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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1335 - (show annotations)
Tue May 28 09:13:59 2013 UTC (6 years, 6 months ago) by ph10
File size: 247389 byte(s)
Final source file tidies for 8.33 release.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.33 28-May-2013
5 --------------------------
6
7 1. Added 'U' to some constants that are compared to unsigned integers, to
8 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9 variables that are added to signed variables, to ensure the result is
10 signed and can be negated.
11
12 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
13 CMake config files.
14
15 3. Revise the creation of config.h.generic so that all boolean macros are
16 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17 overriding via -D on the command line possible.
18
19 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20 to unsigned int is reported to make a quite noticeable speed difference in
21 a specific Windows environment. Testing on Linux did also appear to show
22 some benefit (and it is clearly not harmful). Also fixed the definition of
23 Xop which should be unsigned.
24
25 5. Related to (4), changing the definition of the intermediate variable cc
26 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27 improvements.
28
29 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
30 unnecessary spaces.
31
32 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33 and later.
34
35 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36
37 9. Optimizing fast_forward_start_bits in JIT.
38
39 10. Adding support for callouts in JIT, and fixing some issues revealed
40 during this work. Namely:
41
42 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43
44 (b) Minimum length was not checked before the matching is started.
45
46 11. The value of capture_last that is passed to callouts was incorrect in some
47 cases when there was a capture on one path that was subsequently abandoned
48 after a backtrack. Also, the capture_last value is now reset after a
49 recursion, since all captures are also reset in this case.
50
51 12. The interpreter no longer returns the "too many substrings" error in the
52 case when an overflowing capture is in a branch that is subsequently
53 abandoned after a backtrack.
54
55 13. In the pathological case when an offset vector of size 2 is used, pcretest
56 now prints out the matched string after a yield of 0 or 1.
57
58 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59 Only supported by the JIT compiler at the moment.
60
61 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
62
63 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
64 offset of the starting point of the matching process, provided the offsets
65 vector is large enough.
66
67 17. The \A escape now records a lookbehind value of 1, though its execution
68 does not actually inspect the previous character. This is to ensure that,
69 in partial multi-segment matching, at least one character from the old
70 segment is retained when a new segment is processed. Otherwise, if there
71 are no lookbehinds in the pattern, \A might match incorrectly at the start
72 of a new segment.
73
74 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
75
76 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
77 gives some modest performance improvement in 8-bit mode.
78
79 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
80 be expressed in certain programming languages using Universal Character
81 Names.
82
83 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
84 which points out that "non characters" are not "characters that may not
85 appear in Unicode strings" but rather "characters that are reserved for
86 internal use and have only local meaning".
87
88 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
89 there was a conditional group that depended on an assertion, if the
90 assertion was false, the callout that immediately followed the alternation
91 in the condition was skipped when pcre_exec() was used for matching.
92
93 23. Allow an explicit callout to be inserted before an assertion that is the
94 condition for a conditional group, for compatibility with automatic
95 callouts, which always insert a callout at this point.
96
97 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
98 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
99
100 25. (*PRUNE) is now supported by the JIT compiler.
101
102 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
103
104 27. Fix the case where there are two or more SKIPs with arguments that may be
105 ignored.
106
107 28. (*SKIP) is now supported by the JIT compiler.
108
109 29. (*THEN) is now supported by the JIT compiler.
110
111 30. Update RunTest with additional test selector options.
112
113 31. The way PCRE handles backtracking verbs has been changed in two ways.
114
115 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
116 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
117 backtracking. In some cases this makes it more Perl-compatible, but Perl's
118 rather obscure rules do not always do the same thing.
119
120 (2) Previously, backtracking verbs were confined within assertions. This is
121 no longer the case for positive assertions, except for (*ACCEPT). Again,
122 this sometimes improves Perl compatibility, and sometimes does not.
123
124 32. A number of tests that were in test 2 because Perl did things differently
125 have been moved to test 1, because either Perl or PCRE has changed, and
126 these tests are now compatible.
127
128 32. Backtracking control verbs are now handled in the same way in JIT and
129 interpreter.
130
131 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
132 contained a forward subroutine reference caused a compile error.
133
134 34. Auto-detect and optimize limited repetitions in JIT.
135
136 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
137 blocking (*UTF) etc.
138
139 36. In the interpreter, maximizing pattern repetitions for characters and
140 character types now use tail recursion, which reduces stack usage.
141
142 37. The value of the max lookbehind was not correctly preserved if a compiled
143 and saved regex was reloaded on a host of different endianness.
144
145 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
146 of the compiled pattern block, expand the flags field from 16 to 32 bits
147 because it was almost full.
148
149 39. Try madvise first before posix_madvise.
150
151 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
152 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
153 original change locked out all matches of empty strings. This has been
154 changed so that one match of an empty string per line is recognized.
155 Subsequent searches on the same line (for colouring or for --only-matching,
156 for example) do not recognize empty strings.
157
158 41. Applied a user patch to fix a number of spelling mistakes in comments.
159
160 42. Data lines longer than 65536 caused pcretest to crash.
161
162 43. Clarified the data type for length and startoffset arguments for pcre_exec
163 and pcre_dfa_exec in the function-specific man pages, where they were
164 explicitly stated to be in bytes, never having been updated. I also added
165 some clarification to the pcreapi man page.
166
167 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
168 a segmentation fault.
169
170
171 Version 8.32 30-November-2012
172 -----------------------------
173
174 1. Improved JIT compiler optimizations for first character search and single
175 character iterators.
176
177 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
178 Patch by Daniel Richard G.
179
180 3. Single character iterator optimizations in the JIT compiler.
181
182 4. Improved JIT compiler optimizations for character ranges.
183
184 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
185 Reported by Giuseppe D'Angelo.
186
187 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
188 of a line, was being set incorrectly in cases where .* appeared inside
189 atomic brackets at the start of a pattern, or where there was a subsequent
190 *PRUNE or *SKIP.
191
192 7. Improved instruction cache flush for POWER/PowerPC.
193 Patch by Daniel Richard G.
194
195 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
196 grep:
197
198 (a) There is now no limit to the number of patterns to be matched.
199
200 (b) An error is given if a pattern is too long.
201
202 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
203 are now supported.
204
205 (d) --exclude-from and --include-from (multiple use) have been added.
206
207 (e) Exclusions and inclusions now apply to all files and directories, not
208 just to those obtained from scanning a directory recursively.
209
210 (f) Multiple uses of -f and --file-list are now supported.
211
212 (g) In a Windows environment, the default for -d has been changed from
213 "read" (the GNU grep default) to "skip", because otherwise the presence
214 of a directory in the file list provokes an error.
215
216 (h) The documentation has been revised and clarified in places.
217
218 9. Improve the matching speed of capturing brackets.
219
220 10. Changed the meaning of \X so that it now matches a Unicode extended
221 grapheme cluster.
222
223 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
224 out POSIX threads when JIT support is configured.
225
226 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
227
228 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
229 structure could go wrong in environments where size_t is not the same size
230 as int.
231
232 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
233 set.
234
235 15. The EBCDIC support had decayed; later updates to the code had included
236 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
237 general tidy up of EBCDIC-related issues, and the documentation was also
238 not quite right. There is now a test that can be run on ASCII systems to
239 check some of the EBCDIC-related things (but is it not a full test).
240
241 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
242 in a small tidy to the code.
243
244 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
245
246 18. If the --only-matching (-o) option in pcregrep is specified multiple
247 times, each one causes appropriate output. For example, -o1 -o2 outputs the
248 substrings matched by the 1st and 2nd capturing parentheses. A separating
249 string can be specified by --om-separator (default empty).
250
251 19. Improving the first n character searches.
252
253 20. Turn case lists for horizontal and vertical white space into macros so that
254 they are defined only once.
255
256 21. This set of changes together give more compatible Unicode case-folding
257 behaviour for characters that have more than one other case when UCP
258 support is available.
259
260 (a) The Unicode property table now has offsets into a new table of sets of
261 three or more characters that are case-equivalent. The MultiStage2.py
262 script that generates these tables (the pcre_ucd.c file) now scans
263 CaseFolding.txt instead of UnicodeData.txt for character case
264 information.
265
266 (b) The code for adding characters or ranges of characters to a character
267 class has been abstracted into a generalized function that also handles
268 case-independence. In UTF-mode with UCP support, this uses the new data
269 to handle characters with more than one other case.
270
271 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
272 whose other case is greater than 256 are now correctly matched
273 caselessly. Previously, the high codepoint matched the low one, but not
274 vice versa.
275
276 (d) The processing of \h, \H, \v, and \ in character classes now makes use
277 of the new class addition function, using character lists defined as
278 macros alongside the case definitions of 20 above.
279
280 (e) Caseless back references now work with characters that have more than
281 one other case.
282
283 (f) General caseless matching of characters with more than one other case
284 is supported.
