. Applied Craig's patch to remove the use of push_back().
1984
1985 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1986 matching function regexec().
1987
1988 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1989 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1990 references. PCRE supports relative numbers with this syntax (I don't think
1991 Oniguruma does).
1992
1993 12. Previously, a group with a zero repeat such as (...){0} was completely
1994 omitted from the compiled regex. However, this means that if the group
1995 was called as a subroutine from elsewhere in the pattern, things went wrong
1996 (an internal error was given). Such groups are now left in the compiled
1997 pattern, with a new opcode that causes them to be skipped at execution
1998 time.
1999
2000 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2001 to the way PCRE behaves:
2002
2003 (a) A lone ] character is dis-allowed (Perl treats it as data).
2004
2005 (b) A back reference to an unmatched subpattern matches an empty string
2006 (Perl fails the current match path).
2007
2008 (c) A data ] in a character class must be notated as \] because if the
2009 first data character in a class is ], it defines an empty class. (In
2010 Perl it is not possible to have an empty class.) The empty class []
2011 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2012 The negative empty class [^] matches any one character, independently
2013 of the DOTALL setting.
2014
2015 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2016 non-existent subpattern following a character class starting with ']' and
2017 containing () gave an internal compiling error instead of "reference to
2018 non-existent subpattern". Fortunately, when the pattern did exist, the
2019 compiled code was correct. (When scanning forwards to check for the
2020 existence of the subpattern, it was treating the data ']' as terminating
2021 the class, so got the count wrong. When actually compiling, the reference
2022 was subsequently set up correctly.)
2023
2024 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2025 it was being rejected as not supported by pcre_dfa_exec(), even though
2026 other assertions are supported. I have made pcre_dfa_exec() support
2027 (*FAIL).
2028
2029 16. The implementation of 13c above involved the invention of a new opcode,
2030 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2031 cannot be changed at match time, I realized I could make a small
2032 improvement to matching performance by compiling OP_ALLANY instead of
2033 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2034 on the OP_ANY path.
2035
2036 17. Compiling pcretest on Windows with readline support failed without the
2037 following two fixes: (1) Make the unistd.h include conditional on
2038 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2039
2040 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2041 ncurses library to be included for pcretest when ReadLine support is
2042 requested, but also to allow for it to be overridden. This patch came from
2043 Daniel Bergström.
2044
2045 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2046 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2047 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2048 spotting this.
2049
2050
2051 Version 7.6 28-Jan-08
2052 ---------------------
2053
2054 1. A character class containing a very large number of characters with
2055 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2056 overflow.
2057
2058 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2059 HAVE_LONG_LONG is not defined.
2060
2061 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2062 bring them up to date and include new features. This patch includes:
2063
2064 - Fixed PH's badly added libz and libbz2 support.
2065 - Fixed a problem with static linking.
2066 - Added pcredemo. [But later removed - see 7 below.]
2067 - Fixed dftables problem and added an option.
2068 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2069 HAVE_LONG_LONG.
2070 - Added readline support for pcretest.
2071 - Added an listing of the option settings after cmake has run.
2072
2073 4. A user submitted a patch to Makefile that makes it easy to create
2074 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2075 Makefile.am that cause it to include this special target, without
2076 affecting anything else. Note that the same mingw target plus all
2077 the other distribution libraries and programs are now supported
2078 when configuring with CMake (see 6 below) instead of with
2079 Configure/Make.
2080
2081 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2082 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2083 exported in the Windows port". It has not yet been confirmed that the patch
2084 solves the problem, but it does no harm.
2085
2086 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2087 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2088 with CMake, and also correct the comment about stack recursion.
2089
2090 7. Remove the automatic building of pcredemo from the ./configure system and
2091 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2092 of a program that users should build themselves after PCRE is installed, so
2093 building it automatically is not really right. What is more, it gave
2094 trouble in some build environments.
2095
2096 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2097
2098
2099 Version 7.5 10-Jan-08
2100 ---------------------
2101
2102 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2103 values in parens when parsing an RE using the C++ wrapper."
2104
2105 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2106 Characters greater than 255 were excluded from the class instead of being
2107 included.
2108
2109 3. The same bug as (2) above applied to negated POSIX classes such as
2110 [:^space:].
2111
2112 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2113 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2114 I have changed it.
2115
2116 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2117 first named subpattern) and a construct such as (?&a) would reference the
2118 first named subpattern whose name started with "a" (in other words, the
2119 length check was missing). Both these problems are fixed. "Subpattern name
2120 expected" is now given for (?&) (a zero-length name), and this patch also
2121 makes it give the same error for \k'' (previously it complained that that
2122 was a reference to a non-existent subpattern).
2123
2124 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2125 this is right because (?- can be followed by option settings as well as by
2126 digits. I have, however, made the messages clearer.
2127
2128 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2129 than the number used in the conditional) now cause a compile-time error.
2130 This is actually not compatible with Perl, which accepts such patterns, but
2131 treats the conditional as always being FALSE (as PCRE used to), but it
2132 seems to me that giving a diagnostic is better.
2133
2134 8. Change "alphameric" to the more common word "alphanumeric" in comments
2135 and messages.
2136
2137 9. Fix two occurrences of "backslash" in comments that should have been
2138 "backspace".
2139
2140 10. Remove two redundant lines of code that can never be obeyed (their function
2141 was moved elsewhere).
2142
2143 11. The program that makes PCRE's Unicode character property table had a bug
2144 which caused it to generate incorrect table entries for sequences of
2145 characters that have the same character type, but are in different scripts.
2146 It amalgamated them into a single range, with the script of the first of
2147 them. In other words, some characters were in the wrong script. There were
2148 thirteen such cases, affecting characters in the following ranges:
2149
2150 U+002b0 - U+002c1
2151 U+0060c - U+0060d
2152 U+0061e - U+00612
2153 U+0064b - U+0065e
2154 U+0074d - U+0076d
2155 U+01800 - U+01805
2156 U+01d00 - U+01d77
2157 U+01d9b - U+01dbf
2158 U+0200b - U+0200f
2159 U+030fc - U+030fe
2160 U+03260 - U+0327f
2161 U+0fb46 - U+0fbb1
2162 U+10450 - U+1049d
2163
2164 12. The -o option (show only the matching part of a line) for pcregrep was not
2165 compatible with GNU grep in that, if there was more than one match in a
2166 line, it showed only the first of them. It now behaves in the same way as
2167 GNU grep.
2168
2169 13. If the -o and -v options were combined for pcregrep, it printed a blank
2170 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2171 does the same. The return code can be used to tell if there were any
2172 non-matching lines.
2173
2174 14. Added --file-offsets and --line-offsets to pcregrep.
2175
2176 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2177 infinitely looping recursion. The bug was that positive lookaheads were not
2178 being skipped when checking for a possible empty match (negative lookaheads
2179 and both kinds of lookbehind were skipped).
2180
2181 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2182 inclusion of <windows.h> to before rather than after the definition of
2183 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2184
2185 17. Specifying a possessive quantifier with a specific limit for a Unicode
2186 character property caused pcre_compile() to compile bad code, which led at
2187 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2188 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2189 caused the error; without that there was no problem.
2190
2191 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2192
2193 19. Added --enable-pcretest-libreadline.
2194
2195 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2196 RE::GlobalReplace(). As a result, the number of replacements returned was
2197 double what it should be. I removed one of the increments, but Craig sent a
2198 later patch that removed the other one (the right fix) and added unit tests
2199 that check the return values (which was not done before).
2200
2201 21. Several CMake things:
2202
2203 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2204 the names libpcre and libpcreposix, not just pcre and pcreposix.
2205
2206 (2) The above change means that pcretest and pcregrep are now correctly
2207 linked with the newly-built libraries, not previously installed ones.
2208
2209 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2210
2211 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2212 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2213 UTF-8 newline character). The key issue is that the pattern starts .*;
2214 this means that the match must be either at the beginning, or after a
2215 newline. The bug was in the code for advancing after a failed match and
2216 checking that the new position followed a newline. It was not taking
2217 account of UTF-8 characters correctly.
2218
2219 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2220 character classes. PCRE was not treating the sequence [:...:] as a
2221 character class unless the ... were all letters. Perl, however, seems to
2222 allow any characters between [: and :], though of course it rejects as
2223 unknown any "names" that contain non-letters, because all the known class
2224 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2225 for example, whereas PCRE did not - it did not recognize a POSIX character
2226 class. This seemed a bit dangerous, so the code has been changed to be
2227 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2228 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2229 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2230 Perl does, and where it didn't before.
2231
2232 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2233 Windows environments %n is disabled by default.
2234
2235
2236 Version 7.4 21-Sep-07
2237 ---------------------
2238
2239 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2240 means that a class such as [\s] counted as "explicit reference to CR or
2241 LF". That isn't really right - the whole point of the change was to try to
2242 help when there was an actual mention of one of the two characters. So now
2243 the change happens only if \r or \n (or a literal CR or LF) character is
2244 encountered.
2245
2246 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2247 of both had grown to the point where there were only 3 bits left.
2248 Fortunately, there was spare space in the data structure, and so I have
2249 moved the internal flags into a new 16-bit field to free up more option
2250 bits.
2251
2252 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2253 but did not set the internal JCHANGED flag - either of these is enough to
2254 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2255 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2256 start sets both bits.
2257
2258 4. Added options (at build time, compile time, exec time) to change \R from
2259 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2260
2261 5. doc/pcresyntax.html was missing from the distribution.
2262
2263 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2264 compatibility, even though it is no longer used.
2265
2266 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2267 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2268 windows.h file is present (where different names are used). [This was
2269 reversed later after testing - see 16 below.]
2270
2271 8. Changed all #include <config.h> to #include "config.h". There were also
2272 some further <pcre.h> cases that I changed to "pcre.h".
2273
2274 9. When pcregrep was used with the --colour option, it missed the line ending
2275 sequence off the lines that it output.
2276
2277 10. It was pointed out to me that arrays of string pointers cause lots of
2278 relocations when a shared library is dynamically loaded. A technique of
2279 using a single long string with a table of offsets can drastically reduce
2280 these. I have refactored PCRE in four places to do this. The result is
2281 dramatic:
2282
2283 Originally: 290
2284 After changing UCP table: 187
2285 After changing error message table: 43
2286 After changing table of "verbs" 36
2287 After changing table of Posix names 22
2288
2289 Thanks to the folks working on Gregex for glib for this insight.
2290
2291 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2292 unicode-properties was also set.
2293
2294 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2295
2296 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2297 checked only for CRLF.
2298
2299 14. Added casts to pcretest.c to avoid compiler warnings.
2300
2301 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2302
2303 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2304 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2305 entirely. This removes changes made in 7 above.
2306
2307 17. The CMake files have been updated, and there is now more information about
2308 building with CMake in the NON-UNIX-USE document.
2309
2310
2311 Version 7.3 28-Aug-07
2312 ---------------------
2313
2314 1. In the rejigging of the build system that eventually resulted in 7.1, the
2315 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2316 brackets there is not right, since it causes compilers to look for an
2317 installed pcre.h, not the version that is in the source that is being
2318 compiled (which of course may be different). I have changed it back to:
2319
2320 #include "pcre.h"
2321
2322 I have a vague recollection that the change was concerned with compiling in
2323 different directories, but in the new build system, that is taken care of
2324 by the VPATH setting the Makefile.
2325
2326 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2327 when the subject happened to end in the byte 0x85 (e.g. if the last
2328 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2329 characters but of course it shouldn't be taken as a newline when it is part
2330 of another character. The bug was that, for an unlimited repeat of . in
2331 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2332 characters when looking for a newline.
2333
2334 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2335
2336 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2337 in debug output.
2338
2339 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2340 long printing in the pcrecpp unittest when running under MinGW.
2341
2342 6. ESC_K was left out of the EBCDIC table.
2343
2344 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2345 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2346 limit also applies to "virtual nesting" when a pattern is recursive, and in
2347 this case 1000 isn't so big. I have been able to remove this limit at the
2348 expense of backing off one optimization in certain circumstances. Normally,
2349 when pcre_exec() would call its internal match() function recursively and
2350 immediately return the result unconditionally, it uses a "tail recursion"
2351 feature to save stack. However, when a subpattern that can match an empty
2352 string has an unlimited repetition quantifier, it no longer makes this
2353 optimization. That gives it a stack frame in which to save the data for
2354 checking that an empty string has been matched. Previously this was taken
2355 from the 1000-entry workspace that had been reserved. So now there is no
2356 explicit limit, but more stack is used.
2357
2358 8. Applied Daniel's patches to solve problems with the import/export magic
2359 syntax that is required for Windows, and which was going wrong for the
2360 pcreposix and pcrecpp parts of the library. These were overlooked when this
2361 problem was solved for the main library.
2362
2363 9. There were some crude static tests to avoid integer overflow when computing
2364 the size of patterns that contain repeated groups with explicit upper
2365 limits. As the maximum quantifier is 65535, the maximum group length was
2366 set at 30,000 so that the product of these two numbers did not overflow a
2367 32-bit integer. However, it turns out that people want to use groups that
2368 are longer than 30,000 bytes (though not repeat them that many times).
2369 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2370 made it possible to implement the integer overflow checks in a much more
2371 dynamic way, which I have now done. The artificial limitation on group
2372 length has been removed - we now have only the limit on the total length of
2373 the compiled pattern, which depends on the LINK_SIZE setting.
2374
2375 10. Fixed a bug in the documentation for get/copy named substring when
2376 duplicate names are permitted. If none of the named substrings are set, the
2377 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2378 empty string.
2379
2380 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2381 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2382 because the ] is interpreted as the first data character and the
2383 terminating ] is not found. PCRE has been made compatible with Perl in this
2384 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2385 cause memory overwriting.
2386
2387 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2388 string has been matched (to stop an infinite loop). It was not recognizing
2389 a conditional subpattern that could match an empty string if that
2390 subpattern was within another subpattern. For example, it looped when
2391 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2392 condition was not nested. This bug has been fixed.