285
286 22. Unicode character properties were updated from Unicode 6.2.0
287
288 23. Improved CMake support under Windows. Patch by Daniel Richard G.
289
290 24. Add support for 32-bit character strings, and UTF-32
291
292 25. Major JIT compiler update (code refactoring and bugfixing).
293 Experimental Sparc 32 support is added.
294
295 26. Applied a modified version of Daniel Richard G's patch to create
296 pcre.h.generic and config.h.generic by "make" instead of in the
297 PrepareRelease script.
298
299 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
300 pcre_compile.c when checking for a zero character.
301
302 28. Introducing a native interface for JIT. Through this interface, the compiled
303 machine code can be directly executed. The purpose of this interface is to
304 provide fast pattern matching, so several sanity checks are not performed.
305 However, feature tests are still performed. The new interface provides
306 1.4x speedup compared to the old one.
307
308 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
309 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
310 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
311
312 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
313 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
314 the "old" RFC 2279). Instead, it ended up passing a negative length to
315 pcre_exec().
316
317 31. Add support for GCC's visibility feature to hide internal functions.
318
319 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
320 "unknown -C option" after outputting 0 or 1.
321
322 33. There is now support for generating a code coverage report for the test
323 suite in environments where gcc is the compiler and lcov is installed. This
324 is mainly for the benefit of the developers.
325
326 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
327 unaddressable using valgrind annotations, allowing valgrind to detect
328 invalid memory accesses. This is mainly for the benefit of the developers.
329
330 25. (*UTF) can now be used to start a pattern in any of the three libraries.
331
332 26. Give configure error if --enable-cpp but no C++ compiler found.
333
334
335 Version 8.31 06-July-2012
336 -------------------------
337
338 1. Fixing a wrong JIT test case and some compiler warnings.
339
340 2. Removed a bashism from the RunTest script.
341
342 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
343 to unsigned type, result still unsigned" that was given by an MS compiler
344 on encountering the code "-sizeof(xxx)".
345
346 4. Partial matching support is added to the JIT compiler.
347
348 5. Fixed several bugs concerned with partial matching of items that consist
349 of more than one character:
350
351 (a) /^(..)\1/ did not partially match "aba" because checking references was
352 done on an "all or nothing" basis. This also applied to repeated
353 references.
354
355 (b) \R did not give a hard partial match if \r was found at the end of the
356 subject.
357
358 (c) \X did not give a hard partial match after matching one or more
359 characters at the end of the subject.
360
361 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
362 a partial match for the string "\r".
363
364 (e) When newline was set to CRLF, the metacharacter "." did not recognize
365 a partial match for a CR character at the end of the subject string.
366
367 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
368 running pcretest, the text "(JIT)" added to the output whenever JIT is
369 actually used to run the match.
370
371 7. Individual JIT compile options can be set in pcretest by following -s+[+]
372 or /S+[+] with a digit between 1 and 7.
373
374 8. OP_NOT now supports any UTF character not just single-byte ones.
375
376 9. (*MARK) control verb is now supported by the JIT compiler.
377
378 10. The command "./RunTest list" lists the available tests without actually
379 running any of them. (Because I keep forgetting what they all are.)
380
381 11. Add PCRE_INFO_MAXLOOKBEHIND.
382
383 12. Applied a (slightly modified) user-supplied patch that improves performance
384 when the heap is used for recursion (compiled with --disable-stack-for-
385 recursion). Instead of malloc and free for each heap frame each time a
386 logical recursion happens, frames are retained on a chain and re-used where
387 possible. This sometimes gives as much as 30% improvement.
388
389 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
390 call.
391
392 14. As documented, (*COMMIT) is now confined to within a positive assertion.
393
394 15. It is now possible to link pcretest with libedit as an alternative to
395 libreadline.
396
397 16. (*COMMIT) control verb is now supported by the JIT compiler.
398
399 17. The Unicode data tables have been updated to Unicode 6.1.0.
400
401 18. Added --file-list option to pcregrep.
402
403 19. Added binary file support to pcregrep, including the -a, --binary-files,
404 -I, and --text options.
405
406 20. The madvise function is renamed for posix_madvise for QNX compatibility
407 reasons. Fixed by Giuseppe D'Angelo.
408
409 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
410
411 22. Changed the option for creating symbolic links for 16-bit man pages from
412 -s to -sf so that re-installing does not cause issues.
413
414 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
415
416 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
417 matches in certain environments (the workspace was not being correctly
418 retained). Also added to pcre_dfa_exec() a simple plausibility check on
419 some of the workspace data at the beginning of a restart.
420
421 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
422 was not doing so when it should - probably a typo introduced by SVN 528
423 (change 8.10/14).
424
425 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
426 \w+ when the character tables indicated that \x{c4} was a word character.
427 There were several related cases, all because the tests for doing a table
428 lookup were testing for characters less than 127 instead of 255.
429
430 27. If a pattern contains capturing parentheses that are not used in a match,
431 their slots in the ovector are set to -1. For those that are higher than
432 any matched groups, this happens at the end of processing. In the case when
433 there were back references that the ovector was too small to contain
434 (causing temporary malloc'd memory to be used during matching), and the
435 highest capturing number was not used, memory off the end of the ovector
436 was incorrectly being set to -1. (It was using the size of the temporary
437 memory instead of the true size.)
438
439 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
440 ovector size, it uses memory at the end of the block that it has got.
441
442 29. Check for an overlong MARK name and give an error at compile time. The
443 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
444
445 30. JIT compiler update.
446
447 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
448 Rill for the patch.
449
450 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
451
452 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
453
454 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
455 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
456 was enabled.
457
458 35. Improve JIT code generation for greedy plus quantifier.
459
460 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
461 1 to "aa" instead of to an empty string. The bug affected repeated groups
462 that could potentially match an empty string.
463
464 37. Optimizing single character iterators in JIT.
465
466 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
467 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
468 codepoints that are too big for the mode are faulted, and in a UTF mode,
469 disallowed codepoints are also faulted.
470
471 39. If PCRE was compiled with UTF support, in three places in the DFA
472 matcher there was code that should only have been obeyed in UTF mode, but
473 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
474 processing when bytes with values greater than 127 were present. In 16-bit
475 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
476 both cases the values are those that cannot be the first data item in a UTF
477 character. The three items that might have provoked this were recursions,
478 possessively repeated groups, and atomic groups.
479
480 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
481 and pcregrep, because some OS require shared objects to be explicitly
482 passed to ld, causing the link step to fail if they are not.
483
484 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
485 mode, patterns that started with \h* or \R* might be incorrectly matched.
486
487
488 Version 8.30 04-February-2012
489 -----------------------------
490
491 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
492 name is defined in ctype.h.
493
494 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
495 only in quite long subpatterns.
496
497 3. Removed the function pcre_info(), which has been obsolete and deprecated
498 since it was replaced by pcre_fullinfo() in February 2000.
499
500 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
501 match a (*MARK), and the match failed at the start of the subject, a
502 reference to memory before the start of the subject could occur. This bug
503 was introduced by fix 17 of release 8.21.
504
505 5. A reference to an unset group with zero minimum repetition was giving
506 totally wrong answers (in non-JavaScript-compatibility mode). For example,
507 /(another)?(\1?)test/ matched against "hello world test". This bug was
508 introduced in release 8.13.
509
510 6. Add support for 16-bit character strings (a large amount of work involving
511 many changes and refactorings).
512
513 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
514 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
515 from a file.
516
517 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
518 rounding is not applied in this particular case).
519
520 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
521 if they appear, or are escaped, in patterns.
522
523 10. Get rid of a number of -Wunused-but-set-variable warnings.
524
525 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
526 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
527 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
528 also returns the mark "x". This bug applied to capturing parentheses,
529 non-capturing parentheses, and atomic parentheses. It also applied to some
530 assertions.
531
532 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
533 information out of configure.ac instead of relying on pcre.h.generic, which
534 is not stored in the repository.
535
536 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
537 -lreadline.
538
539 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
540
541 15. Applied Graycode's patch to put the top-level frame on the stack rather
542 than the heap when not using the stack for recursion. This gives a
543 performance improvement in many cases when recursion is not deep.
544
545 16. Experimental code added to "pcretest -C" to output the stack frame size.
546
547
548 Version 8.21 12-Dec-2011
549 ------------------------
550
551 1. Updating the JIT compiler.
552
553 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
554 are added as well.
555
556 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
557 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
558 calling _pcre_jit_exec. Some extra comments are added.
559
560 4. (*MARK) settings inside atomic groups that do not contain any capturing
561 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
562 was introduced by change 18 for 8.20.