2393
2394 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2395 past the start of the subject in the presence of bytes with the top bit
2396 set, for example "\x8aBCD".
2397
2398 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2399 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2400
2401 14. Optimized (?!) to (*FAIL).
2402
2403 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2404 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2405 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2406 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2407 does: it's just the validity check that is more restrictive.
2408
2409 16. Inserted checks for integer overflows during escape sequence (backslash)
2410 processing, and also fixed erroneous offset values for syntax errors during
2411 backslash processing.
2412
2413 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2414 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2415
2416 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2417 caused an overrun.
2418
2419 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2420 something other than just ASCII characters) inside a group that had an
2421 unlimited repeat caused a loop at compile time (while checking to see
2422 whether the group could match an empty string).
2423
2424 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2425 [\P{Any}] did so. (Error in the code for printing property names.)
2426
2427 21. An orphan \E inside a character class could cause a crash.
2428
2429 22. A repeated capturing bracket such as (A)? could cause a wild memory
2430 reference during compilation.
2431
2432 23. There are several functions in pcre_compile() that scan along a compiled
2433 expression for various reasons (e.g. to see if it's fixed length for look
2434 behind). There were bugs in these functions when a repeated \p or \P was
2435 present in the pattern. These operators have additional parameters compared
2436 with \d, etc, and these were not being taken into account when moving along
2437 the compiled data. Specifically:
2438
2439 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2440 length.
2441
2442 (b) An item such as \pL+ within a repeated group could cause crashes or
2443 loops.
2444
2445 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2446 "reference to non-existent subpattern" error.
2447
2448 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2449
2450 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2451 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2452
2453 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2454
2455 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2456 character were causing crashes (broken optimization).
2457
2458 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2459 \p or \P) caused a compile-time loop.
2460
2461 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2462 break. For example, the unstudied pattern [\r\n]A does not match the string
2463 "\r\nA" because change 7.0/46 below moves the current point on by two
2464 characters after failing to match at the start. However, the pattern \nA
2465 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2466 the same is true. There doesn't seem any very clean way out of this, but
2467 what I have chosen to do makes the common cases work: PCRE now takes note
2468 of whether there can be an explicit match for \r or \n anywhere in the
2469 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2470 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2471 pattern has explicit CR or LF references.
2472
2473 29. Added (*CR) etc for changing newline setting at start of pattern.
2474
2475
2476 Version 7.2 19-Jun-07
2477 ---------------------
2478
2479 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2480 which is apparently normally available under Windows.
2481
2482 2. Re-jig the pcregrep tests with different newline settings in an attempt
2483 to make them independent of the local environment's newline setting.
2484
2485 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2486
2487 4. Some of the "internals" tests were previously cut out when the link size
2488 was not 2, because the output contained actual offsets. The recent new
2489 "Z" feature of pcretest means that these can be cut out, making the tests
2490 usable with all link sizes.
2491
2492 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2493 stack recursion. This gives a massive performance boost under BSD, but just
2494 a small improvement under Linux. However, it saves one field in the frame
2495 in all cases.
2496
2497 6. Added more features from the forthcoming Perl 5.10:
2498
2499 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2500 recursion call. It refers to the nth most recently opened parentheses.
2501
2502 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2503 to be opened parentheses.
2504
2505 (c) Conditions that refer to capturing parentheses can be specified
2506 relatively, for example, (?(-2)... or (?(+3)...
2507
2508 (d) \K resets the start of the current match so that everything before
2509 is not part of it.
2510
2511 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2512
2513 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2514 reference syntax.
2515
2516 (g) (?| introduces a group in which the numbering of parentheses in each
2517 alternative starts with the same number.
2518
2519 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2520
2521 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2522 PCRE_INFO_JCHANGED.
2523
2524 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2525 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2526 for detecting groups that can match an empty string.
2527
2528 9. A pattern with a very large number of alternatives (more than several
2529 hundred) was running out of internal workspace during the pre-compile
2530 phase, where pcre_compile() figures out how much memory will be needed. A
2531 bit of new cunning has reduced the workspace needed for groups with
2532 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2533 workspace instead of running out of the 4096 that are available.
2534
2535 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2536
2537 11. Applied patch from Google to remove an optimization that didn't quite work.
2538 The report of the bug said:
2539
2540 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2541 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2542 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2543
2544 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2545 it matched the wrong number of bytes.
2546
2547
2548 Version 7.1 24-Apr-07
2549 ---------------------
2550
2551 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2552 that is more "standard", making use of automake and other Autotools. There
2553 is some re-arrangement of the files and adjustment of comments consequent
2554 on this.
2555
2556 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2557 for recursive directory scanning broke on some systems because the files
2558 are not scanned in any specific order and on different systems the order
2559 was different. A call to "sort" has been inserted into RunGrepTest for the
2560 approprate test as a short-term fix. In the longer term there may be an
2561 alternative.
2562
2563 3. I had an email from Eric Raymond about problems translating some of PCRE's
2564 man pages to HTML (despite the fact that I distribute HTML pages, some
2565 people do their own conversions for various reasons). The problems
2566 concerned the use of low-level troff macros .br and .in. I have therefore
2567 removed all such uses from the man pages (some were redundant, some could
2568 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2569 HTML has been updated to handle .nf/.fi and to complain if it encounters
2570 .br or .in.
2571
2572 4. Updated comments in configure.ac that get placed in config.h.in and also
2573 arranged for config.h to be included in the distribution, with the name
2574 config.h.generic, for the benefit of those who have to compile without
2575 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2576
2577 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2578 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2579 makevp.bat for the current PCRE, using the additional files
2580 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2581
2582 6. A Windows user reported a minor discrepancy with test 2, which turned out
2583 to be caused by a trailing space on an input line that had got lost in his
2584 copy. The trailing space was an accident, so I've just removed it.
2585
2586 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2587 that is needed.
2588
2589 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2590 as "const" (a) because they are and (b) because it helps the PHP
2591 maintainers who have recently made a script to detect big data structures
2592 in the php code that should be moved to the .rodata section. I remembered
2593 to update Builducptable as well, so it won't revert if ucptable.h is ever
2594 re-created.
2595
2596 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2597 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2598 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2599 support is not required. This saves 1.5-2K of code, which is important in
2600 some applications.
2601
2602 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2603 so as not to refer to the tables, even though these functions will never be
2604 called when UTF-8 support is disabled. Otherwise there are problems with a
2605 shared library.
2606
2607 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2608
2609 (a) It was defining its arguments as char * instead of void *.
2610
2611 (b) It was assuming that all moves were upwards in memory; this was true
2612 a long time ago when I wrote it, but is no longer the case.
2613
2614 The emulated memove() is provided for those environments that have neither
2615 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2616 is clearly not the case, as these two bugs were recently reported.
2617
2618 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2619 and Detrail to create the HTML documentation, the .txt form of the man
2620 pages, and it removes trailing spaces from listed files. It also creates
2621 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2622 case, it wraps all the #defines with #ifndefs. This script should be run
2623 before "make dist".
2624
2625 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2626 with Unicode property support.
2627
2628 (a) For a maximizing quantifier, if the two different cases of the
2629 character were of different lengths in their UTF-8 codings (there are
2630 some cases like this - I found 11), and the matching function had to
2631 back up over a mixture of the two cases, it incorrectly assumed they
2632 were both the same length.
2633
2634 (b) When PCRE was configured to use the heap rather than the stack for
2635 recursion during matching, it was not correctly preserving the data for
2636 the other case of a UTF-8 character when checking ahead for a match
2637 while processing a minimizing repeat. If the check also involved
2638 matching a wide character, but failed, corruption could cause an
2639 erroneous result when trying to check for a repeat of the original
2640 character.
2641
2642 13. Some tidying changes to the testing mechanism:
2643
2644 (a) The RunTest script now detects the internal link size and whether there
2645 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2646 values substituted by "configure". (The RunGrepTest script already did
2647 this for UTF-8.) The configure.ac script no longer substitutes the
2648 relevant variables.
2649
2650 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2651 with length and offset values. This means that the output is different
2652 for different internal link sizes. Test 2 is skipped for link sizes
2653 other than 2 because of this, bypassing the problem. Unfortunately,
2654 there was also a test in test 3 (the locale tests) that used /B and
2655 failed for link sizes other than 2. Rather than cut the whole test out,
2656 I have added a new /Z option to pcretest that replaces the length and
2657 offset values with spaces. This is now used to make test 3 independent
2658 of link size. (Test 2 will be tidied up later.)
2659
2660 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2661 segmentation fault instead of returning the appropriate error message.
2662
2663 15. In multiline mode when the newline sequence was set to "any", the pattern
2664 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2665 This doesn't seem right; it now treats the CRLF combination as the line
2666 ending, and so does not match in that case. It's only a pattern such as ^$
2667 that would hit this one: something like ^ABC$ would have failed after \r
2668 and then tried again after \r\n.
2669
2670 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2671 in an attempt to make files that differ only in their line terminators
2672 compare equal. This works on Linux.
2673
2674 17. Under certain error circumstances pcregrep might try to free random memory
2675 as it exited. This is now fixed, thanks to valgrind.
2676
2677 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2678 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2679 was because its rules for how to advance for /g after matching an empty
2680 string at the end of a line did not allow for this case. They now check for
2681 it specially.
2682
2683 20. pcretest is supposed to handle patterns and data of any length, by
2684 extending its buffers when necessary. It was getting this wrong when the
2685 buffer for a data line had to be extended.
2686
2687 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2688 CRLF as a newline sequence.
2689
2690 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2691 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2692 I have nevertheless tidied it up.
2693
2694 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2695
2696 24. Added a man page for pcre-config.
2697
2698
2699 Version 7.0 19-Dec-06
2700 ---------------------
2701
2702 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2703 moving to gcc 4.1.1.
2704
2705 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2706 sys/time.h, which is documented as needed for this function. It doesn't
2707 seem to matter on Linux, but it showed up on some releases of OS X.
2708
2709 3. It seems that there are systems where bytes whose values are greater than
2710 127 match isprint() in the "C" locale. The "C" locale should be the
2711 default when a C program starts up. In most systems, only ASCII printing
2712 characters match isprint(). This difference caused the output from pcretest
2713 to vary, making some of the tests fail. I have changed pcretest so that:
2714
2715 (a) When it is outputting text in the compiled version of a pattern, bytes
2716 other than 32-126 are always shown as hex escapes.
2717
2718 (b) When it is outputting text that is a matched part of a subject string,
2719 it does the same, unless a different locale has been set for the match
2720 (using the /L modifier). In this case, it uses isprint() to decide.
2721
2722 4. Fixed a major bug that caused incorrect computation of the amount of memory
2723 required for a compiled pattern when options that changed within the
2724 pattern affected the logic of the preliminary scan that determines the
2725 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2726 that the computed length was too small. The symptoms of this bug were
2727 either the PCRE error "internal error: code overflow" from pcre_compile(),
2728 or a glibc crash with a message such as "pcretest: free(): invalid next
2729 size (fast)". Examples of patterns that provoked this bug (shown in
2730 pcretest format) are:
2731
2732 /(?-x: )/x
2733 /(?x)(?-x: \s*#\s*)/
2734 /((?i)[\x{c0}])/8
2735 /(?i:[\x{c0}])/8
2736
2737 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2738 is now done differently.
2739
2740 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2741 wrapper classes; (b) implement a new function in the C++ scanner that is
2742 more efficient than the old way of doing things because it avoids levels of
2743 recursion in the regex matching; (c) add a paragraph to the documentation
2744 for the FullMatch() function.
2745
2746 6. The escape sequence \n was being treated as whatever was defined as
2747 "newline". Not only was this contrary to the documentation, which states
2748 that \n is character 10 (hex 0A), but it also went horribly wrong when
2749 "newline" was defined as CRLF. This has been fixed.
2750
2751 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2752 was being set to -1 for the "end of line" case (supposedly a value that no
2753 character can have). Though this value is never used (the check for end of
2754 line is "zero bytes in current character"), it caused compiler complaints.
2755 I've changed it to 0xffffffff.
2756
2757 8. In pcre_version.c, the version string was being built by a sequence of
2758 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2759 string (as it is for production releases) called a macro with an empty
2760 argument. The C standard says the result of this is undefined. The gcc
2761 compiler treats it as an empty string (which was what was wanted) but it is
2762 reported that Visual C gives an error. The source has been hacked around to
2763 avoid this problem.
2764
2765 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2766 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2767 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2768 of them did).
2769
2770 10. Originally, pcretest opened its input and output without "b"; then I was
2771 told that "b" was needed in some environments, so it was added for release
2772 5.0 to both the input and output. (It makes no difference on Unix-like
2773 systems.) Later I was told that it is wrong for the input on Windows. I've
2774 now abstracted the modes into two macros, to make it easier to fiddle with
2775 them, and removed "b" from the input mode under Windows.
2776
2777 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2778
2779 12. Added -help and --help to pcretest as an official way of being reminded
2780 of the options.
2781
2782 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2783 and pcrecpp.cc because they annoy compilers at high warning levels.
2784
2785 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2786
2787 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2788 scripts are not C programs :-) and which was not noticed because it works
2789 on Linux.
2790
2791 16. pcretest is supposed to handle any length of pattern and data line (as one
2792 line or as a continued sequence of lines) by extending its input buffer if
2793 necessary. This feature was broken for very long pattern lines, leading to
2794 a string of junk being passed to pcre_compile() if the pattern was longer
2795 than about 50K.
2796
2797 17. I have done a major re-factoring of the way pcre_compile() computes the
2798 amount of memory needed for a compiled pattern. Previously, there was code
2799 that made a preliminary scan of the pattern in order to do this. That was
2800 OK when PCRE was new, but as the facilities have expanded, it has become
2801 harder and harder to keep it in step with the real compile phase, and there
2802 have been a number of bugs (see for example, 4 above). I have now found a
2803 cunning way of running the real compile function in a "fake" mode that
2804 enables it to compute how much memory it would need, while actually only
2805 ever using a few hundred bytes of working memory and without too many
2806 tests of the mode. This should make future maintenance and development
2807 easier. A side effect of this work is that the limit of 200 on the nesting
2808 depth of parentheses has been removed (though this was never a serious
2809 limitation, I suspect). However, there is a downside: pcre_compile() now
2810 runs more slowly than before (30% or more, depending on the pattern). I
2811 hope this isn't a big issue. There is no effect on runtime performance.
2812
2813 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2814 newline (only possible for the last line of a file) and it was a
2815 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2816
2817 19. Added additional timing features to pcretest. (1) The -tm option now times
2818 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2819 separate command line item, by a number that specifies the number of
2820 repeats to use when timing. The default is 50000; this gives better
2821 precision, but takes uncomfortably long for very large patterns.
2822
2823 20. Extended pcre_study() to be more clever in cases where a branch of a
2824 subpattern has no definite first character. For example, (a*|b*)[cd] would
2825 previously give no result from pcre_study(). Now it recognizes that the
2826 first character must be a, b, c, or d.
2827
2828 21. There was an incorrect error "recursive call could loop indefinitely" if
2829 a subpattern (or the entire pattern) that was being tested for matching an
2830 empty string contained only one non-empty item after a nested subpattern.
2831 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2832 incorrectly, because the \d was being skipped in the check.
2833
2834 22. The pcretest program now has a new pattern option /B and a command line
2835 option -b, which is equivalent to adding /B to every pattern. This causes
2836 it to show the compiled bytecode, without the additional information that
2837 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2838 is the same as /B/I).
2839
2840 23. A new optimization is now able automatically to treat some sequences such
2841 as a*b as a*+b. More specifically, if something simple (such as a character
2842 or a simple class like \d) has an unlimited quantifier, and is followed by
2843 something that cannot possibly match the quantified thing, the quantifier
2844 is automatically "possessified".
2845
2846 24. A recursive reference to a subpattern whose number was greater than 39
2847 went wrong under certain circumstances in UTF-8 mode. This bug could also
2848 have affected the operation of pcre_study().
2849
2850 25. Realized that a little bit of performance could be had by replacing
2851 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2852
2853 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2854
2855 27. Possessive quantifiers such as a++ were previously implemented by turning
2856 them into atomic groups such as ($>a+). Now they have their own opcodes,
2857 which improves performance. This includes the automatically created ones
2858 from 23 above.
2859
2860 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2861 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2862 the first matched character to be a colon. This applied both to named and
2863 numbered groups.
2864
2865 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2866
2867 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2868 building PCRE on Windows easier, so I have included it in the distribution.
2869
2870 31. There is now a check in pcretest against a ridiculously large number being
2871 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2872 loop, the loop is abandoned.
2873
2874 32. Forward references to subpatterns in conditions such as (?(2)...) where
2875 subpattern 2 is defined later cause pcre_compile() to search forwards in
2876 the pattern for the relevant set of parentheses. This search went wrong
2877 when there were unescaped parentheses in a character class, parentheses
2878 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2879
2880 33. "Subroutine" calls and backreferences were previously restricted to
2881 referencing subpatterns earlier in the regex. This restriction has now
2882 been removed.
2883
2884 34. Added a number of extra features that are going to be in Perl 5.10. On the
2885 whole, these are just syntactic alternatives for features that PCRE had
2886 previously implemented using the Python syntax or my own invention. The
2887 other formats are all retained for compatibility.
2888
2889 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2890 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2891 also .NET compatible.
2892
2893 (b) A recursion or subroutine call to a named group can now be defined as
2894 (?&name) as well as (?P>name).
2895
2896 (c) A backreference to a named group can now be defined as \k<name> or
2897 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2898 5.10, are also .NET compatible.
2899
2900 (d) A conditional reference to a named group can now use the syntax
2901 (?(<name>) or (?('name') as well as (?(name).
2902
2903 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2904 groups (named and numbered) that are never evaluated inline, but can be
2905 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2906 is always false. There may be only one alternative in such a group.
2907
2908 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2909 as the simple (?(R). The condition is true only if the most recent
2910 recursion is that of the given number or name. It does not search out
2911 through the entire recursion stack.
2912
2913 (g) The escape \gN or \g{N} has been added, where N is a positive or
2914 negative number, specifying an absolute or relative reference.
2915
2916 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2917 some "unreachable code" warnings.
2918
2919 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2920 things, this adds five new scripts.
2921
2922 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2923 There were also incompatibilities regarding the handling of \Q..\E inside
2924 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2925 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2926
2927 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2928 matches an empty string, and forcibly breaks the loop. There were bugs in
2929 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2930 against aaaa the result was just "a" rather than "aaaa", for example. Two
2931 separate and independent bugs (that affected different cases) have been
2932 fixed.
2933
2934 39. Refactored the code to abolish the use of different opcodes for small
2935 capturing bracket numbers. This is a tidy that I avoided doing when I
2936 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2937 The new approach is not only tidier, it makes it possible to reduce the
2938 memory needed to fix the previous bug (38).
2939
2940 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2941 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2942 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2943 mode.
2944
2945 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2946 report.
2947
2948 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2949 copy construction and assignment in the C++ wrapper.
2950
2951 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2952 couple of bugs that could have given wrong results in the "--newline=crlf"
2953 case.
2954
2955 44. Added a number of casts and did some reorganization of signed/unsigned int
2956 variables following suggestions from Dair Grant. Also renamed the variable
2957 "this" as "item" because it is a C++ keyword.
2958
2959 45. Arranged for dftables to add
2960
2961 #include "pcre_internal.h"
2962
2963 to pcre_chartables.c because without it, gcc 4.x may remove the array
2964 definition from the final binary if PCRE is built into a static library and
2965 dead code stripping is activated.
2966
2967 46. For an unanchored pattern, if a match attempt fails at the start of a
2968 newline sequence, and the newline setting is CRLF or ANY, and the next two
2969 characters are CRLF, advance by two characters instead of one.
2970
2971
2972 Version 6.7 04-Jul-06
2973 ---------------------
2974
2975 1. In order to handle tests when input lines are enormously long, pcretest has
2976 been re-factored so that it automatically extends its buffers when
2977 necessary. The code is crude, but this _is_ just a test program. The
2978 default size has been increased from 32K to 50K.
2979
2980 2. The code in pcre_study() was using the value of the re argument before
2981 testing it for NULL. (Of course, in any sensible call of the function, it
2982 won't be NULL.)
2983
2984 3. The memmove() emulation function in pcre_internal.h, which is used on
2985 systems that lack both memmove() and bcopy() - that is, hardly ever -
2986 was missing a "static" storage class specifier.
2987
2988 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2989 containing an extended class (one that cannot be represented by a bitmap
2990 because it contains high-valued characters or Unicode property items, e.g.
2991 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2992 pattern, but PCRE should not loop if you do not (it no longer does).
2993 [Detail: two cases were found: (a) a repeated subpattern containing an
2994 extended class; (b) a recursive reference to a subpattern that followed a
2995 previous extended class. It wasn't skipping over the extended class
2996 correctly when UTF-8 mode was not set.]
2997
2998 5. A negated single-character class was not being recognized as fixed-length
2999 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3000 compile error "lookbehind assertion is not fixed length".
3001
3002 6. The RunPerlTest auxiliary script was showing an unexpected difference
3003 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3004 write a Perl script that can interpret lines of an input file either as
3005 byte characters or as UTF-8, which is what "perltest" was being required to
3006 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3007 can't do is switch easily at run time between having the "use utf8;" pragma
3008 or not. In the end, I fudged it by using the RunPerlTest script to insert
3009 "use utf8;" explicitly for the UTF-8 tests.
3010
3011 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3012 the end of the subject string, contrary to the documentation and to what
3013 Perl does. This was true of both matching functions. Now it matches only at
3014 the start of the subject and immediately after *internal* newlines.
3015
3016 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3017 a pointer to an int instead of a pointer to an unsigned long int. This
3018 caused problems on 64-bit systems.
3019
3020 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3021 instance of the 'standard' template library not being so standard".
3022
3023 10. There was no check on the number of named subpatterns nor the maximum
3024 length of a subpattern name. The product of these values is used to compute
3025 the size of the memory block for a compiled pattern. By supplying a very
3026 long subpattern name and a large number of named subpatterns, the size
3027 computation could be caused to overflow. This is now prevented by limiting
3028 the length of names to 32 characters, and the number of named subpatterns
3029 to 10,000.
3030
3031 11. Subpatterns that are repeated with specific counts have to be replicated in
3032 the compiled pattern. The size of memory for this was computed from the
3033 length of the subpattern and the repeat count. The latter is limited to
3034 65535, but there was no limit on the former, meaning that integer overflow
3035 could in principle occur. The compiled length of a repeated subpattern is
3036 now limited to 30,000 bytes in order to prevent this.
3037
3038 12. Added the optional facility to have named substrings with the same name.
3039
3040 13. Added the ability to use a named substring as a condition, using the
3041 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3042 are numbers (not recommended). Forward references are permitted.
3043
3044 14. Added forward references in named backreferences (if you see what I mean).
3045
3046 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3047 pattern could run off the end of the subject. For example, the pattern
3048 "(?s)(.{1,5})"8 did this with the subject "ab".
3049
3050 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3051 PCRE_CASELESS was set when matching characters that were quantified with ?
3052 or *.
3053
3054 17. A character class other than a single negated character that had a minimum
3055 but no maximum quantifier - for example [ab]{6,} - was not handled
3056 correctly by pce_dfa_exec(). It would match only one character.
3057
3058 18. A valid (though odd) pattern that looked like a POSIX character
3059 class but used an invalid character after [ (for example [[,abc,]]) caused
3060 pcre_compile() to give the error "Failed: internal error: code overflow" or
3061 in some cases to crash with a glibc free() error. This could even happen if
3062 the pattern terminated after [[ but there just happened to be a sequence of
3063 letters, a binary zero, and a closing ] in the memory that followed.
3064
3065 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3066 over the years. Originally (before any Unicode support), just the bottom 8
3067 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3068 output from "man perlunicode" includes this:
3069
3070 The regular expression compiler produces polymorphic opcodes. That
3071 is, the pattern adapts to the data and automatically switches to
3072 the Unicode character scheme when presented with Unicode data--or
3073 instead uses a traditional byte scheme when presented with byte
3074 data.
3075
3076 Sadly, a wide octal escape does not cause a switch, and in a string with
3077 no other multibyte characters, these octal escapes are treated as before.
3078 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3079 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3080 Unicode string.
3081
3082 I have not perpetrated such confusion in PCRE. Up till now, it took just
3083 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3084 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3085 translate to the appropriate multibyte character.
3086
3087 29. Applied some refactoring to reduce the number of warnings from Microsoft
3088 and Borland compilers. This has included removing the fudge introduced
3089 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3090 a warning about an unused variable.
3091
3092 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3093 characters since release 4.0, because Perl (from release 5.004) does not.
3094 [Or at least, is documented not to: some releases seem to be in conflict
3095 with the documentation.] However, when a pattern was studied with
3096 pcre_study() and all its branches started with \s, PCRE still included VT
3097 as a possible starting character. Of course, this did no harm; it just
3098 caused an unnecessary match attempt.
3099
3100 22. Removed a now-redundant internal flag bit that recorded the fact that case
3101 dependency changed within the pattern. This was once needed for "required
3102 byte" processing, but is no longer used. This recovers a now-scarce options
3103 bit. Also moved the least significant internal flag bit to the most-
3104 significant bit of the word, which was not previously used (hangover from
3105 the days when it was an int rather than a uint) to free up another bit for
3106 the future.
3107
3108 23. Added support for CRLF line endings as well as CR and LF. As well as the
3109 default being selectable at build time, it can now be changed at runtime
3110 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3111 specify that it is scanning data with non-default line endings.
3112
3113 24. Changed the definition of CXXLINK to make it agree with the definition of
3114 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3115
3116 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3117 recursions. This makes a big different to stack usage for some patterns.
3118
3119 26. If a subpattern containing a named recursion or subroutine reference such
3120 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3121 the space required for the compiled pattern went wrong and gave too small a
3122 value. Depending on the environment, this could lead to "Failed: internal
3123 error: code overflow at offset 49" or "glibc detected double free or
3124 corruption" errors.
3125
3126 27. Applied patches from Google (a) to support the new newline modes and (b) to
3127 advance over multibyte UTF-8 characters in GlobalReplace.
3128
3129 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3130 difference for some implementation of PCRE in some Windows version.
3131
3132 29. Added some extra testing facilities to pcretest:
3133
3134 \q<number> in a data line sets the "match limit" value
3135 \Q<number> in a data line sets the "match recursion limt" value
3136 -S <number> sets the stack size, where <number> is in megabytes
3137
3138 The -S option isn't available for Windows.
3139
3140
3141 Version 6.6 06-Feb-06
3142 ---------------------
3143
3144 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3145 in pcreposix.h. I have copied the definition from pcre.h.
3146
3147 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3148 because pcre.h is no longer a built file.
3149
3150 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3151 not normally included in the compiled code.
3152
3153
3154 Version 6.5 01-Feb-06
3155 ---------------------
3156
3157 1. When using the partial match feature with pcre_dfa_exec(), it was not
3158 anchoring the second and subsequent partial matches at the new starting
3159 point. This could lead to incorrect results. For example, with the pattern
3160 /1234/, partially matching against "123" and then "a4" gave a match.
3161
3162 2. Changes to pcregrep:
3163
3164 (a) All non-match returns from pcre_exec() were being treated as failures
3165 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3166 error message is output. Some extra information is given for the
3167 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3168 probably the only errors that are likely to be caused by users (by
3169 specifying a regex that has nested indefinite repeats, for instance).