563
564 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
565 ECMA-262 standard.
566
567 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
568 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
569 This bug was probably introduced by change 9 of 8.13.
570
571 7. While fixing 6 above, I noticed that a number of other items were being
572 incorrectly rejected as "not fixed length". This arose partly because newer
573 opcodes had not been added to the fixed-length checking code. I have (a)
574 corrected the bug and added tests for these items, and (b) arranged for an
575 error to occur if an unknown opcode is encountered while checking for fixed
576 length instead of just assuming "not fixed length". The items that were
577 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
578 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
579 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
580
581 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
582 being incorrectly compiled and would have given unpredicatble results.
583
584 9. A possessively repeated subpattern with minimum repeat count greater than
585 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
586 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
587 the first (A) could occur when it should not.
588
589 10. Add a cast and remove a redundant test from the code.
590
591 11. JIT should use pcre_malloc/pcre_free for allocation.
592
593 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
594 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
595 is anything other than /usr, -L is still shown).
596
597 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
598
599 14. Perl does not support \N without a following name in a [] class; PCRE now
600 also gives an error.
601
602 15. If a forward reference was repeated with an upper limit of around 2000,
603 it caused the error "internal error: overran compiling workspace". The
604 maximum number of forward references (including repeats) was limited by the
605 internal workspace, and dependent on the LINK_SIZE. The code has been
606 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
607 the default depends on LINK_SIZE. There is a new upper limit (for safety)
608 of around 200,000 forward references. While doing this, I also speeded up
609 the filling in of repeated forward references.
610
611 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
612 incorrectly expecting the subject to contain another "a" after the start.
613
614 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
615 in the match, the SKIP should be ignored. This was not happening; instead
616 the SKIP was being treated as NOMATCH. For patterns such as
617 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
618 tested.
619
620 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
621 now much more compatible with Perl, in particular in cases where the result
622 is a non-match for a non-anchored pattern. For example, if
623 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
624 "m", where previously it did not return a name. A side effect of this
625 change is that for partial matches, the last encountered mark name is
626 returned, as for non matches. A number of tests that were previously not
627 Perl-compatible have been moved into the Perl-compatible test files. The
628 refactoring has had the pleasing side effect of removing one argument from
629 the match() function, thus reducing its stack requirements.
630
631 19. If the /S+ option was used in pcretest to study a pattern using JIT,
632 subsequent uses of /S (without +) incorrectly behaved like /S+.
633
634 21. Retrieve executable code size support for the JIT compiler and fixing
635 some warnings.
636
637 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
638 not work when the shorter character appeared right at the end of the
639 subject string.
640
641 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
642 systems.
643
644 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
645 output it when the /M option is used in pcretest.
646
647 25. The CheckMan script was not being included in the distribution. Also, added
648 an explicit "perl" to run Perl scripts from the PrepareRelease script
649 because this is reportedly needed in Windows.
650
651 26. If study data was being save in a file and studying had not found a set of
652 "starts with" bytes for the pattern, the data written to the file (though
653 never used) was taken from uninitialized memory and so caused valgrind to
654 complain.
655
656 27. Updated RunTest.bat as provided by Sheri Pierce.
657
658 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
659
660 29. Computation of memory usage for the table of capturing group names was
661 giving an unnecessarily large value.
662
663
664 Version 8.20 21-Oct-2011
665 ------------------------
666
667 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
668 a POSIX class. After further experiments with Perl, which convinced me that
669 Perl has bugs and confusions, a closing square bracket is no longer allowed
670 in a POSIX name. This bug also affected patterns with classes that started
671 with full stops.
672
673 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
674 captured substring, but while checking the failing first alternative,
675 substring 1 is temporarily captured. If the output vector supplied to
676 pcre_exec() was not big enough for this capture, the yield of the function
677 was still zero ("insufficient space for captured substrings"). This cannot
678 be totally fixed without adding another stack variable, which seems a lot
679 of expense for a edge case. However, I have improved the situation in cases
680 such as /(a)(b)x|abc/ matched against "abc", where the return code
681 indicates that fewer than the maximum number of slots in the ovector have
682 been set.
683
684 3. Related to (2) above: when there are more back references in a pattern than
685 slots in the output vector, pcre_exec() uses temporary memory during
686 matching, and copies in the captures as far as possible afterwards. It was
687 using the entire output vector, but this conflicts with the specification
688 that only 2/3 is used for passing back captured substrings. Now it uses
689 only the first 2/3, for compatibility. This is, of course, another edge
690 case.
691
692 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
693 main code base, and can be used by building with --enable-jit. When this is
694 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
695 runtime --no-jit option is given.
696
697 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
698 ovector, the return from the function was zero, implying that there were
699 other matches that did not fit. The correct "exactly full" value is now
700 returned.
701
702 6. If a subpattern that was called recursively or as a subroutine contained
703 (*PRUNE) or any other control that caused it to give a non-standard return,
704 invalid errors such as "Error -26 (nested recursion at the same subject
705 position)" or even infinite loops could occur.
706
707 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
708 computing the minimum length on reaching *ACCEPT, and so ended up with the
709 wrong value of 1 rather than 0. Further investigation indicates that
710 computing a minimum subject length in the presence of *ACCEPT is difficult
711 (think back references, subroutine calls), and so I have changed the code
712 so that no minimum is registered for a pattern that contains *ACCEPT.
713
714 8. If (*THEN) was present in the first (true) branch of a conditional group,
715 it was not handled as intended. [But see 16 below.]
716
717 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
718 Sheri Pierce.
719
720 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
721 the first byte in a match must be "a".
722
723 11. Change 17 for 8.13 increased the recursion depth for patterns like
724 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
725 pattern contains any instances of (*THEN). If it does not, the old
726 optimizations are restored. It would be nice to do this on a per-group
727 basis, but at the moment that is not feasible.
728
729 12. In some environments, the output of pcretest -C is CRLF terminated. This
730 broke RunTest's code that checks for the link size. A single white space
731 character after the value is now allowed for.
732
733 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
734 For "fr", it uses the Windows-specific input and output files.
735
736 14. If (*THEN) appeared in a group that was called recursively or as a
737 subroutine, it did not work as intended. [But see next item.]
738
739 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
740 pattern fragments (but not containing any | characters). If A and B are
741 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
742 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
743 D. In other words, Perl considers parentheses that do not contain any |
744 characters to be part of a surrounding alternative, whereas PCRE was
745 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
746 differently. PCRE now behaves in the same way as Perl, except in the case
747 of subroutine/recursion calls such as (?1) which have in any case always
748 been different (but PCRE had them first :-).
749
750 16. Related to 15 above: Perl does not treat the | in a conditional group as
751 creating alternatives. Such a group is treated in the same way as an
752 ordinary group without any | characters when processing (*THEN). PCRE has
753 been changed to match Perl's behaviour.
754
755 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
756 RunGrepTest script failed.
757
758 18. Change 22 for version 13 caused atomic groups to use more stack. This is
759 inevitable for groups that contain captures, but it can lead to a lot of
760 stack use in large patterns. The old behaviour has been restored for atomic
761 groups that do not contain any capturing parentheses.
762
763 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
764 suppress the check for a minimum subject length at run time. (If it was
765 given to pcre_exec() or pcre_dfa_exec() it did work.)
766
767 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
768 fail to work when decoding hex characters in data strings in EBCDIC
769 environments.
770
771 21. It appears that in at least one Mac OS environment, the isxdigit() function
772 is implemented as a macro that evaluates to its argument more than once,
773 contravening the C 90 Standard (I haven't checked a later standard). There
774 was an instance in pcretest which caused it to go wrong when processing
775 \x{...} escapes in subject strings. The has been rewritten to avoid using
776 things like p++ in the argument of isxdigit().
777
778
779 Version 8.13 16-Aug-2011
780 ------------------------
781
782 1. The Unicode data tables have been updated to Unicode 6.0.0.
783
784 2. Two minor typos in pcre_internal.h have been fixed.
785
786 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
787 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
788 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
789
790 4. There were a number of related bugs in the code for matching backrefences
791 caselessly in UTF-8 mode when codes for the characters concerned were
792 different numbers of bytes. For example, U+023A and U+2C65 are an upper
793 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
794 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
795 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
796 2-byte code at the end of the subject (it thought there wasn't enough data
797 left).
798
799 5. Comprehensive information about what went wrong is now returned by
800 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
801 as the output vector has at least 2 elements. The offset of the start of
802 the failing character and a reason code are placed in the vector.
803
804 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
805 now returned is for the first byte of the failing character, instead of the
806 last byte inspected. This is an incompatible change, but I hope it is small
807 enough not to be a problem. It makes the returned offset consistent with
808 pcre_exec() and pcre_dfa_exec().
809
810 7. pcretest now gives a text phrase as well as the error number when
811 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
812 failure, the offset and reason code are output.
813
814 8. When \R was used with a maximizing quantifier it failed to skip backwards
815 over a \r\n pair if the subsequent match failed. Instead, it just skipped
816 back over a single character (\n). This seems wrong (because it treated the
817 two characters as a single entity when going forwards), conflicts with the
818 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
819 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
820 has been changed.