3170 If there are more than 20 of these errors, pcregrep is abandoned.
3171
3172 (b) A binary zero was treated as data while matching, but terminated the
3173 output line if it was written out. This has been fixed: binary zeroes
3174 are now no different to any other data bytes.
3175
3176 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3177 used to set a locale for matching. The --locale=xxxx long option has
3178 been added (no short equivalent) to specify a locale explicitly on the
3179 pcregrep command, overriding the environment variables.
3180
3181 (d) When -B was used with -n, some line numbers in the output were one less
3182 than they should have been.
3183
3184 (e) Added the -o (--only-matching) option.
3185
3186 (f) If -A or -C was used with -c (count only), some lines of context were
3187 accidentally printed for the final match.
3188
3189 (g) Added the -H (--with-filename) option.
3190
3191 (h) The combination of options -rh failed to suppress file names for files
3192 that were found from directory arguments.
3193
3194 (i) Added the -D (--devices) and -d (--directories) options.
3195
3196 (j) Added the -F (--fixed-strings) option.
3197
3198 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3199
3200 (l) Added the --colo(u)r option.
3201
3202 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3203 is not present by default.
3204
3205 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3206 items such as (?R) or (?1), when the recursion could match a number of
3207 alternatives. If it matched one of the alternatives, but subsequently,
3208 outside the recursion, there was a failure, the code tried to back up into
3209 the recursion. However, because of the way PCRE is implemented, this is not
3210 possible, and the result was an incorrect result from the match.
3211
3212 In order to prevent this happening, the specification of recursion has
3213 been changed so that all such subpatterns are automatically treated as
3214 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3215
3216 4. I had overlooked the fact that, in some locales, there are characters for
3217 which isalpha() is true but neither isupper() nor islower() are true. In
3218 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3219 and ordfeminine) are like this. This affected the treatment of \w and \W
3220 when they appeared in character classes, but not when they appeared outside
3221 a character class. The bit map for "word" characters is now created
3222 separately from the results of isalnum() instead of just taking it from the
3223 upper, lower, and digit maps. (Plus the underscore character, of course.)
3224
3225 5. The above bug also affected the handling of POSIX character classes such as
3226 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3227 permanent tables. Instead, the bit maps for such a class were previously
3228 created as the appropriate unions of the upper, lower, and digit bitmaps.
3229 Now they are created by subtraction from the [[:word:]] class, which has
3230 its own bitmap.
3231
3232 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3233 It is created by subtracting the vertical space characters (\x09, \x0a,
3234 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3235 subtraction was done in the overall bitmap for a character class, meaning
3236 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3237 be recognized. This bug has been fixed.
3238
3239 7. Patches from the folks at Google:
3240
3241 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3242 real life, but is still worth protecting against".
3243
3244 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3245 regular expressions".
3246
3247 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3248 have it.
3249
3250 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3251 "configure" and the latter not, in order to fix a problem somebody had
3252 with compiling the Arg class on HP-UX.
3253
3254 (e) Improve the error-handling of the C++ wrapper a little bit.
3255
3256 (f) New tests for checking recursion limiting.
3257
3258 8. The pcre_memmove() function, which is used only if the environment does not
3259 have a standard memmove() function (and is therefore rarely compiled),
3260 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3261 returning a result (though PCRE never actually uses the result).
3262
3263 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3264 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3265 returned instead of calling malloc() with an overflowing number that would
3266 most likely cause subsequent chaos.
3267
3268 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3269
3270 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3271 with this option is matched, the nmatch and pmatch options of regexec() are
3272 ignored.
3273
3274 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3275 provided in case anyone wants to the the POSIX interface with UTF-8
3276 strings.
3277
3278 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3279 C++ linking (needed for some HP-UX environments).
3280
3281 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3282 (unused parameter) and in the pcre_printint() function (omitted "default"
3283 switch label when the default is to do nothing).
3284
3285 15. Added some code to make it possible, when PCRE is compiled as a C++
3286 library, to replace subject pointers for pcre_exec() with a smart pointer
3287 class, thus making it possible to process discontinuous strings.
3288
3289 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3290 much the same function. They were added by different people who were trying
3291 to make PCRE easy to compile on non-Unix systems. It has been suggested
3292 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3293 for compiling on Windows systems. I have therefore replaced it with
3294 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3295 defaults to "extern" for C or "extern C" for C++, which works fine on
3296 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3297 SCOPE with something explicit in config.h. In addition:
3298
3299 (a) pcreposix.h still had just "extern" instead of either of these macros;
3300 I have replaced it with PCRE_DATA_SCOPE.
3301
3302 (b) Functions such as _pcre_xclass(), which are internal to the library,
3303 but external in the C sense, all had PCRE_EXPORT in their definitions.
3304 This is apparently wrong for the Windows case, so I have removed it.
3305 (It makes no difference on Unix-like systems.)
3306
3307 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3308 of recursive calls to match(). This is different to MATCH_LIMIT because
3309 that limits the total number of calls to match(), not all of which increase
3310 the depth of recursion. Limiting the recursion depth limits the amount of
3311 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3312 when PCRE is compiled, and changed at run time. A patch from Google adds
3313 this functionality to the C++ interface.
3314
3315 18. Changes to the handling of Unicode character properties:
3316
3317 (a) Updated the table to Unicode 4.1.0.
3318
3319 (b) Recognize characters that are not in the table as "Cn" (undefined).
3320
3321 (c) I revised the way the table is implemented to a much improved format
3322 which includes recognition of ranges. It now supports the ranges that
3323 are defined in UnicodeData.txt, and it also amalgamates other
3324 characters into ranges. This has reduced the number of entries in the
3325 table from around 16,000 to around 3,000, thus reducing its size
3326 considerably. I realized I did not need to use a tree structure after
3327 all - a binary chop search is just as efficient. Having reduced the
3328 number of entries, I extended their size from 6 bytes to 8 bytes to
3329 allow for more data.
3330
3331 (d) Added support for Unicode script names via properties such as \p{Han}.
3332
3333 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3334 matching that character.
3335
3336 20. When matching a repeated Unicode property with a minimum greater than zero,
3337 (for example \pL{2,}), PCRE could look past the end of the subject if it
3338 reached it while seeking the minimum number of characters. This could
3339 happen only if some of the characters were more than one byte long, because
3340 there is a check for at least the minimum number of bytes.
3341
3342 21. Refactored the implementation of \p and \P so as to be more general, to
3343 allow for more different types of property in future. This has changed the
3344 compiled form incompatibly. Anybody with saved compiled patterns that use
3345 \p or \P will have to recompile them.
3346
3347 22. Added "Any" and "L&" to the supported property types.
3348
3349 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3350 but give a compile time error if the value is greater than 0xff.
3351
3352 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3353 accidentally not being installed or uninstalled.
3354
3355 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3356 made were to insert the current release number. This seemed silly, because
3357 it made things harder for people building PCRE on systems that don't run
3358 "configure". I have turned pcre.h into a distributed file, no longer built
3359 by "configure", with the version identification directly included. There is
3360 no longer a pcre.h.in file.
3361
3362 However, this change necessitated a change to the pcre-config script as
3363 well. It is built from pcre-config.in, and one of the substitutions was the
3364 release number. I have updated configure.ac so that ./configure now finds
3365 the release number by grepping pcre.h.
3366
3367 26. Added the ability to run the tests under valgrind.
3368
3369
3370 Version 6.4 05-Sep-05
3371 ---------------------
3372
3373 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3374 "--" to be printed when multiple files were scanned, even when none of the
3375 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3376 consider it to be a bug, and have restored the previous behaviour.
3377
3378 2. A couple of code tidies to get rid of compiler warnings.
3379
3380 3. The pcretest program used to cheat by referring to symbols in the library
3381 whose names begin with _pcre_. These are internal symbols that are not
3382 really supposed to be visible externally, and in some environments it is
3383 possible to suppress them. The cheating is now confined to including
3384 certain files from the library's source, which is a bit cleaner.
3385
3386 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3387 file's purpose clearer.
3388
3389 5. Reorganized pcre_ucp_findchar().
3390
3391
3392 Version 6.3 15-Aug-05
3393 ---------------------
3394
3395 1. The file libpcre.pc.in did not have general read permission in the tarball.
3396
3397 2. There were some problems when building without C++ support:
3398
3399 (a) If C++ support was not built, "make install" and "make test" still
3400 tried to test it.
3401
3402 (b) There were problems when the value of CXX was explicitly set. Some
3403 changes have been made to try to fix these, and ...
3404
3405 (c) --disable-cpp can now be used to explicitly disable C++ support.
3406
3407 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3408 backslash in a target when C++ was disabled. This confuses some
3409 versions of "make", apparently. Using an intermediate variable solves
3410 this. (Same for CPP_LOBJ.)
3411
3412 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3413 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3414 necessary on certain architectures.
3415
3416 4. Added a setting of -export-symbols-regex to the link command to remove
3417 those symbols that are exported in the C sense, but actually are local
3418 within the library, and not documented. Their names all begin with
3419 "_pcre_". This is not a perfect job, because (a) we have to except some
3420 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3421 available (and never for static libraries). I have made a note to try to
3422 find a way round (a) in the future.
3423
3424
3425 Version 6.2 01-Aug-05
3426 ---------------------
3427
3428 1. There was no test for integer overflow of quantifier values. A construction
3429 such as {1111111111111111} would give undefined results. What is worse, if
3430 a minimum quantifier for a parenthesized subpattern overflowed and became
3431 negative, the calculation of the memory size went wrong. This could have
3432 led to memory overwriting.
3433
3434 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3435
3436 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3437 operating environments where this matters.
3438
3439 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3440 PCRE options from within the C++ wrapper.
3441
3442 5. Named capturing subpatterns were not being correctly counted when a pattern
3443 was compiled. This caused two problems: (a) If there were more than 100
3444 such subpatterns, the calculation of the memory needed for the whole
3445 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3446 back references of the form \12, where the number was greater than 9, were
3447 not recognized as back references, even though there were sufficient
3448 previous subpatterns.
3449
3450 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3451 versions of gcc, e.g. 2.95.4.
3452
3453
3454 Version 6.1 21-Jun-05
3455 ---------------------
3456
3457 1. There was one reference to the variable "posix" in pcretest.c that was not
3458 surrounded by "#if !defined NOPOSIX".
3459
3460 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3461 the cross-check on the old pcre_info() function, for the benefit of the
3462 cut-down version of PCRE that is currently imported into Exim.
3463
3464 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3465 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3466 patterns that start (?i) but I don't think that matters. The use of (?i) is
3467 just an example; this all applies to the other options as well.
3468
3469 4. Since libtool seems to echo the compile commands it is issuing, the output
3470 from "make" can be reduced a bit by putting "@" in front of each libtool
3471 compile command.
3472
3473 5. Patch from the folks at Google for configure.in to be a bit more thorough
3474 in checking for a suitable C++ installation before trying to compile the
3475 C++ stuff. This should fix a reported problem when a compiler was present,
3476 but no suitable headers.
3477
3478 6. The man pages all had just "PCRE" as their title. I have changed them to
3479 be the relevant file name. I have also arranged that these names are
3480 retained in the file doc/pcre.txt, which is a concatenation in text format
3481 of all the man pages except the little individual ones for each function.
3482
3483 7. The NON-UNIX-USE file had not been updated for the different set of source
3484 files that come with release 6. I also added a few comments about the C++
3485 wrapper.
3486
3487
3488 Version 6.0 07-Jun-05
3489 ---------------------
3490
3491 1. Some minor internal re-organization to help with my DFA experiments.
3492
3493 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3494 didn't matter for the library itself when fully configured, but did matter
3495 when compiling without UCP support, or within Exim, where the ucp files are
3496 not imported.
3497
3498 3. Refactoring of the library code to split up the various functions into
3499 different source modules. The addition of the new DFA matching code (see
3500 below) to a single monolithic source would have made it really too
3501 unwieldy, quite apart from causing all the code to be include in a
3502 statically linked application, when only some functions are used. This is
3503 relevant even without the DFA addition now that patterns can be compiled in
3504 one application and matched in another.
3505
3506 The downside of splitting up is that there have to be some external
3507 functions and data tables that are used internally in different modules of
3508 the library but which are not part of the API. These have all had their
3509 names changed to start with "_pcre_" so that they are unlikely to clash
3510 with other external names.
3511
3512 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3513 a different (DFA) algorithm. Although it is slower than the original
3514 function, it does have some advantages for certain types of matching
3515 problem.
3516
3517 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3518 including restarting after a partial match.
3519
3520 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3521 defined when compiling for Windows was sent to me. I have put it into the
3522 code, though I have no means of testing or verifying it.
3523
3524 7. Added the pcre_refcount() auxiliary function.
3525
3526 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3527 match before or at the first newline in the subject string. In pcretest,
3528 the /f option on a pattern can be used to set this.
3529
3530 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3531 would behave wrongly. This has been present in PCRE since release 4.0.
3532
3533 10. A number of changes to the pcregrep command:
3534
3535 (a) Refactored how -x works; insert ^(...)$ instead of setting
3536 PCRE_ANCHORED and checking the length, in preparation for adding
3537 something similar for -w.
3538
3539 (b) Added the -w (match as a word) option.
3540
3541 (c) Refactored the way lines are read and buffered so as to have more
3542 than one at a time available.
3543
3544 (d) Implemented a pcregrep test script.
3545
3546 (e) Added the -M (multiline match) option. This allows patterns to match
3547 over several lines of the subject. The buffering ensures that at least
3548 8K, or the rest of the document (whichever is the shorter) is available
3549 for matching (and similarly the previous 8K for lookbehind assertions).
3550
3551 (f) Changed the --help output so that it now says
3552
3553 -w, --word-regex(p)
3554
3555 instead of two lines, one with "regex" and the other with "regexp"
3556 because that confused at least one person since the short forms are the
3557 same. (This required a bit of code, as the output is generated
3558 automatically from a table. It wasn't just a text change.)
3559
3560 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3561 option but starts with a hyphen. Could be a pattern or a path name
3562 starting with a hyphen, for instance.
3563
3564 (h) "-" can be given as a file name to represent stdin.
3565
3566 (i) When file names are being printed, "(standard input)" is used for
3567 the standard input, for compatibility with GNU grep. Previously
3568 "<stdin>" was used.
3569
3570 (j) The option --label=xxx can be used to supply a name to be used for
3571 stdin when file names are being printed. There is no short form.
3572
3573 (k) Re-factored the options decoding logic because we are going to add
3574 two more options that take data. Such options can now be given in four
3575 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3576
3577 (l) Added the -A, -B, and -C options for requesting that lines of context
3578 around matches be printed.
3579
3580 (m) Added the -L option to print the names of files that do not contain
3581 any matching lines, that is, the complement of -l.
3582
3583 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3584 continue to scan other files.
3585
3586 (o) The -s option was incorrectly implemented. For compatibility with other
3587 greps, it now suppresses the error message for a non-existent or non-
3588 accessible file (but not the return code). There is a new option called
3589 -q that suppresses the output of matching lines, which was what -s was
3590 previously doing.
3591
3592 (p) Added --include and --exclude options to specify files for inclusion
3593 and exclusion when recursing.
3594
3595 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3596 Hopefully, it now does.
3597
3598 12. Missing cast in pcre_study().
3599
3600 13. Added an "uninstall" target to the makefile.
3601
3602 14. Replaced "extern" in the function prototypes in Makefile.in with
3603 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3604 world, but is set differently for Windows.
3605
3606 15. Added a second compiling function called pcre_compile2(). The only
3607 difference is that it has an extra argument, which is a pointer to an
3608 integer error code. When there is a compile-time failure, this is set
3609 non-zero, in addition to the error test pointer being set to point to an
3610 error message. The new argument may be NULL if no error number is required
3611 (but then you may as well call pcre_compile(), which is now just a
3612 wrapper). This facility is provided because some applications need a
3613 numeric error indication, but it has also enabled me to tidy up the way
3614 compile-time errors are handled in the POSIX wrapper.
3615
3616 16. Added VPATH=.libs to the makefile; this should help when building with one
3617 prefix path and installing with another. (Or so I'm told by someone who
3618 knows more about this stuff than I do.)
3619
3620 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3621 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3622 match everything, including newlines. This is not POSIX-compatible, but
3623 somebody wanted the feature. From pcretest it can be activated by using
3624 both the P and the s flags.
3625
3626 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3627
3628 19. libpcre.pc was being incorrectly installed as executable.
3629
3630 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3631 it now also looks for '\r' so that it will work unmodified on Windows.
3632
3633 21. Added Google's contributed C++ wrapper to the distribution.
3634
3635 22. Added some untidy missing memory free() calls in pcretest, to keep
3636 Electric Fence happy when testing.
3637
3638
3639
3640 Version 5.0 13-Sep-04
3641 ---------------------
3642
3643 1. Internal change: literal characters are no longer packed up into items
3644 containing multiple characters in a single byte-string. Each character
3645 is now matched using a separate opcode. However, there may be more than one
3646 byte in the character in UTF-8 mode.
3647
3648 2. The pcre_callout_block structure has two new fields: pattern_position and
3649 next_item_length. These contain the offset in the pattern to the next match
3650 item, and its length, respectively.
3651
3652 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3653 insertion of callouts before each pattern item. Added the /C option to
3654 pcretest to make use of this.
3655
3656 4. On the advice of a Windows user, the lines
3657
3658 #if defined(_WIN32) || defined(WIN32)
3659 _setmode( _fileno( stdout ), 0x8000 );
3660 #endif /* defined(_WIN32) || defined(WIN32) */
3661
3662 have been added to the source of pcretest. This apparently does useful
3663 magic in relation to line terminators.
3664
3665 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3666 for the benefit of those environments where the "b" makes a difference.
3667
3668 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3669 to know about it. I have put a hack into configure.in that adds in code
3670 to set GCC=yes if CC=icc. This seems to end up at a point in the
3671 generated configure script that is early enough to affect the setting of
3672 compiler options, which is what is needed, but I have no means of testing
3673 whether it really works. (The user who reported this had patched the
3674 generated configure script, which of course I cannot do.)
3675
3676 LATER: After change 22 below (new libtool files), the configure script
3677 seems to know about icc (and also ecc). Therefore, I have commented out
3678 this hack in configure.in.
3679
3680 7. Added support for pkg-config (2 patches were sent in).
3681
3682 8. Negated POSIX character classes that used a combination of internal tables
3683 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3684 [[:^ascii]]. Typically, they would match almost any characters. The other
3685 POSIX classes were not broken in this way.
3686
3687 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3688 to find the match, as PCRE was deluded into thinking that the match had to
3689 start at the start point or following a newline. The same bug applied to
3690 patterns with negative forward assertions or any backward assertions
3691 preceding ".*" at the start, unless the pattern required a fixed first
3692 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3693
3694 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3695 starting at the last subject character, bytes beyond the end of the subject
3696 string were read.
3697
3698 11. Renamed the variable "class" as "classbits" to make life easier for C++
3699 users. (Previously there was a macro definition, but it apparently wasn't
3700 enough.)
3701
3702 12. Added the new field "tables" to the extra data so that tables can be passed
3703 in at exec time, or the internal tables can be re-selected. This allows
3704 a compiled regex to be saved and re-used at a later time by a different
3705 program that might have everything at different addresses.
3706
3707 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3708 -R library as well as a -L library.
3709
3710 14. The debugging options of pcretest (-d on the command line or D on a
3711 pattern) showed incorrect output for anything following an extended class
3712 that contained multibyte characters and which was followed by a quantifier.
3713
3714 15. Added optional support for general category Unicode character properties
3715 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3716 support. It adds about 90K to the size of the library. The meanings of the
3717 inbuilt class escapes such as \d and \s have NOT been changed.
3718
3719 16. Updated pcredemo.c to include calls to free() to release the memory for the
3720 compiled pattern.
3721
3722 17. The generated file chartables.c was being created in the source directory
3723 instead of in the building directory. This caused the build to fail if the
3724 source directory was different from the building directory, and was
3725 read-only.
3726
3727 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3728 file. No doubt somebody will tell me if they don't make sense... Also added
3729 Dan Mooney's comments about building on OpenVMS.
3730
3731 19. Added support for partial matching via the PCRE_PARTIAL option for
3732 pcre_exec() and the \P data escape in pcretest.
3733
3734 20. Extended pcretest with 3 new pattern features:
3735
3736 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3737 write the compiled pattern to the file whose name is "rest-of-line".
3738 This is a straight binary dump of the data, with the saved pointer to
3739 the character tables forced to be NULL. The study data, if any, is
3740 written too. After writing, pcretest reads a new pattern.
3741
3742 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3743 compiled pattern from the given file. There must not be any
3744 occurrences of "<" in the file name (pretty unlikely); if there are,
3745 pcretest will instead treat the initial "<" as a pattern delimiter.
3746 After reading in the pattern, pcretest goes on to read data lines as
3747 usual.
3748
3749 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3750 and 16-bit fields in a compiled pattern, to simulate a pattern that
3751 was compiled on a host of opposite endianness.
3752
3753 21. The pcre-exec() function can now cope with patterns that were compiled on
3754 hosts of opposite endianness, with this restriction:
3755
3756 As for any compiled expression that is saved and used later, the tables
3757 pointer field cannot be preserved; the extra_data field in the arguments
3758 to pcre_exec() should be used to pass in a tables address if a value
3759 other than the default internal tables were used at compile time.
3760
3761 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3762 now diagnosed as an error. Previously, most of the time, a negative number
3763 would have been treated as zero, but if in addition "ovector" was passed as
3764 NULL, a crash could occur.
3765
3766 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3767 new versions from the libtool 1.5 distribution (the last one is a copy of
3768 a file called libtool.m4). This seems to have fixed the need to patch
3769 "configure" to support Darwin 1.3 (which I used to do). However, I still
3770 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3771 workstation).
3772
3773 24. Changed the PCRE licence to be the more standard "BSD" licence.
3774
3775
3776 Version 4.5 01-Dec-03
3777 ---------------------
3778
3779 1. There has been some re-arrangement of the code for the match() function so
3780 that it can be compiled in a version that does not call itself recursively.
3781 Instead, it keeps those local variables that need separate instances for
3782 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3783 needs to "recurse". Keeping track of where control must go is done by means
3784 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3785 hide most of the details from the main code, and operates only if
3786 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3787 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3788 operating.
3789
3790 To make it easier for callers to provide specially tailored get/free
3791 functions for this usage, two new functions, pcre_stack_malloc, and
3792 pcre_stack_free, are used. They are always called in strict stacking order,
3793 and the size of block requested is always the same.
3794
3795 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3796 PCRE has been compiled to use the stack or the heap for recursion. The
3797 -C option of pcretest uses this to show which version is compiled.
3798
3799 A new data escape \S, is added to pcretest; it causes the amounts of store
3800 obtained and freed by both kinds of malloc/free at match time to be added
3801 to the output.
3802
3803 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3804 what's available on my current Linux desktop machine.
3805
3806 3. When matching a UTF-8 string, the test for a valid string at the start has
3807 been extended. If start_offset is not zero, PCRE now checks that it points
3808 to a byte that is the start of a UTF-8 character. If not, it returns
3809 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3810 this is necessary because there may be backward assertions in the pattern.
3811 When matching the same subject several times, it may save resources to use
3812 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3813
3814 4. The code for checking the validity of UTF-8 strings has been tightened so
3815 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3816 containing "overlong sequences".
3817
3818 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3819 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3820 should have been "&", but it just so happened that all the cases this let
3821 through by mistake were picked up later in the function.
3822
3823 6. I had used a variable called "isblank" - this is a C99 function, causing
3824 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3825
3826 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3827 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3828 so the version is shown; (c) stop "make test" echoing "./RunTest".
3829
3830 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3831
3832 9. The prototype for memmove() for systems that don't have it was using
3833 size_t, but the inclusion of the header that defines size_t was later. I've
3834 moved the #includes for the C headers earlier to avoid this.
3835
3836 10. Added some adjustments to the code to make it easier to compiler on certain
3837 special systems:
3838
3839 (a) Some "const" qualifiers were missing.
3840 (b) Added the macro EXPORT before all exported functions; by default this
3841 is defined to be empty.
3842 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3843 that it reads its output file name as an argument instead of writing
3844 to the standard output and assuming this can be redirected.
3845
3846 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3847 class containing characters with values greater than 255, PCRE compilation
3848 went into a loop.
3849
3850 12. A recursive reference to a subpattern that was within another subpattern
3851 that had a minimum quantifier of zero caused PCRE to crash. For example,
3852 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3853 recursion. If the recursively-called subpattern itself had a zero repeat,
3854 that was OK.
3855
3856 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3857 buffer into which it was copied (for escape processing) was still set at
3858 1024, so long lines caused crashes.
3859
3860 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3861 "internal error: code overflow...". This applied to any character class
3862 that was followed by a possessive quantifier.
3863
3864 15. Modified the Makefile to add libpcre.la as a prerequisite for
3865 libpcreposix.la because I was told this is needed for a parallel build to
3866 work.
3867
3868 16. If a pattern that contained .* following optional items at the start was
3869 studied, the wrong optimizing data was generated, leading to matching
3870 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3871 matching string must start with a or b or c. The correct conclusion for
3872 this pattern is that a match can start with any character.
3873
3874
3875 Version 4.4 13-Aug-03
3876 ---------------------
3877
3878 1. In UTF-8 mode, a character class containing characters with values between
3879 127 and 255 was not handled correctly if the compiled pattern was studied.
3880 In fixing this, I have also improved the studying algorithm for such
3881 classes (slightly).
3882
3883 2. Three internal functions had redundant arguments passed to them. Removal
3884 might give a very teeny performance improvement.
3885
3886 3. Documentation bug: the value of the capture_top field in a callout is *one
3887 more than* the number of the hightest numbered captured substring.
3888
3889 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3890 in incorrectly linking with a previously installed version. They now link
3891 explicitly with libpcre.la.
3892
3893 5. configure.in no longer needs to recognize Cygwin specially.
3894
3895 6. A problem in pcre.in for Windows platforms is fixed.
3896
3897 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3898 pcretest, it used to include the size of the study block as part of its
3899 output. Unfortunately, the structure contains a field that has a different
3900 size on different hardware architectures. This meant that the tests that
3901 showed this size failed. As the block is currently always of a fixed size,
3902 this information isn't actually particularly useful in pcretest output, so
3903 I have just removed it.
3904
3905 8. Three pre-processor statements accidentally did not start in column 1.
3906 Sadly, there are *still* compilers around that complain, even though
3907 standard C has not required this for well over a decade. Sigh.
3908
3909 9. In pcretest, the code for checking callouts passed small integers in the
3910 callout_data field, which is a void * field. However, some picky compilers
3911 complained about the casts involved for this on 64-bit systems. Now
3912 pcretest passes the address of the small integer instead, which should get
3913 rid of the warnings.
3914
3915 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3916 both compile and run time, and gives an error if an invalid UTF-8 sequence
3917 is found. There is a option for disabling this check in cases where the
3918 string is known to be correct and/or the maximum performance is wanted.
3919
3920 11. In response to a bug report, I changed one line in Makefile.in from
3921
3922 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3923 to
3924 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3925
3926 to look similar to other lines, but I have no way of telling whether this
3927 is the right thing to do, as I do not use Windows. No doubt I'll get told
3928 if it's wrong...
3929
3930
3931 Version 4.3 21-May-03
3932 ---------------------
3933
3934 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3935 Makefile.
3936
3937 2. Some refactoring to improve the quality of the code:
3938
3939 (i) The utf8_table... variables are now declared "const".