821
822 9. Some internal refactoring has changed the processing so that the handling
823 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
824 time (the PCRE_DOTALL option was changed this way some time ago: version
825 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
826 which was always a bit of a fudge. It also means that there is one less
827 argument for the match() function, which reduces its stack requirements
828 slightly. This change also fixes an incompatibility with Perl: the pattern
829 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
830
831 10. More internal refactoring has drastically reduced the number of recursive
832 calls to match() for possessively repeated groups such as (abc)++ when
833 using pcre_exec().
834
835 11. While implementing 10, a number of bugs in the handling of groups were
836 discovered and fixed:
837
838 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
839 (a|)*(?1) gave a compile-time internal error.
840 ((a|)+)+ did not notice that the outer group could match an empty string.
841 (^a|^)+ was not marked as anchored.
842 (.*a|.*)+ was not marked as matching at start or after a newline.
843
844 12. Yet more internal refactoring has removed another argument from the match()
845 function. Special calls to this function are now indicated by setting a
846 value in a variable in the "match data" data block.
847
848 13. Be more explicit in pcre_study() instead of relying on "default" for
849 opcodes that mean there is no starting character; this means that when new
850 ones are added and accidentally left out of pcre_study(), testing should
851 pick them up.
852
853 14. The -s option of pcretest has been documented for ages as being an old
854 synonym of -m (show memory usage). I have changed it to mean "force study
855 for every regex", that is, assume /S for every regex. This is similar to -i
856 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
857 using it. It makes it easier to run collections of tests with and without
858 study enabled, and thereby test pcre_study() more easily. All the standard
859 tests are now run with and without -s (but some patterns can be marked as
860 "never study" - see 20 below).
861
862 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
863 restoration of the capturing data to the outer values was not happening
864 correctly.
865
866 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
867 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
868 pattern had matched an empty string, and so incorrectly returned a no
869 match.
870
871 17. There was optimizing code for the last branch of non-capturing parentheses,
872 and also for the obeyed branch of a conditional subexpression, which used
873 tail recursion to cut down on stack usage. Unfortunately, now that there is
874 the possibility of (*THEN) occurring in these branches, tail recursion is
875 no longer possible because the return has to be checked for (*THEN). These
876 two optimizations have therefore been removed. [But see 8.20/11 above.]
877
878 18. If a pattern containing \R was studied, it was assumed that \R always
879 matched two bytes, thus causing the minimum subject length to be
880 incorrectly computed because \R can also match just one byte.
881
882 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
883 was incorrectly computed.
884
885 20. If /S is present twice on a test pattern in pcretest input, it now
886 *disables* studying, thereby overriding the use of -s on the command line
887 (see 14 above). This is necessary for one or two tests to keep the output
888 identical in both cases.
889
890 21. When (*ACCEPT) was used in an assertion that matched an empty string and
891 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
892
893 22. When an atomic group that contained a capturing parenthesis was
894 successfully matched, but the branch in which it appeared failed, the
895 capturing was not being forgotten if a higher numbered group was later
896 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
897 group 1 to "a", when in fact it should be unset. This applied to multi-
898 branched capturing and non-capturing groups, repeated or not, and also to
899 positive assertions (capturing in negative assertions does not happen
900 in PCRE) and also to nested atomic groups.
901
902 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
903 subject after a captured substring, to make it easier to tell which of a
904 number of identical substrings has been captured.
905
906 24. The way atomic groups are processed by pcre_exec() has been changed so that
907 if they are repeated, backtracking one repetition now resets captured
908 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
909 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
910 "aaa". Previously, it would have been "a". As part of this code
911 refactoring, the way recursive calls are handled has also been changed.
912
913 25. If an assertion condition captured any substrings, they were not passed
914 back unless some other capturing happened later. For example, if
915 (?(?=(a))a) was matched against "a", no capturing was returned.
916
917 26. When studying a pattern that contained subroutine calls or assertions,
918 the code for finding the minimum length of a possible match was handling
919 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
920 group 1 called group 2 while simultaneously a separate group 2 called group
921 1). A stack overflow occurred in this case. I have fixed this by limiting
922 the recursion depth to 10.
923
924 27. Updated RunTest.bat in the distribution to the version supplied by Tom
925 Fortmann. This supports explicit test numbers on the command line, and has
926 argument validation and error reporting.
927
928 28. An instance of \X with an unlimited repeat could fail if at any point the
929 first character it looked at was a mark character.
930
931 29. Some minor code refactoring concerning Unicode properties and scripts
932 should reduce the stack requirement of match() slightly.
933
934 30. Added the '=' option to pcretest to check the setting of unused capturing
935 slots at the end of the pattern, which are documented as being -1, but are
936 not included in the return count.
937
938 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
939 compiled something random. Now it gives a compile-time error (as does
940 Perl).
941
942 32. A *MARK encountered during the processing of a positive assertion is now
943 recorded and passed back (compatible with Perl).
944
945 33. If --only-matching or --colour was set on a pcregrep call whose pattern
946 had alternative anchored branches, the search for a second match in a line
947 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
948 matched the line "0102" twice. The same bug affected patterns that started
949 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
950 twice.
951
952 34. Previously, PCRE did not allow quantification of assertions. However, Perl
953 does, and because of capturing effects, quantifying parenthesized
954 assertions may at times be useful. Quantifiers are now allowed for
955 parenthesized assertions.
956
957 35. A minor code tidy in pcre_compile() when checking options for \R usage.
958
959 36. \g was being checked for fancy things in a character class, when it should
960 just be a literal "g".
961
962 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
963 appearance of a nested POSIX class supersedes an apparent external class.
964 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
965 unescaped square brackets may also appear as part of class names. For
966 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
967 more like Perl. (But see 8.20/1 above.)
968
969 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
970 was because it thought it was \N{name}, which is not supported).
971
972 39. Add minix to OS list not supporting the -S option in pcretest.
973
974 40. PCRE tries to detect cases of infinite recursion at compile time, but it
975 cannot analyze patterns in sufficient detail to catch mutual recursions
976 such as ((?1))((?2)). There is now a runtime test that gives an error if a
977 subgroup is called recursively as a subpattern for a second time at the
978 same position in the subject string. In previous releases this might have
979 been caught by the recursion limit, or it might have run out of stack.
980
981 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
982 happen only once. PCRE was, however incorrectly giving a compile time error
983 "recursive call could loop indefinitely" because it cannot analyze the
984 pattern in sufficient detail. The compile time test no longer happens when
985 PCRE is compiling a conditional subpattern, but actual runaway loops are
986 now caught at runtime (see 40 above).
987
988 42. It seems that Perl allows any characters other than a closing parenthesis
989 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
990 has been changed to be the same.
991
992 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
993 as not to get warnings when autogen.sh is called. Also changed
994 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
995
996 44. To help people who use pcregrep to scan files containing exceedingly long
997 lines, the following changes have been made:
998
999 (a) The default value of the buffer size parameter has been increased from
1000 8K to 20K. (The actual buffer used is three times this size.)
1001
1002 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1003 PCRE is built.
1004
1005 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1006 to be set at run time.
1007
1008 (d) Numerical values in pcregrep options can be followed by K or M, for
1009 example --buffer-size=50K.
1010
1011 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1012 given and the return code is set to 2.
1013
1014 45. Add a pointer to the latest mark to the callout data block.
1015
1016 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1017 partial match of an empty string instead of no match. This was specific to
1018 the use of ".".
1019
1020 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1021 complete match instead of a partial match. This bug was dependent on both
1022 the PCRE_UTF8 and PCRE_DOTALL options being set.
1023
1024 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1025 starting byte set, because \b was not being ignored.
1026
1027
1028 Version 8.12 15-Jan-2011
1029 ------------------------
1030
1031 1. Fixed some typos in the markup of the man pages, and wrote a script that
1032 checks for such things as part of the documentation building process.
1033
1034 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1035 --match-limit and --recursion-limit options (added for 8.11). In
1036 particular, this made one of the standard tests fail. (The integer value
1037 went into the wrong half of a long int.)
1038
1039 3. If the --colour option was given to pcregrep with -v (invert match), it
1040 did strange things, either producing crazy output, or crashing. It should,
1041 of course, ignore a request for colour when reporting lines that do not
1042 match.
1043
1044 4. Another pcregrep bug caused similar problems if --colour was specified with
1045 -M (multiline) and the pattern match finished with a line ending.
1046
1047 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1048 matched in multiline mode, the following line was shown as part of the
1049 match. This seems wrong, so I have changed it.
1050
1051 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1052 the check for further matches in the same line (so they could be coloured)
1053 to overrun the end of the current line. If another match was found, it was
1054 incorrectly shown (and then shown again when found in the next line).