3940
3941 (ii) The code for \cx, which used the "case flipping" table to upper case
3942 lower case letters, now just substracts 32. This is ASCII-specific,
3943 but the whole concept of \cx is ASCII-specific, so it seems
3944 reasonable.
3945
3946 (iii) PCRE was using its character types table to recognize decimal and
3947 hexadecimal digits in the pattern. This is silly, because it handles
3948 only 0-9, a-f, and A-F, but the character types table is locale-
3949 specific, which means strange things might happen. A private
3950 table is now used for this - though it costs 256 bytes, a table is
3951 much faster than multiple explicit tests. Of course, the standard
3952 character types table is still used for matching digits in subject
3953 strings against \d.
3954
3955 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3956 ending in _t are). So I've renamed it as ESC_tee.
3957
3958 3. The first argument for regexec() in the POSIX wrapper should have been
3959 defined as "const".
3960
3961 4. Changed pcretest to use malloc() for its buffers so that they can be
3962 Electric Fenced for debugging.
3963
3964 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3965 to read one or more bytes before the start of the subject string. Often this
3966 had no effect on PCRE's behaviour, but in some circumstances it could
3967 provoke a segmentation fault.
3968
3969 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3970 to try to read one or more bytes before the start of the subject string.
3971
3972 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3973 UTF-8 support could misbehave in various ways if the subject string
3974 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3975 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3976 back over UTF-8 characters.)
3977
3978
3979 Version 4.2 14-Apr-03
3980 ---------------------
3981
3982 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3983
3984 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3985 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3986 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3987 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3988 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3989 and BUILD_EXEEXT
3990 Note: automatic setting of the BUILD variables is not yet working
3991 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3992 compile-time but not at link-time
3993 [LINK]: use for linking executables only
3994 make different versions for Windows and non-Windows
3995 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3996 libraries
3997 [LINK_FOR_BUILD]: new variable
3998 [OBJEXT]: use throughout
3999 [EXEEXT]: use throughout
4000 <winshared>: new target
4001 <wininstall>: new target
4002 <dftables.o>: use native compiler
4003 <dftables>: use native linker
4004 <install>: handle Windows platform correctly
4005 <clean>: ditto
4006 <check>: ditto
4007 copy DLL to top builddir before testing
4008
4009 As part of these changes, -no-undefined was removed again. This was reported
4010 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4011 in any case.
4012
4013 3. Some tidies to get rid of compiler warnings:
4014
4015 . In the match_data structure, match_limit was an unsigned long int, whereas
4016 match_call_count was an int. I've made them both unsigned long ints.
4017
4018 . In pcretest the fact that a const uschar * doesn't automatically cast to
4019 a void * provoked a warning.
4020
4021 . Turning on some more compiler warnings threw up some "shadow" variables
4022 and a few more missing casts.
4023
4024 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4025 option, a class that contained a single character with a value between 128
4026 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4027
4028 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4029 option, a class that contained several characters, but with at least one
4030 whose value was between 128 and 255 caused PCRE to crash.
4031
4032
4033 Version 4.1 12-Mar-03
4034 ---------------------
4035
4036 1. Compiling with gcc -pedantic found a couple of places where casts were
4037 needed, and a string in dftables.c that was longer than standard compilers are
4038 required to support.
4039
4040 2. Compiling with Sun's compiler found a few more places where the code could
4041 be tidied up in order to avoid warnings.
4042
4043 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4044 first of these names is deprecated in the latest Autoconf in favour of the name
4045 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4046 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4047 analogy I have changed it to CFLAGS_FOR_BUILD.
4048
4049 4. Added -no-undefined to the linking command in the Makefile, because this is
4050 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4051 linking step for the pcreposix library.
4052
4053 5. PCRE was failing to diagnose the case of two named groups with the same
4054 name.
4055
4056 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4057 literal character that is needed in the subject for a match, and scans along to
4058 ensure that it is present before embarking on the full matching process. This
4059 saves time in cases of nested unlimited repeats that are never going to match.
4060 Problem: the scan can take a lot of time if the subject is very long (e.g.
4061 megabytes), thus penalizing straightforward matches. It is now done only if the
4062 amount of subject to be scanned is less than 1000 bytes.
4063
4064 7. A lesser problem with the same optimization is that it was recording the
4065 first character of an anchored pattern as "needed", thus provoking a search
4066 right along the subject, even when the first match of the pattern was going to
4067 fail. The "needed" character is now not set for anchored patterns, unless it
4068 follows something in the pattern that is of non-fixed length. Thus, it still
4069 fulfils its original purpose of finding quick non-matches in cases of nested
4070 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4071
4072
4073 Version 4.0 17-Feb-03
4074 ---------------------
4075
4076 1. If a comment in an extended regex that started immediately after a meta-item
4077 extended to the end of string, PCRE compiled incorrect data. This could lead to
4078 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4079
4080 2. Moved to autoconf 2.53 and libtool 1.4.2.
4081
4082 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4083 the special perltest8 script is no longer needed - all the tests can be run
4084 from a single perltest script.
4085
4086 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4087 by \s. It has now been removed in PCRE. This means it isn't recognized as
4088 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4089 class [:space:] *does* include VT, thereby creating a mess.
4090
4091 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4092 space and tab.
4093
4094 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4095 its new features into the main test script, reducing the number of scripts.
4096
4097 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4098 were backward compatible, and made the (?i) apply to the whole pattern, as if
4099 /i were given. Now it behaves more logically, and applies the option setting
4100 only to what follows. PCRE has been changed to follow suit. However, if it
4101 finds options settings right at the start of the pattern, it extracts them into
4102 the global options, as before. Thus, they show up in the info data.
4103
4104 8. Added support for the \Q...\E escape sequence. Characters in between are
4105 treated as literals. This is slightly different from Perl in that $ and @ are
4106 also handled as literals inside the quotes. In Perl, they will cause variable
4107 interpolation. Note the following examples:
4108
4109 Pattern PCRE matches Perl matches
4110
4111 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4112 \Qabc\$xyz\E abc\$xyz abc\$xyz
4113 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4114
4115 For compatibility with Perl, \Q...\E sequences are recognized inside character
4116 classes as well as outside them.
4117
4118 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4119 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4120 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4121 signed/unsigned warnings.
4122
4123 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4124 option for pcretest, so I've replaced it by a simple function that does just
4125 that job.
4126
4127 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4128 "pcregrep -".
4129
4130 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4131 Java package. This provides some syntactic sugar for simple cases of what my
4132 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4133 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4134 item, you can use this simplified notation. Note that only makes sense with
4135 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4136 greediness, whatever the setting of the PCRE_UNGREEDY option.
4137
4138 13. A change of greediness default within a pattern was not taking effect at
4139 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4140 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4141 was abstracted outside.
4142
4143 14. PCRE now supports the \G assertion. It is true when the current matching
4144 position is at the start point of the match. This differs from \A when the
4145 starting offset is non-zero. Used with the /g option of pcretest (or similar
4146 code), it works in the same way as it does for Perl's /g option. If all
4147 alternatives of a regex begin with \G, the expression is anchored to the start
4148 match position, and the "anchored" flag is set in the compiled expression.
4149
4150 15. Some bugs concerning the handling of certain option changes within patterns
4151 have been fixed. These applied to options other than (?ims). For example,
4152 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4153 been the other way round. Some of this was related to change 7 above.
4154
4155 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4156 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4157 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4158 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4159
4160 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4161 mode. Unlike ".", it always matches newline, whatever the setting of
4162 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4163 assertions. Perl allows it, but it doesn't (in general) work because it can't
4164 calculate the length of the lookbehind. At least, that's the case for Perl
4165 5.8.0 - I've been told they are going to document that it doesn't work in
4166 future.
4167
4168 18. Added an error diagnosis for escapes that PCRE does not support: these are
4169 \L, \l, \N, \P, \p, \U, \u, and \X.
4170
4171 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4172 reading past the end of the pattern in cases such as /[abcd/.
4173
4174 20. PCRE was getting more memory than necessary for patterns with classes that
4175 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4176
4177 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4178 compiling PCRE for use with Virtual Pascal.
4179
4180 22. Small fix to the Makefile to make it work properly if the build is done
4181 outside the source tree.
4182
4183 23. Added a new extension: a condition to go with recursion. If a conditional
4184 subpattern starts with (?(R) the "true" branch is used if recursion has
4185 happened, whereas the "false" branch is used only at the top level.
4186
4187 24. When there was a very long string of literal characters (over 255 bytes
4188 without UTF support, over 250 bytes with UTF support), the computation of how
4189 much memory was required could be incorrect, leading to segfaults or other
4190 strange effects.
4191
4192 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4193 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4194 there was a subsequent back reference to those brackets. This meant that, for
4195 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4196 possible to check for precisely this case. All we can do is abandon the
4197 optimization if .* occurs inside capturing brackets when there are any back
4198 references whatsoever. (See below for a better fix that came later.)
4199
4200 26. The handling of the optimization for finding the first character of a
4201 non-anchored pattern, and for finding a character that is required later in the
4202 match were failing in some cases. This didn't break the matching; it just
4203 failed to optimize when it could. The way this is done has been re-implemented.
4204
4205 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4206
4207 28. Added a new feature that provides some of the functionality that Perl
4208 provides with (?{...}). The facility is termed a "callout". The way it is done
4209 in PCRE is for the caller to provide an optional function, by setting
4210 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4211 global variable. By default it is unset, which disables all calling out. To get
4212 the function called, the regex must include (?C) at appropriate points. This
4213 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4214 This provides a means of identifying different callout points. When PCRE
4215 reaches such a point in the regex, if pcre_callout has been set, the external
4216 function is called. It is provided with data in a structure called
4217 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4218 matching continues; if it returns a non-zero value, the match at the current
4219 point fails. However, backtracking will occur if possible. [This was changed
4220 later and other features added - see item 49 below.]
4221
4222 29. pcretest is upgraded to test the callout functionality. It provides a
4223 callout function that displays information. By default, it shows the start of
4224 the match and the current position in the text. There are some new data escapes
4225 to vary what happens:
4226
4227 \C+ in addition, show current contents of captured substrings
4228 \C- do not supply a callout function
4229 \C!n return 1 when callout number n is reached
4230 \C!n!m return 1 when callout number n is reached for the mth time
4231
4232 30. If pcregrep was called with the -l option and just a single file name, it
4233 output "<stdin>" if a match was found, instead of the file name.
4234
4235 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4236 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4237 pcre_exec(). This saves a malloc/free per call. The default value of
4238 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4239 when configuring.
4240
4241 32. The default maximum size of a compiled pattern is 64K. There have been a
4242 few cases of people hitting this limit. The code now uses macros to handle the
4243 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4244 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4245 configuring. Tests 2 and 5 work only with 2-byte links because they output
4246 debugging information about compiled patterns.
4247
4248 33. Internal code re-arrangements:
4249
4250 (a) Moved the debugging function for printing out a compiled regex into
4251 its own source file (printint.c) and used #include to pull it into
4252 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4253 separate copies.
4254
4255 (b) Defined the list of op-code names for debugging as a macro in
4256 internal.h so that it is next to the definition of the opcodes.
4257
4258 (c) Defined a table of op-code lengths for simpler skipping along compiled
4259 code. This is again a macro in internal.h so that it is next to the
4260 definition of the opcodes.
4261
4262 34. Added support for recursive calls to individual subpatterns, along the
4263 lines of Robin Houston's patch (but implemented somewhat differently).
4264
4265 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4266 allow it to read and process whole directories in Win32. This code was
4267 contributed by Lionel Fourquaux; it has not been tested by me.
4268
4269 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4270 used to name a group. Names consist of alphanumerics and underscores, and must
4271 be unique. Back references use the syntax (?P=name) and recursive calls use
4272 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4273 numbers. The function pcre_fullinfo() can be used after compilation to extract
4274 a name/number map. There are three relevant calls:
4275
4276 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4277 PCRE_INFO_NAMECOUNT yields the number of entries
4278 PCRE_INFO_NAMETABLE yields a pointer to the map.
4279
4280 The map is a vector of fixed-size entries. The size of each entry depends on
4281 the length of the longest name used. The first two bytes of each entry are the
4282 group number, most significant byte first. There follows the corresponding
4283 name, zero terminated. The names are in alphabetical order.
4284
4285 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4286 case instead of 255. Making it the same both with and without UTF-8 support
4287 means that the same test output works with both.
4288
4289 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4290 calling malloc() with a zero argument.
4291
4292 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4293 optimization. I've improved things by keeping a bitmap of backreferences with
4294 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4295 fact referenced, the optimization can be applied. It is unlikely that a
4296 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4297 the match to follow \n) will appear inside brackets with a number greater than
4298 31, but if it does, any back reference > 31 suppresses the optimization.
4299
4300 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4301 of disabling numbered capturing parentheses. Any opening parenthesis that is
4302 not followed by ? behaves as if it were followed by ?: but named parentheses
4303 can still be used for capturing (and they will acquire numbers in the usual
4304 way).
4305
4306 41. Redesigned the return codes from the match() function into yes/no/error so
4307 that errors can be passed back from deep inside the nested calls. A malloc
4308 failure while inside a recursive subpattern call now causes the
4309 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4310
4311 42. It is now possible to set a limit on the number of times the match()
4312 function is called in a call to pcre_exec(). This facility makes it possible to
4313 limit the amount of recursion and backtracking, though not in a directly
4314 obvious way, because the match() function is used in a number of different
4315 circumstances. The count starts from zero for each position in the subject
4316 string (for non-anchored patterns). The default limit is, for compatibility, a
4317 large number, namely 10 000 000. You can change this in two ways:
4318
4319 (a) When configuring PCRE before making, you can use --with-match-limit=n
4320 to set a default value for the compiled library.
4321
4322 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4323 a different value is set. See 45 below.
4324
4325 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4326
4327 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4328 of things that can be changed at compile time. The first argument specifies
4329 what is wanted and the second points to where the information is to be placed.
4330 The current list of available information is:
4331
4332 PCRE_CONFIG_UTF8
4333
4334 The output is an integer that is set to one if UTF-8 support is available;
4335 otherwise it is set to zero.
4336
4337 PCRE_CONFIG_NEWLINE
4338
4339 The output is an integer that it set to the value of the code that is used for
4340 newline. It is either LF (10) or CR (13).
4341
4342 PCRE_CONFIG_LINK_SIZE
4343
4344 The output is an integer that contains the number of bytes used for internal
4345 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4346
4347 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4348
4349 The output is an integer that contains the threshold above which the POSIX
4350 interface uses malloc() for output vectors. See item 31 above.
4351
4352 PCRE_CONFIG_MATCH_LIMIT
4353
4354 The output is an unsigned integer that contains the default limit of the number
4355 of match() calls in a pcre_exec() execution. See 42 above.
4356
4357 44. pcretest has been upgraded by the addition of the -C option. This causes it
4358 to extract all the available output from the new pcre_config() function, and to
4359 output it. The program then exits immediately.
4360
4361 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4362 order to support additional features. One way would have been to define
4363 pcre_exec2() (for example) with extra arguments, but this would not have been
4364 extensible, and would also have required all calls to the original function to
4365 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4366 is used for passing in "extra" data from pcre_study().
4367
4368 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4369 contains the following fields:
4370
4371 flags a bitmap indicating which of the following fields are set
4372 study_data opaque data from pcre_study()
4373 match_limit a way of specifying a limit on match() calls for a specific
4374 call to pcre_exec()
4375 callout_data data for callouts (see 49 below)
4376
4377 The flag bits are also defined in pcre.h, and are
4378
4379 PCRE_EXTRA_STUDY_DATA
4380 PCRE_EXTRA_MATCH_LIMIT
4381 PCRE_EXTRA_CALLOUT_DATA
4382
4383 The pcre_study() function now returns one of these new pcre_extra blocks, with
4384 the actual study data pointed to by the study_data field, and the
4385 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4386 before. That is, this change is entirely upwards-compatible and requires no
4387 change to existing code.
4388
4389 If you want to pass in additional data to pcre_exec(), you can either place it
4390 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4391 block.
4392
4393 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4394 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4395 times with different match limits, until it finds the minimum value needed for
4396 pcre_exec() to complete. The value is then output. This can be instructive; for
4397 most simple matches the number is quite small, but for pathological cases it
4398 gets very large very quickly.
4399
4400 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4401 returns the size of the data block pointed to by the study_data field in a
4402 pcre_extra block, that is, the value that was passed as the argument to
4403 pcre_malloc() when PCRE was getting memory in which to place the information
4404 created by pcre_study(). The fourth argument should point to a size_t variable.
4405 pcretest has been extended so that this information is shown after a successful
4406 pcre_study() call when information about the compiled regex is being displayed.
4407
4408 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4409 because what follows is always an absolute path. (Later: it turns out that this
4410 is more than cosmetic for MinGW, because it doesn't like empty path
4411 components.)
4412
4413 49. Some changes have been made to the callout feature (see 28 above):
4414
4415 (i) A callout function now has three choices for what it returns:
4416
4417 0 => success, carry on matching
4418 > 0 => failure at this point, but backtrack if possible
4419 < 0 => serious error, return this value from pcre_exec()
4420
4421 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4422 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4423 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4424 use by callout functions. It will never be used by PCRE itself.
4425
4426 (ii) The pcre_extra structure (see 45 above) has a void * field called
4427 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4428 pcre_callout_block structure has a field of the same name. The contents of
4429 the field passed in the pcre_extra structure are passed to the callout
4430 function in the corresponding field in the callout block. This makes it
4431 easier to use the same callout-containing regex from multiple threads. For
4432 testing, the pcretest program has a new data escape
4433
4434 \C*n pass the number n (may be negative) as callout_data
4435
4436 If the callout function in pcretest receives a non-zero value as
4437 callout_data, it returns that value.
4438
4439 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4440 there were some redundant $(CFLAGS) in commands that are now specified as
4441 $(LINK), which already includes $(CFLAGS).
4442
4443 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4444 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4445 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4446 one-byte characters throughout. Note that case-insensitive matching applies
4447 only to characters whose values are less than 256. PCRE doesn't support the
4448 notion of cases for higher-valued characters.
4449
4450 (i) A character class whose characters are all within 0-255 is handled as
4451 a bit map, and the map is inverted for negative classes. Previously, a
4452 character > 255 always failed to match such a class; however it should
4453 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4454
4455 (ii) A negated character class with a single character < 255 is coded as
4456 "not this character" (OP_NOT). This wasn't working properly when the test
4457 character was multibyte, either singly or repeated.
4458
4459 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4460 mode, for example: \x{100}{2,3}.
4461
4462 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4463 singly or repeated) now correctly test multibyte characters. However,
4464 PCRE doesn't recognize any characters with values greater than 255 as
4465 digits, spaces, or word characters. Such characters always match \D, \S,
4466 and \W, and never match \d, \s, or \w.
4467
4468 (v) Classes may now contain characters and character ranges with values
4469 greater than 255. For example: [ab\x{100}-\x{400}].
4470
4471 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4472 PCRE in UTF-8 mode.
4473
4474 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4475 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4476 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4477 value.)
4478
4479 53. The single man page has become too large. I have therefore split it up into
4480 a number of separate man pages. These also give rise to individual HTML pages;
4481 these are now put in a separate directory, and there is an index.html page that
4482 lists them all. Some hyperlinking between the pages has been installed.
4483
4484 54. Added convenience functions for handling named capturing parentheses.
4485
4486 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4487 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4488 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4489 are faulted.
4490
4491 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4492 calling configure. These values are used when compiling the dftables.c program
4493 which is run to generate the source of the default character tables. They
4494 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4495 you will need to set these values.
4496
4497 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4498
4499
4500 Version 3.9 02-Jan-02
4501 ---------------------
4502
4503 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4504
4505 2. If --disable-static was given, the building process failed when trying to
4506 build pcretest and pcregrep. (For some reason it was using libtool to compile
4507 them, which is not right, as they aren't part of the library.)
4508
4509
4510 Version 3.8 18-Dec-01
4511 ---------------------
4512
4513 1. The experimental UTF-8 code was completely screwed up. It was packing the
4514 bytes in the wrong order. How dumb can you get?
4515
4516
4517 Version 3.7 29-Oct-01
4518 ---------------------
4519
4520 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4521 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4522 this didn't happen under Solaris 8, where I normally test things.
4523
4524 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4525 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4526 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4527 this fix an hour or so after the initial 3.7 release.)
4528
4529
4530 Version 3.6 23-Oct-01
4531 ---------------------
4532
4533 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4534 offsets passed as NULL with zero offset count.
4535
4536 2. The config.guess and config.sub files had not been updated when I moved to
4537 the latest autoconf.
4538
4539
4540 Version 3.5 15-Aug-01
4541 ---------------------
4542
4543 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4544 had been forgotten.
4545
4546 2. By using declared but undefined structures, we can avoid using "void"
4547 definitions in pcre.h while keeping the internal definitions of the structures
4548 private.
4549
4550 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4551 user point of view, this means that both static and shared libraries are built
4552 by default, but this can be individually controlled. More of the work of
4553 handling this static/shared cases is now inside libtool instead of PCRE's make
4554 file.
4555
4556 4. The pcretest utility is now installed along with pcregrep because it is
4557 useful for users (to test regexs) and by doing this, it automatically gets
4558 relinked by libtool. The documentation has been turned into a man page, so
4559 there are now .1, .txt, and .html versions in /doc.
4560
4561 5. Upgrades to pcregrep:
4562 (i) Added long-form option names like gnu grep.
4563 (ii) Added --help to list all options with an explanatory phrase.
4564 (iii) Added -r, --recursive to recurse into sub-directories.
4565 (iv) Added -f, --file to read patterns from a file.
4566
4567 6. pcre_exec() was referring to its "code" argument before testing that
4568 argument for NULL (and giving an error if it was NULL).
4569
4570 7. Upgraded Makefile.in to allow for compiling in a different directory from
4571 the source directory.
4572
4573 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4574 options bits, the pointer it was passed was to an int instead of to an unsigned
4575 long int. This mattered only on 64-bit systems.
4576
4577 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4578 generated) instead of pcre.in, which it its source. Also made the same change
4579 in several of the .c files.
4580
4581 10. A new release of gcc defines printf() as a macro, which broke pcretest
4582 because it had an ifdef in the middle of a string argument for printf(). Fixed
4583 by using separate calls to printf().
4584
4585 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4586 script, to force use of CR or LF instead of \n in the source. On non-Unix
4587 systems, the value can be set in config.h.
4588
4589 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4590 absolute limit. Changed the text of the error message to make this clear, and
4591 likewise updated the man page.
4592
4593 13. The limit of 99 on the number of capturing subpatterns has been removed.
4594 The new limit is 65535, which I hope will not be a "real" limit.
4595
4596
4597 Version 3.4 22-Aug-00
4598 ---------------------
4599
4600 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4601
4602 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4603
4604
4605 Version 3.3 01-Aug-00
4606 ---------------------
4607
4608 1. If an octal character was given, but the value was greater than \377, it
4609 was not getting masked to the least significant bits, as documented. This could
4610 lead to crashes in some systems.
4611
4612 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4613 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4614
4615 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4616 These just pass their arguments on to (pcre_free)(), but they are provided
4617 because some uses of PCRE bind it to non-C systems that can call its functions,
4618 but cannot call free() or pcre_free() directly.
4619
4620 4. Add "make test" as a synonym for "make check". Corrected some comments in
4621 the Makefile.
4622
4623 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4624 Makefile.
4625
4626 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4627 command called pgrep for grepping around the active processes.
4628
4629 7. Added the beginnings of support for UTF-8 character strings.
4630
4631 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4632 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4633 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4634 out for the ar command.)
4635
4636
4637 Version 3.2 12-May-00
4638 ---------------------
4639
4640 This is purely a bug fixing release.
4641
4642 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4643 of ZA. This was just one example of several cases that could provoke this bug,
4644 which was introduced by change 9 of version 2.00. The code for breaking
4645 infinite loops after an iteration that matches an empty string was't working
4646 correctly.
4647
4648 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4649 when matched against abbab (for example). After matching an empty string, it
4650 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4651 caused it to match further down the string than it should.
4652
4653 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4654 was there because it doesn't seem to be needed, and it causes trouble on some
4655 systems, as it is not a Standard C header. It has been removed.
4656
4657 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4658 were reported on the Macintosh. The changes were from
4659
4660 while ((c = *(++ptr)) != 0 && c != '\n');
4661 to
4662 while ((c = *(++ptr)) != 0 && c != '\n') ;
4663
4664 Totally extraordinary, but if that's what it takes...
4665
4666 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4667 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4668 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4669 assumes the way PCRE uses memmove() (always moving upwards).
4670
4671 6. PCRE is being used in one environment where strchr() is not available. There
4672 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4673 faster code anyway.
4674
4675
4676 Version 3.1 09-Feb-00
4677 ---------------------
4678
4679 The only change in this release is the fixing of some bugs in Makefile.in for
4680 the "install" target:
4681
4682 (1) It was failing to install pcreposix.h.
4683
4684 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4685
4686
4687 Version 3.0 01-Feb-00
4688 ---------------------
4689
4690 1. Add support for the /+ modifier to perltest (to output $` like it does in
4691 pcretest).
4692
4693 2. Add support for the /g modifier to perltest.
4694
4695 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4696 matches null strings.
4697
4698 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4699 pattern. Perl treats empty patterns specially - it reuses the most recent
4700 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4701 effect.
4702
4703 5. The POSIX interface was broken in that it was just handing over the POSIX
4704 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4705 required a bigger vector, with some working space on the end. This means that
4706 the POSIX wrapper now has to get and free some memory, and copy the results.
4707
4708 6. Added some simple autoconf support, placing the test data and the
4709 documentation in separate directories, re-organizing some of the
4710 information files, and making it build pcre-config (a GNU standard). Also added
4711 libtool support for building PCRE as a shared library, which is now the
4712 default.
4713
4714 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4715 09 are not valid octal constants. Single digits will be used for minor values
4716 less than 10.
4717
4718 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4719 existing programs that set these in the POSIX interface can use PCRE without
4720 modification.
4721
4722 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4723 return all that pcre_info() returns, plus additional data. The pcre_info()
4724 function is retained for compatibility, but is considered to be obsolete.
4725
4726 10. Added experimental recursion feature (?R) to handle one common case that
4727 Perl 5.6 will be able to do with (?p{...}).
4728
4729 11. Added support for POSIX character classes like [:alpha:], which Perl is
4730 adopting.
4731
4732
4733 Version 2.08 31-Aug-99
4734 ----------------------
4735
4736 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4737 trying to match at the startoffset position, but instead was moving forward to
4738 the next newline as if a previous match had failed.
4739
4740 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4741 and could get into a loop if a null string was matched other than at the start
4742 of the subject.
4743
4744 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4745 be distinguished at compile time, and for completeness also added PCRE_DATE.
4746
4747 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4748 in GnuWin32 environments.
4749
4750
4751 Version 2.07 29-Jul-99
4752 ----------------------
4753
4754 1. The documentation is now supplied in plain text form and HTML as well as in
4755 the form of man page sources.
4756
4757 2. C++ compilers don't like assigning (void *) values to other pointer types.
4758 In particular this affects malloc(). Although there is no problem in Standard
4759 C, I've put in casts to keep C++ compilers happy.
4760
4761 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4762 should be (const char *).
4763
4764 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4765 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4766 However, I haven't made this a standard facility. The documentation doesn't
4767 mention it, and the Makefile doesn't support it.
4768
4769 5. The Makefile now contains an "install" target, with editable destinations at
4770 the top of the file. The pcretest program is not installed.
4771
4772 6. pgrep -V now gives the PCRE version number and date.
4773
4774 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4775 causing the entire string to be ignored, instead of just the last character.