1055
1056 7. If pcregrep was compiled under Windows, there was a reference to the
1057 function pcregrep_exit() before it was defined. I am assuming this was
1058 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1059 reported by a user. I've moved the definition above the reference.
1060
1061
1062 Version 8.11 10-Dec-2010
1063 ------------------------
1064
1065 1. (*THEN) was not working properly if there were untried alternatives prior
1066 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1067 backtracked to try for "b" instead of moving to the next alternative branch
1068 at the same level (in this case, to look for "c"). The Perl documentation
1069 is clear that when (*THEN) is backtracked onto, it goes to the "next
1070 alternative in the innermost enclosing group".
1071
1072 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1073 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1074 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1075 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1076 (*THEN).
1077
1078 3. If \s appeared in a character class, it removed the VT character from
1079 the class, even if it had been included by some previous item, for example
1080 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1081 of \s, but is part of the POSIX "space" class.)
1082
1083 4. A partial match never returns an empty string (because you can always
1084 match an empty string at the end of the subject); however the checking for
1085 an empty string was starting at the "start of match" point. This has been
1086 changed to the "earliest inspected character" point, because the returned
1087 data for a partial match starts at this character. This means that, for
1088 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1089 (previously it gave "no match").
1090
1091 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1092 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1093 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1094 has an implication that the given string is incomplete (because a partial
1095 match is preferred over a full match). For this reason, these items now
1096 give a partial match in this situation. [Aside: previously, the one case
1097 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1098 match rather than a full match, which was wrong by the old rules, but is
1099 now correct.]
1100
1101 6. There was a bug in the handling of #-introduced comments, recognized when
1102 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1103 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1104 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1105 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1106 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1107 places in pcre_compile().
1108
1109 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1110 comments when looking ahead for named forward references to subpatterns,
1111 the only newline sequence it recognized was NL. It now handles newlines
1112 according to the set newline convention.
1113
1114 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1115 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1116 cater for a lack of strerror(). These oversights have been fixed.
1117
1118 9. Added --match-limit and --recursion-limit to pcregrep.
1119
1120 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1121
1122 11. When the -o option was used, pcregrep was setting a return code of 1, even
1123 when matches were found, and --line-buffered was not being honoured.
1124
1125 12. Added an optional parentheses number to the -o and --only-matching options
1126 of pcregrep.
1127
1128 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1129 can match an empty string. The code to do it in pcretest and pcredemo
1130 needed fixing:
1131
1132 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1133 only one byte after an empty string match just before CRLF (this case
1134 just got forgotten; "any" and "anycrlf" were OK).
1135
1136 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1137 mode when an empty string match preceded an ASCII character followed by
1138 a non-ASCII character. (The code for advancing by one character rather
1139 than one byte was nonsense.)
1140
1141 (c) The pcredemo.c sample program did not have any code at all to handle
1142 the cases when CRLF is a valid newline sequence.
1143
1144 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1145 as a starting offset was within the subject string. There is now a new
1146 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1147 negative or greater than the length of the string. In order to test this,
1148 pcretest is extended to allow the setting of negative starting offsets.
1149
1150 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1151 starting offset points to the beginning of a UTF-8 character was
1152 unnecessarily clumsy. I tidied it up.
1153
1154 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1155 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1156
1157 17. Nobody had reported that the --include_dir option, which was added in
1158 release 7.7 should have been called --include-dir (hyphen, not underscore)
1159 for compatibility with GNU grep. I have changed it to --include-dir, but
1160 left --include_dir as an undocumented synonym, and the same for
1161 --exclude-dir, though that is not available in GNU grep, at least as of
1162 release 2.5.4.
1163
1164 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1165 characters from a string of bytes) have been redefined so as not to use
1166 loops, in order to improve performance in some environments. At the same
1167 time, I abstracted some of the common code into auxiliary macros to save
1168 repetition (this should not affect the compiled code).
1169
1170 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1171 compile-time error is now given if \c is not followed by an ASCII
1172 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1173 different, and any byte value is allowed.)
1174
1175 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1176 START_OPTIMIZE option, which is now allowed at compile time - but just
1177 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1178 to pcregrep and other applications that have no direct access to PCRE
1179 options. The new /Y option in pcretest sets this option when calling
1180 pcre_compile().
1181
1182 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1183 back references. Groups containing recursive back references were forced to
1184 be atomic by that change, but in the case of named groups, the amount of
1185 memory required was incorrectly computed, leading to "Failed: internal
1186 error: code overflow". This has been fixed.
1187
1188 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1189 pcretest.c, to avoid build problems in some Borland environments.
1190
1191
1192 Version 8.10 25-Jun-2010
1193 ------------------------
1194
1195 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1196 THEN.
1197
1198 2. (*ACCEPT) was not working when inside an atomic group.
1199
1200 3. Inside a character class, \B is treated as a literal by default, but
1201 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1202 causes the error). The code is unchanged, but I tidied the documentation.
1203
1204 4. Inside a character class, PCRE always treated \R and \X as literals,
1205 whereas Perl faults them if its -w option is set. I have changed PCRE so
1206 that it faults them when PCRE_EXTRA is set.
1207
1208 5. Added support for \N, which always matches any character other than
1209 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1210
1211 6. When compiling pcregrep with newer versions of gcc which may have
1212 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1213 declared with attribute warn_unused_result" were given. Just casting the
1214 result to (void) does not stop the warnings; a more elaborate fudge is
1215 needed. I've used a macro to implement this.
1216
1217 7. Minor change to pcretest.c to avoid a compiler warning.
1218
1219 8. Added four artifical Unicode properties to help with an option to make
1220 \s etc use properties (see next item). The new properties are: Xan
1221 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1222
1223 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1224 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1225 this option. Modified pcretest to add /W to test this facility. Added
1226 REG_UCP to make it available via the POSIX interface.
1227
1228 10. Added --line-buffered to pcregrep.
1229
1230 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1231 studied, and the match started with a letter with a code point greater than
1232 127 whose first byte was different to the first byte of the other case of
1233 the letter, the other case of this starting letter was not recognized
1234 (#976).
1235
1236 12. If a pattern that was studied started with a repeated Unicode property
1237 test, for example, \p{Nd}+, there was the theoretical possibility of
1238 setting up an incorrect bitmap of starting bytes, but fortunately it could
1239 not have actually happened in practice until change 8 above was made (it
1240 added property types that matched character-matching opcodes).
1241
1242 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1243 possible starting bytes for non-anchored patterns.
1244
1245 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1246 \R, and also a number of cases that involve Unicode properties, both
1247 explicit and implicit when PCRE_UCP is set.
1248
1249 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1250 input, it could crash or give wrong results if characters with values
1251 greater than 0xc0 were present in the subject string. (Detail: it assumed
1252 UTF-8 input when processing these items.)
1253
1254 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1255 size_t is 64-bit (#991).
1256
1257 17. Added a check for running out of memory when PCRE is compiled with
1258 --disable-stack-for-recursion (#990).
1259
1260 18. If the last data line in a file for pcretest does not have a newline on
1261 the end, a newline was missing in the output.
1262
1263 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1264 less than 128) in its various bitmaps. However, there is a facility for
1265 generating tables according to the current locale when PCRE is compiled. It
1266 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1267 characters, are recognized by isspace() and therefore were getting set in
1268 these tables, and indeed these tables seem to approximate to ISO 8859. This
1269 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1270 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1271 which of course cannot start UTF-8 characters. I have changed the code so
1272 that only real ASCII characters (less than 128) and the correct starting
1273 bytes for UTF-8 encodings are set for characters greater than 127 when in
1274 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1275 altogether.)
1276
1277 20. Added the /T option to pcretest so as to be able to run tests with non-
1278 standard character tables, thus making it possible to include the tests
1279 used for 19 above in the standard set of tests.
1280
1281 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1282 reference to a subpattern the other side of a comment that contains an
1283 opening parenthesis caused either an internal compiling error, or a
1284 reference to the wrong subpattern.
1285
1286
1287 Version 8.02 19-Mar-2010
1288 ------------------------
1289
1290 1. The Unicode data tables have been updated to Unicode 5.2.0.
1291
1292 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1293 configured.
1294
1295 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1296 original author of that file, following a query about its status.
1297
1298 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1299 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1300
1301 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1302 quantifier applied to a forward-referencing subroutine call, could compile
1303 incorrect code or give the error "internal error: previously-checked
1304 referenced subpattern not found".
1305
1306 6. Both MS Visual Studio and Symbian OS have problems with initializing
1307 variables to point to external functions. For these systems, therefore,
1308 pcre_malloc etc. are now initialized to local functions that call the
1309 relevant global functions.
1310
1311 7. There were two entries missing in the vectors called coptable and poptable
1312 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1313 I've fixed the data, and added a kludgy way of testing at compile time that
1314 the lengths are correct (equal to the number of opcodes).
1315
1316 8. Following on from 7, I added a similar kludge to check the length of the
1317 eint vector in pcreposix.c.
1318
1319 9. Error texts for pcre_compile() are held as one long string to avoid too
1320 much relocation at load time. To find a text, the string is searched,
1321 counting zeros. There was no check for running off the end of the string,
1322 which could happen if a new error number was added without updating the
1323 string.
1324
1325 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1326
1327 11. \K was not working if it appeared in an atomic group or in a group that
1328 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1329 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1330 the assertion is positive, but not if it is negative.
1331
1332 12. Change 11 fortuitously reduced the size of the stack frame used in the
1333 "match()" function of pcre_exec.c by one pointer. Forthcoming
1334 implementation of support for (*MARK) will need an extra pointer on the
1335 stack; I have reserved it now, so that the stack frame size does not
1336 decrease.
1337
1338 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1339 item in branch that calls a recursion is a subroutine call - as in the
1340 second branch in the above example - was incorrectly given the compile-
1341 time error "recursive call could loop indefinitely" because pcre_compile()
1342 was not correctly checking the subroutine for matching a non-empty string.
1343
1344 14. The checks for overrunning compiling workspace could trigger after an
1345 overrun had occurred. This is a "should never occur" error, but it can be
1346 triggered by pathological patterns such as hundreds of nested parentheses.
1347 The checks now trigger 100 bytes before the end of the workspace.
1348
1349 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1350
1351
1352 Version 8.01 19-Jan-2010
1353 ------------------------
1354
1355 1. If a pattern contained a conditional subpattern with only one branch (in
1356 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1357 computed the wrong minimum data length (which is of course zero for such
1358 subpatterns). This could cause incorrect "no match" results.
1359
1360 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1361 the pattern is reset in the first branch, pcre_compile() failed with
1362 "internal error: code overflow at offset...". This happened only when
1363 the reset was to the original external option setting. (An optimization
1364 abstracts leading options settings into an external setting, which was the
1365 cause of this.)
1366
1367 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1368 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1369 assertion pattern did not match (meaning that the assertion was true), it
1370 was incorrectly treated as false if the SKIP had been reached during the
1371 matching. This also applied to assertions used as conditions.
1372
1373 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1374 assertion subpattern, including such a pattern used as a condition,
1375 unpredictable results occurred, instead of the error return
1376 PCRE_ERROR_DFA_UITEM.
1377
1378 5. The C++ GlobalReplace function was not working like Perl for the special
1379 situation when an empty string is matched. It now does the fancy magic
1380 stuff that is necessary.
1381
1382 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1383 removed. (These were left over from very, very early versions of PCRE.)
1384
1385 7. Some cosmetic changes to the code to make life easier when compiling it
1386 as part of something else:
1387
1388 (a) Change DEBUG to PCRE_DEBUG.
1389
1390 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1391 called "current" as "current_branch", to prevent a collision with the
1392 Linux macro when compiled as a kernel module.
1393
1394 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1395 prevent a collision with the Linux macro when compiled as a kernel
1396 module.
1397
1398 8. In pcre_compile() there are some checks for integer overflows that used to
1399 cast potentially large values to (double). This has been changed to that
1400 when building, a check for int64_t is made, and if it is found, it is used
1401 instead, thus avoiding the use of floating point arithmetic. (There is no
1402 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1403 double.
1404
1405 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1406 2005 (difference between two addresses compared to an unsigned value).
1407
1408 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1409 custom one, because of the following reported problem in Windows:
1410
1411 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1412 under Win32.
1413 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1414 therefore missing the function definition.
1415 - The compiler thus generates a "C" signature for the test function.
1416 - The linker fails to find the "C" function.
1417 - PCRE fails to configure if asked to do so against libbz2.
1418
1419 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1420 messages were output:
1421
1422 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1423 rerunning libtoolize, to keep the correct libtool macros in-tree.
1424 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1425
1426 I have done both of these things.
1427
1428 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1429 most of the time, it *can* run out if it is given a pattern that contains a
1430 runaway infinite recursion. I updated the discussion in the pcrestack man
1431 page.
1432
1433 13. Now that we have gone to the x.xx style of version numbers, the minor
1434 version may start with zero. Using 08 or 09 is a bad idea because users
1435 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1436 interpreted as invalid octal numbers. I've updated the previous comment in
1437 configure.ac, and also added a check that gives an error if 08 or 09 are
1438 used.
1439
1440 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1441 causing partial matching to fail when the end of the subject matched \W
1442 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1443
1444 15. There were some discrepancies between the declarations in pcre_internal.h
1445 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1446 their definitions. The declarations used "const uschar *" and the
1447 definitions used USPTR. Even though USPTR is normally defined as "const
1448 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1449 reported that: "This difference in casting confuses some C++ compilers, for
1450 example, SunCC recognizes above declarations as different functions and
1451 generates broken code for hbpcre." I have changed the declarations to use
1452 USPTR.
1453
1454 16. GNU libtool is named differently on some systems. The autogen.sh script now
1455 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1456 (FreeBSD).
1457
1458 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1459 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1460 comment: "Figure out how to create a longlong from a string: strtoll and
1461 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1462 instance, but it only takes 2 args instead of 3!"
1463
1464 18. A subtle bug concerned with back references has been fixed by a change of
1465 specification, with a corresponding code fix. A pattern such as
1466 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1467 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1468 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1469 same bug. Such groups have to be quantified to be useful, or contained
1470 inside another quantified group. (If there's no repetition, the reference
1471 can never match.) The problem arises because, having left the group and
1472 moved on to the rest of the pattern, a later failure that backtracks into
1473 the group uses the captured value from the final iteration of the group
1474 rather than the correct earlier one. I have fixed this in PCRE by forcing
1475 any group that contains a reference to itself to be an atomic group; that
1476 is, there cannot be any backtracking into it once it has completed. This is
1477 similar to recursive and subroutine calls.
1478
1479
1480 Version 8.00 19-Oct-09
1481 ----------------------
1482
1483 1. The table for translating pcre_compile() error codes into POSIX error codes
1484 was out-of-date, and there was no check on the pcre_compile() error code
1485 being within the table. This could lead to an OK return being given in
1486 error.
1487
1488 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1489 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1490 in a Windows environment.
1491
1492 3. The pcregrep --count option prints the count for each file even when it is
1493 zero, as does GNU grep. However, pcregrep was also printing all files when
1494 --files-with-matches was added. Now, when both options are given, it prints
1495 counts only for those files that have at least one match. (GNU grep just
1496 prints the file name in this circumstance, but including the count seems
1497 more useful - otherwise, why use --count?) Also ensured that the
1498 combination -clh just lists non-zero counts, with no names.
1499
1500 4. The long form of the pcregrep -F option was incorrectly implemented as
1501 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1502 but it seems right to fix it, and I didn't think it was worth preserving
1503 the old behaviour.
1504
1505 5. The command line items --regex=pattern and --regexp=pattern were not
1506 recognized by pcregrep, which required --regex pattern or --regexp pattern
1507 (with a space rather than an '='). The man page documented the '=' forms,
1508 which are compatible with GNU grep; these now work.
1509
1510 6. No libpcreposix.pc file was created for pkg-config; there was just
1511 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1512
1513 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1514 when UCP support is not needed, by modifying the Python script that
1515 generates it from Unicode data files. This should not matter if the module
1516 is correctly used as a library, but I received one complaint about 50K of
1517 unwanted data. My guess is that the person linked everything into his
1518 program rather than using a library. Anyway, it does no harm.
1519
1520 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1521 was a minimum greater than 1 for a wide character in a possessive
1522 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1523 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1524 character. Chaos in the form of incorrect output or a compiling loop could
1525 result.
1526
1527 9. The restrictions on what a pattern can contain when partial matching is
1528 requested for pcre_exec() have been removed. All patterns can now be
1529 partially matched by this function. In addition, if there are at least two
1530 slots in the offset vector, the offset of the earliest inspected character
1531 for the match and the offset of the end of the subject are set in them when
1532 PCRE_ERROR_PARTIAL is returned.
1533
1534 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1535 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1536 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1537 and may be more useful for multi-segment matching.
1538
1539 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1540 used to be given if ever the end of the subject was reached; now it is
1541 given only if matching could not proceed because another character was
1542 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1543 string "Z", which now yields "no match" instead of "partial match". In the
1544 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1545 final character ended with (*FAIL).
1546
1547 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1548 if the pattern had a "must contain" character that was already found in the
1549 earlier partial match, unless partial matching was again requested. For
1550 example, with the pattern /dog.(body)?/, the "must contain" character is
1551 "g". If the first part-match was for the string "dog", restarting with
1552 "sbody" failed. This bug has been fixed.