4776
4777 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4778 non-matching string, it can take a very, very long time, even for strings of
4779 quite modest length, because of the nested recursion. PCRE now does better in
4780 some of these cases. It does this by remembering the last required literal
4781 character in the pattern, and pre-searching the subject to ensure it is present
4782 before running the real match. In other words, it applies a heuristic to detect
4783 some types of certain failure quickly, and in the above example, if presented
4784 with a string that has no trailing " it gives "no match" very quickly.
4785
4786 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4787 other alternatives are tried instead.
4788
4789
4790 Version 2.06 09-Jun-99
4791 ----------------------
4792
4793 1. Change pcretest's output for amount of store used to show just the code
4794 space, because the remainder (the data block) varies in size between 32-bit and
4795 64-bit systems.
4796
4797 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4798 start matching at. This allows lookbehinds to work when searching for multiple
4799 occurrences in a string.
4800
4801 3. Added additional options to pcretest for testing multiple occurrences:
4802
4803 /+ outputs the rest of the string that follows a match
4804 /g loops for multiple occurrences, using the new startoffset argument
4805 /G loops for multiple occurrences by passing an incremented pointer
4806
4807 4. PCRE wasn't doing the "first character" optimization for patterns starting
4808 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4809 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4810 the letter 'x'. On long subject strings, this gives a significant speed-up.
4811
4812
4813 Version 2.05 21-Apr-99
4814 ----------------------
4815
4816 1. Changed the type of magic_number from int to long int so that it works
4817 properly on 16-bit systems.
4818
4819 2. Fixed a bug which caused patterns starting with .* not to work correctly
4820 when the subject string contained newline characters. PCRE was assuming
4821 anchoring for such patterns in all cases, which is not correct because .* will
4822 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4823 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4824 must be retried after every newline in the subject.
4825
4826
4827 Version 2.04 18-Feb-99
4828 ----------------------
4829
4830 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4831 computation of the store needed to hold the pattern was incorrect (too large).
4832 If such patterns were nested a few deep, this could multiply and become a real
4833 problem.
4834
4835 2. Added /M option to pcretest to show the memory requirement of a specific
4836 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4837
4838 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4839 compiled in such a way that the backtracking after subsequent failure was
4840 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4841 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4842
4843
4844 Version 2.03 02-Feb-99
4845 ----------------------
4846
4847 1. Fixed typo and small mistake in man page.
4848
4849 2. Added 4th condition (GPL supersedes if conflict) and created separate
4850 LICENCE file containing the conditions.
4851
4852 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4853 Perl, that is the internal \ allows the delimiter to be included in the
4854 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4855 the final delimiter, add \ to the end of the pattern (to test the error).
4856
4857 4. Added the convenience functions for extracting substrings after a successful
4858 match. Updated pcretest to make it able to test these functions.
4859
4860
4861 Version 2.02 14-Jan-99
4862 ----------------------
4863
4864 1. Initialized the working variables associated with each extraction so that
4865 their saving and restoring doesn't refer to uninitialized store.
4866
4867 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4868 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4869 fix the problem.
4870
4871 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4872 calls, and wasn't printing the correct value for compiling calls. Increased the
4873 default value of LOOPREPEAT, and the number of significant figures in the
4874 times.
4875
4876 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4877
4878 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4879 a building problem on Windows NT with a FAT file system.
4880
4881
4882 Version 2.01 21-Oct-98
4883 ----------------------
4884
4885 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4886 to character tables built by pcre_maketables() in the current locale. If NULL
4887 is passed, the default tables are used.
4888
4889
4890 Version 2.00 24-Sep-98
4891 ----------------------
4892
4893 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4894 it any more.
4895
4896 2. Allow quantification of (?>) groups, and make it work correctly.
4897
4898 3. The first character computation wasn't working for (?>) groups.
4899
4900 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4901 end of the subject) and add 5.005's \z, which really does match only at the
4902 very end of the subject.
4903
4904 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4905
4906 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4907 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4908 localized options. All options to pcre_study() were also removed.
4909
4910 7. Add other new features from 5.005:
4911
4912 $(?<= positive lookbehind
4913 $(?<! negative lookbehind
4914 (?imsx-imsx) added the unsetting capability
4915 such a setting is global if at outer level; local otherwise
4916 (?imsx-imsx:) non-capturing groups with option setting
4917 (?(cond)re|re) conditional pattern matching
4918
4919 A backreference to itself in a repeated group matches the previous
4920 captured string.
4921
4922 8. General tidying up of studying (both automatic and via "study")
4923 consequential on the addition of new assertions.
4924
4925 9. As in 5.005, unlimited repeated groups that could match an empty substring
4926 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4927 runtime if any iteration does actually match an empty substring.
4928
4929 10. Include the RunTest script in the distribution.
4930
4931 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4932 discrepancies, some of which were old and were also with respect to 5.004. They
4933 have now been fixed.
4934
4935
4936 Version 1.09 28-Apr-98
4937 ----------------------
4938
4939 1. A negated single character class followed by a quantifier with a minimum
4940 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4941 program crashes, or just wrong answers. This did not apply to negated classes
4942 containing more than one character, or to minima other than one.
4943
4944
4945 Version 1.08 27-Mar-98
4946 ----------------------
4947
4948 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4949
4950 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4951 latter must appear before anything that relies on it in the pattern.
4952
4953
4954 Version 1.07 16-Feb-98
4955 ----------------------
4956
4957 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4958 repeat of a potentially empty string).
4959
4960
4961 Version 1.06 23-Jan-98
4962 ----------------------
4963
4964 1. Added Markus Oberhumer's little patches for C++.
4965
4966 2. Literal strings longer than 255 characters were broken.
4967
4968
4969 Version 1.05 23-Dec-97
4970 ----------------------
4971
4972 1. Negated character classes containing more than one character were failing if
4973 PCRE_CASELESS was set at run time.
4974
4975
4976 Version 1.04 19-Dec-97
4977 ----------------------
4978
4979 1. Corrected the man page, where some "const" qualifiers had been omitted.
4980
4981 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4982 input syntax.
4983
4984 3. Fixed memory leak which occurred when a regex with back references was
4985 matched with an offsets vector that wasn't big enough. The temporary memory
4986 that is used in this case wasn't being freed if the match failed.
4987
4988 4. Tidied pcretest to ensure it frees memory that it gets.
4989
4990 5. Temporary memory was being obtained in the case where the passed offsets
4991 vector was exactly big enough.
4992
4993 6. Corrected definition of offsetof() from change 5 below.
4994
4995 7. I had screwed up change 6 below and broken the rules for the use of
4996 setjmp(). Now fixed.
4997
4998
4999 Version 1.03 18-Dec-97
5000 ----------------------
5001
5002 1. A erroneous regex with a missing opening parenthesis was correctly
5003 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5004 on some systems.
5005
5006 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5007 it was reported that one broken compiler failed on the former because "code" is
5008 also an independent variable.
5009
5010 3. The erroneous regex a[]b caused an array overrun reference.
5011
5012 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5013 fail on data ending with that character. (It was going on too far, and checking
5014 the next character, typically a binary zero.) This was specific to the
5015 optimized code for single-character negative classes.
5016
5017 5. Added a contributed patch from the TIN world which does the following:
5018
5019 + Add an undef for memmove, in case the the system defines a macro for it.
5020
5021 + Add a definition of offsetof(), in case there isn't one. (I don't know
5022 the reason behind this - offsetof() is part of the ANSI standard - but
5023 it does no harm).
5024
5025 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5026 most of the places where whitespace preceded '#'. I have given up and
5027 allowed the remaining 2 cases to be at the margin.
5028
5029 + Rename some variables in pcre to eliminate shadowing. This seems very
5030 pedantic, but does no harm, of course.
5031
5032 6. Moved the call to setjmp() into its own function, to get rid of warnings
5033 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5034
5035 7. Constructs such as \d{8,} were compiling into the equivalent of
5036 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5037 outcome, but in this particular case used more store than had been allocated,
5038 which caused the bug to be discovered because it threw up an internal error.
5039
5040 8. The debugging code in both pcre and pcretest for outputting the compiled
5041 form of a regex was going wrong in the case of back references followed by
5042 curly-bracketed repeats.
5043
5044
5045 Version 1.02 12-Dec-97
5046 ----------------------
5047
5048 1. Typos in pcre.3 and comments in the source fixed.
5049
5050 2. Applied a contributed patch to get rid of places where it used to remove
5051 'const' from variables, and fixed some signed/unsigned and uninitialized
5052 variable warnings.
5053
5054 3. Added the "runtest" target to Makefile.
5055
5056 4. Set default compiler flag to -O2 rather than just -O.
5057
5058
5059 Version 1.01 19-Nov-97
5060 ----------------------
5061
5062 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5063 like /([ab]*)*/, that is, for classes with more than one character in them.
5064
5065 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5066 as /((?>a*))*/ (a PCRE_EXTRA facility).
5067
5068
5069 Version 1.00 18-Nov-97
5070 ----------------------
5071
5072 1. Added compile-time macros to support systems such as SunOS4 which don't have
5073 memmove() or strerror() but have other things that can be used instead.
5074
5075 2. Arranged that "make clean" removes the executables.
5076
5077
5078 Version 0.99 27-Oct-97
5079 ----------------------
5080
5081 1. Fixed bug in code for optimizing classes with only one character. It was
5082 initializing a 32-byte map regardless, which could cause it to run off the end
5083 of the memory it had got.
5084
5085 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5086
5087
5088 Version 0.98 22-Oct-97
5089 ----------------------
5090
5091 1. Fixed bug in code for handling temporary memory usage when there are more
5092 back references than supplied space in the ovector. This could cause segfaults.
5093
5094
5095 Version 0.97 21-Oct-97
5096 ----------------------
5097
5098 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5099
5100 2. Optimized negated single characters not to use a bit map.
5101
5102 3. Brought error texts together as macro definitions; clarified some of them;
5103 fixed one that was wrong - it said "range out of order" when it meant "invalid
5104 escape sequence".
5105
5106 4. Changed some char * arguments to const char *.
5107
5108 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5109
5110 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5111 pcretest.
5112
5113
5114 Version 0.96 16-Oct-97
5115 ----------------------
5116
5117 1. Added a simple "pgrep" utility to the distribution.
5118
5119 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5120 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5121 where "ddd" means "one or more decimal digits".
5122
5123 3. Fixed serious bug. If a pattern had a back reference, but the call to
5124 pcre_exec() didn't supply a large enough ovector to record the related
5125 identifying subpattern, the match always failed. PCRE now remembers the number
5126 of the largest back reference, and gets some temporary memory in which to save
5127 the offsets during matching if necessary, in order to ensure that
5128 backreferences always work.
5129
5130 4. Increased the compatibility with Perl in a number of ways:
5131
5132 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5133 to request this handling. The option can be set at compile or exec time.
5134
5135 (b) $ matches before a terminating newline by default; an option
5136 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5137 mode). The option can be set at compile or exec time.
5138
5139 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5140 the same as Perl's. If the decimal number it represents is less than 10
5141 or there aren't that many previous left capturing parentheses, an octal
5142 escape is read. Inside a character class, it's always an octal escape,
5143 even if it is a single digit.
5144
5145 (d) An escaped but undefined alphabetic character is taken as a literal,
5146 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5147 escapes.
5148
5149 (e) {0} is now permitted. (The previous item is removed from the compiled
5150 pattern).
5151
5152 5. Changed all the names of code files so that the basic parts are no longer
5153 than 10 characters, and abolished the teeny "globals.c" file.
5154
5155 6. Changed the handling of character classes; they are now done with a 32-byte
5156 bit map always.
5157
5158 7. Added the -d and /D options to pcretest to make it possible to look at the
5159 internals of compilation without having to recompile pcre.
5160
5161
5162 Version 0.95 23-Sep-97
5163 ----------------------
5164
5165 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5166 \x20 at the start of a run of normal characters. These were being treated as
5167 real characters, instead of the source characters being re-checked.
5168
5169
5170 Version 0.94 18-Sep-97
5171 ----------------------
5172
5173 1. The functions are now thread-safe, with the caveat that the global variables
5174 containing pointers to malloc() and free() or alternative functions are the
5175 same for all threads.
5176
5177 2. Get pcre_study() to generate a bitmap of initial characters for non-
5178 anchored patterns when this is possible, and use it if passed to pcre_exec().
5179
5180
5181 Version 0.93 15-Sep-97
5182 ----------------------
5183
5184 1. /(b)|(:+)/ was computing an incorrect first character.
5185
5186 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5187 but not actually doing anything yet.
5188
5189 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5190 as Perl does (e.g. [-az] or [az-]).
5191
5192 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5193 all possible positions.
5194
5195 5. Split up into different modules to avoid including unneeded functions in a
5196 compiled binary. However, compile and exec are still in one module. The "study"
5197 function is split off.
5198
5199 6. The character tables are now in a separate module whose source is generated
5200 by an auxiliary program - but can then be edited by hand if required. There are
5201 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5202 toupper() in the code.
5203
5204 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5205 make them global. Abolish the function for setting them, as the caller can now
5206 set them directly.
5207
5208
5209 Version 0.92 11-Sep-97
5210 ----------------------
5211
5212 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5213 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5214
5215 2. Caseless matching was not working in character classes if the characters in
5216 the pattern were in upper case.
5217
5218 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5219
5220 4. Make PCRE_ANCHORED public and accept as a compile option.
5221
5222 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5223 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5224 pass them.
5225
5226 6. Give an error if bad option bits passed at compile or run time.
5227
5228 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5229 pcretest to cause it to pass that flag.
5230
5231 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5232 options, and the first character, if set.
5233
5234 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5235
5236
5237 Version 0.91 10-Sep-97
5238 ----------------------
5239
5240 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5241 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5242
5243 2. PCRE was looping on encountering an indefinitely repeated back reference to
5244 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5245 Perl does - treats the match as successful.
5246
5247 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5