1553
1554 13. The string returned by pcre_dfa_exec() after a partial match has been
1555 changed so that it starts at the first inspected character rather than the
1556 first character of the match. This makes a difference only if the pattern
1557 starts with a lookbehind assertion or \b or \B (\K is not supported by
1558 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1559 matching functions compatible, and I think it's the right thing to do.
1560
1561 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1562 so that the demonstration program is easily available in environments where
1563 PCRE has not been installed from source.
1564
1565 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1566 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1567 library.
1568
1569 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1570 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1571 is not the first non-POSIX option to be added. Clearly some people find
1572 these options useful.
1573
1574 17. If a caller to the POSIX matching function regexec() passes a non-zero
1575 value for nmatch with a NULL value for pmatch, the value of
1576 nmatch is forced to zero.
1577
1578 18. RunGrepTest did not have a test for the availability of the -u option of
1579 the diff command, as RunTest does. It now checks in the same way as
1580 RunTest, and also checks for the -b option.
1581
1582 19. If an odd number of negated classes containing just a single character
1583 interposed, within parentheses, between a forward reference to a named
1584 subpattern and the definition of the subpattern, compilation crashed with
1585 an internal error, complaining that it could not find the referenced
1586 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1587 [The bug was that it was starting one character too far in when skipping
1588 over the character class, thus treating the ] as data rather than
1589 terminating the class. This meant it could skip too much.]
1590
1591 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1592 /g option in pcretest when the pattern contains \K, which makes it possible
1593 to have an empty string match not at the start, even when the pattern is
1594 anchored. Updated pcretest and pcredemo to use this option.
1595
1596 21. If the maximum number of capturing subpatterns in a recursion was greater
1597 than the maximum at the outer level, the higher number was returned, but
1598 with unset values at the outer level. The correct (outer level) value is
1599 now given.
1600
1601 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1602 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1603 make it do so. The string so far is captured, making this feature
1604 compatible with Perl.
1605
1606 23. The tests have been re-organized, adding tests 11 and 12, to make it
1607 possible to check the Perl 5.10 features against Perl 5.10.
1608
1609 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1610 pattern matches a fixed length string. PCRE did not allow this; now it
1611 does. Neither allows recursion.
1612
1613 25. I finally figured out how to implement a request to provide the minimum
1614 length of subject string that was needed in order to match a given pattern.
1615 (It was back references and recursion that I had previously got hung up
1616 on.) This code has now been added to pcre_study(); it finds a lower bound
1617 to the length of subject needed. It is not necessarily the greatest lower
1618 bound, but using it to avoid searching strings that are too short does give
1619 some useful speed-ups. The value is available to calling programs via
1620 pcre_fullinfo().
1621
1622 26. While implementing 25, I discovered to my embarrassment that pcretest had
1623 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1624 study optimizations had never been tested with that matching function.
1625 Oops. What is worse, even when it was passed study data, there was a bug in
1626 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1627 were also very few tests of studied patterns with pcre_dfa_exec().
1628
1629 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1630 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1631 on the other side of the coin, they are no longer allowed to have different
1632 names, because these cannot be distinguished in PCRE, and this has caused
1633 confusion. (This is a difference from Perl.)
1634
1635 28. When duplicate subpattern names are present (necessarily with different
1636 numbers, as required by 27 above), and a test is made by name in a
1637 conditional pattern, either for a subpattern having been matched, or for
1638 recursion in such a pattern, all the associated numbered subpatterns are
1639 tested, and the overall condition is true if the condition is true for any
1640 one of them. This is the way Perl works, and is also more like the way
1641 testing by number works.
1642
1643
1644 Version 7.9 11-Apr-09
1645 ---------------------
1646
1647 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1648 (pcretest), all targets were linked against these libraries. This included
1649 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1650 libraries. This caused unwanted dependencies to be created. This problem
1651 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1652 pcretest is linked with readline.
1653
1654 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1655 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1656 moved outside it again, because FALSE and TRUE are already defined in AIX,
1657 but BOOL is not.
1658
1659 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1660 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1661
1662 4. The pcregrep documentation said spaces were inserted as well as colons (or
1663 hyphens) following file names and line numbers when outputting matching
1664 lines. This is not true; no spaces are inserted. I have also clarified the
1665 wording for the --colour (or --color) option.
1666
1667 5. In pcregrep, when --colour was used with -o, the list of matching strings
1668 was not coloured; this is different to GNU grep, so I have changed it to be
1669 the same.
1670
1671 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1672 each matching line was coloured. Now it goes on to look for further matches
1673 of any of the test patterns, which is the same behaviour as GNU grep.
1674
1675 7. A pattern that could match an empty string could cause pcregrep to loop; it
1676 doesn't make sense to accept an empty string match in pcregrep, so I have
1677 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1678 seems to be how GNU grep behaves. [But see later change 40 for release
1679 8.33.]
1680
1681 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1682 start or after a newline", because the conditional assertion was not being
1683 correctly handled. The rule now is that both the assertion and what follows
1684 in the first alternative must satisfy the test.
1685
1686 9. If auto-callout was enabled in a pattern with a conditional group whose
1687 condition was an assertion, PCRE could crash during matching, both with
1688 pcre_exec() and pcre_dfa_exec().
1689
1690 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1691 used for matching.
1692
1693 11. Unicode property support in character classes was not working for
1694 characters (bytes) greater than 127 when not in UTF-8 mode.
1695
1696 12. Added the -M command line option to pcretest.
1697
1698 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1699
1700 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1701
1702 16. Added comments and documentation about mis-use of no_arg in the C++
1703 wrapper.
1704
1705 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1706 from Martin Jerabek that uses macro names for all relevant character and
1707 string constants.
1708
1709 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1710 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1711 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1712 these, but not everybody uses configure.
1713
1714 19. A conditional group that had only one branch was not being correctly
1715 recognized as an item that could match an empty string. This meant that an
1716 enclosing group might also not be so recognized, causing infinite looping
1717 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1718 with the subject "ab", where knowledge that the repeated group can match
1719 nothing is needed in order to break the loop.
1720
1721 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1722 exec(), but without supplying a callout function, matching went wrong.
1723
1724 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1725 leak if the size of the offset vector was greater than 30. When the vector
1726 is smaller, the saved offsets during recursion go onto a local stack
1727 vector, but for larger vectors malloc() is used. It was failing to free
1728 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1729 error, in fact).
1730
1731 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1732 heapframe that is used only when UTF-8 support is enabled. This caused no
1733 problem, but was untidy.
1734
1735 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1736 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1737 included within another project.
1738
1739 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1740 slightly modified by me:
1741
1742 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1743 not building pcregrep.
1744
1745 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1746 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1747
1748 25. Forward references, both numeric and by name, in patterns that made use of
1749 duplicate group numbers, could behave incorrectly or give incorrect errors,
1750 because when scanning forward to find the reference group, PCRE was not
1751 taking into account the duplicate group numbers. A pattern such as
1752 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1753
1754 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1755 the feature of a custom pointer more persuasive (as requested by a user).
1756
1757 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1758 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1759 pre-defined.
1760
1761 28. Added support for (*UTF8) at the start of a pattern.
1762
1763 29. Arrange for flags added by the "release type" setting in CMake to be shown
1764 in the configuration summary.
1765
1766
1767 Version 7.8 05-Sep-08
1768 ---------------------
1769
1770 1. Replaced UCP searching code with optimized version as implemented for Ad
1771 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1772 stage table and inline lookup instead of a function, giving speed ups of 2
1773 to 5 times on some simple patterns that I tested. Permission was given to
1774 distribute the MultiStage2.py script that generates the tables (it's not in
1775 the tarball, but is in the Subversion repository).
1776
1777 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1778 scripts.
1779
1780 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1781 a group with a zero qualifier. The result of the study could be incorrect,
1782 or the function might crash, depending on the pattern.
1783
1784 4. Caseless matching was not working for non-ASCII characters in back
1785 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1786 It now works when Unicode Property Support is available.
1787
1788 5. In pcretest, an escape such as \x{de} in the data was always generating
1789 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1790 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1791 truncation.
1792
1793 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1794
1795 7. Added two (int) casts to pcregrep when printing the difference of two
1796 pointers, in case they are 64-bit values.
1797
1798 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1799 test 2 if it fails.
1800
1801 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1802 and a #define of that name to empty if it is not externally set. This is to
1803 allow users of MSVC to set it if necessary.
1804
1805 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1806 the convenience functions in the pcre_get.c source file.
1807
1808 11. An option change at the start of a pattern that had top-level alternatives
1809 could cause overwriting and/or a crash. This command provoked a crash in
1810 some environments:
1811
1812 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1813
1814 This potential security problem was recorded as CVE-2008-2371.
1815
1816 12. For a pattern where the match had to start at the beginning or immediately
1817 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1818 pcre_dfa_exec() could read past the end of the passed subject if there was
1819 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1820 pcretest so that it places the subject at the end of its malloc-ed buffer.
1821
1822 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1823 exec() might read past the end of the data buffer in UTF-8 mode.
1824
1825 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1826 the data contained the byte 0x85 as part of a UTF-8 character within its
1827 first line. This applied both to normal and DFA matching.
1828
1829 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1830 /^[^d]*?$/8 failed to match "abc".
1831
1832 16. Added a missing copyright notice to pcrecpp_internal.h.
1833
1834 17. Make it more clear in the documentation that values returned from
1835 pcre_exec() in ovector are byte offsets, not character counts.
1836
1837 18. Tidied a few places to stop certain compilers from issuing warnings.
1838
1839 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1840 supplied by Stefan Weber. I made a further small update for 7.8 because
1841 there is a change of source arrangements: the pcre_searchfuncs.c module is
1842 replaced by pcre_ucd.c.
1843
1844
1845 Version 7.7 07-May-08
1846 ---------------------
1847
1848 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1849 a string to a long long, pretend we don't even have a long long." This is
1850 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1851
1852 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1853 pre-7.6 versions, which defined a global no_arg variable instead of putting
1854 it in the RE class. (See also #8 below.)
1855
1856 3. Remove a line of dead code, identified by coverity and reported by Nuno
1857 Lopes.
1858
1859 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1860
1861 (1) The include/exclude patterns were being applied to the whole pathnames
1862 of files, instead of just to the final components.
1863
1864 (2) If there was more than one level of directory, the subdirectories were
1865 skipped unless they satisfied the include/exclude conditions. This is
1866 inconsistent with GNU grep (and could even be seen as contrary to the
1867 pcregrep specification - which I improved to make it absolutely clear).
1868 The action now is always to scan all levels of directory, and just
1869 apply the include/exclude patterns to regular files.
1870
1871 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1872 --exclude_dir in the tests to avoid scanning .svn directories.
1873
1874 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1875 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1876 doesn't support NULs in patterns.
1877
1878 7. Added some missing "const"s to declarations of static tables in
1879 pcre_compile.c and pcre_dfa_exec.c.
1880
1881 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1882 caused by fix #2 above. (Subsequently also a second patch to fix the
1883 first patch. And a third patch - this was a messy problem.)
1884
1885 9. Applied Craig's patch to remove the use of push_back().
1886
1887 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1888 matching function regexec().
1889
1890 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1891 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1892 references. PCRE supports relative numbers with this syntax (I don't think
1893 Oniguruma does).
1894
1895 12. Previously, a group with a zero repeat such as (...){0} was completely
1896 omitted from the compiled regex. However, this means that if the group
1897 was called as a subroutine from elsewhere in the pattern, things went wrong
1898 (an internal error was given). Such groups are now left in the compiled
1899 pattern, with a new opcode that causes them to be skipped at execution
1900 time.
1901
1902 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1903 to the way PCRE behaves:
1904
1905 (a) A lone ] character is dis-allowed (Perl treats it as data).
1906
1907 (b) A back reference to an unmatched subpattern matches an empty string
1908 (Perl fails the current match path).
1909
1910 (c) A data ] in a character class must be notated as \] because if the
1911 first data character in a class is ], it defines an empty class. (In
1912 Perl it is not possible to have an empty class.) The empty class []
1913 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1914 The negative empty class [^] matches any one character, independently
1915 of the DOTALL setting.
1916
1917 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1918 non-existent subpattern following a character class starting with ']' and
1919 containing () gave an internal compiling error instead of "reference to
1920 non-existent subpattern". Fortunately, when the pattern did exist, the
1921 compiled code was correct. (When scanning forwards to check for the
1922 existence of the subpattern, it was treating the data ']' as terminating
1923 the class, so got the count wrong. When actually compiling, the reference
1924 was subsequently set up correctly.)
1925
1926 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1927 it was being rejected as not supported by pcre_dfa_exec(), even though
1928 other assertions are supported. I have made pcre_dfa_exec() support
1929 (*FAIL).
1930
1931 16. The implementation of 13c above involved the invention of a new opcode,
1932 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1933 cannot be changed at match time, I realized I could make a small
1934 improvement to matching performance by compiling OP_ALLANY instead of
1935 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1936 on the OP_ANY path.
1937
1938 17. Compiling pcretest on Windows with readline support failed without the
1939 following two fixes: (1) Make the unistd.h include conditional on
1940 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1941
1942 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1943 ncurses library to be included for pcretest when ReadLine support is
1944 requested, but also to allow for it to be overridden. This patch came from
1945 Daniel Bergström.
1946
1947 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1948 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1949 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1950 spotting this.
1951
1952
1953 Version 7.6 28-Jan-08
1954 ---------------------
1955
1956 1. A character class containing a very large number of characters with
1957 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1958 overflow.
1959
1960 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1961 HAVE_LONG_LONG is not defined.
1962
1963 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1964 bring them up to date and include new features. This patch includes:
1965
1966 - Fixed PH's badly added libz and libbz2 support.
1967 - Fixed a problem with static linking.
1968 - Added pcredemo. [But later removed - see 7 below.]
1969 - Fixed dftables problem and added an option.
1970 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1971 HAVE_LONG_LONG.
1972 - Added readline support for pcretest.
1973 - Added an listing of the option settings after cmake has run.
1974
1975 4. A user submitted a patch to Makefile that makes it easy to create
1976 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1977 Makefile.am that cause it to include this special target, without
1978 affecting anything else. Note that the same mingw target plus all
1979 the other distribution libraries and programs are now supported
1980 when configuring with CMake (see 6 below) instead of with
1981 Configure/Make.
1982
1983 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1984 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1985 exported in the Windows port". It has not yet been confirmed that the patch
1986 solves the problem, but it does no harm.
1987
1988 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1989 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1990 with CMake, and also correct the comment about stack recursion.
1991
1992 7. Remove the automatic building of pcredemo from the ./configure system and
1993 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1994 of a program that users should build themselves after PCRE is installed, so
1995 building it automatically is not really right. What is more, it gave
1996 trouble in some build environments.
1997
1998 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1999
2000
2001 Version 7.5 10-Jan-08
2002 ---------------------
2003
2004 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2005 values in parens when parsing an RE using the C++ wrapper."
2006
2007 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2008 Characters greater than 255 were excluded from the class instead of being
2009 included.
2010
2011 3. The same bug as (2) above applied to negated POSIX classes such as
2012 [:^space:].
2013
2014 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2015 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2016 I have changed it.
2017
2018 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2019 first named subpattern) and a construct such as (?&a) would reference the
2020 first named subpattern whose name started with "a" (in other words, the
2021 length check was missing). Both these problems are fixed. "Subpattern name
2022 expected" is now given for (?&) (a zero-length name), and this patch also
2023 makes it give the same error for \k'' (previously it complained that that
2024 was a reference to a non-existent subpattern).
2025
2026 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2027 this is right because (?- can be followed by option settings as well as by
2028 digits. I have, however, made the messages clearer.
2029
2030 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2031 than the number used in the conditional) now cause a compile-time error.
2032 This is actually not compatible with Perl, which accepts such patterns, but
2033 treats the conditional as always being FALSE (as PCRE used to), but it
2034 seems to me that giving a diagnostic is better.
2035
2036 8. Change "alphameric" to the more common word "alphanumeric" in comments
2037 and messages.
2038
2039 9. Fix two occurrences of "backslash" in comments that should have been
2040 "backspace".
2041
2042 10. Remove two redundant lines of code that can never be obeyed (their function
2043 was moved elsewhere).
2044
2045 11. The program that makes PCRE's Unicode character property table had a bug
2046 which caused it to generate incorrect table entries for sequences of
2047 characters that have the same character type, but are in different scripts.
2048 It amalgamated them into a single range, with the script of the first of
2049 them. In other words, some characters were in the wrong script. There were
2050 thirteen such cases, affecting characters in the following ranges:
2051
2052 U+002b0 - U+002c1
2053 U+0060c - U+0060d
2054 U+0061e - U+00612
2055 U+0064b - U+0065e
2056 U+0074d - U+0076d
2057 U+01800 - U+01805
2058 U+01d00 - U+01d77
2059 U+01d9b - U+01dbf
2060 U+0200b - U+0200f
2061 U+030fc - U+030fe
2062 U+03260 - U+0327f
2063 U+0fb46 - U+0fbb1
2064 U+10450 - U+1049d
2065
2066 12. The -o option (show only the matching part of a line) for pcregrep was not
2067 compatible with GNU grep in that, if there was more than one match in a
2068 line, it showed only the first of them. It now behaves in the same way as
2069 GNU grep.
2070
2071 13. If the -o and -v options were combined for pcregrep, it printed a blank
2072 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2073 does the same. The return code can be used to tell if there were any
2074 non-matching lines.
2075
2076 14. Added --file-offsets and --line-offsets to pcregrep.
2077
2078 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2079 infinitely looping recursion. The bug was that positive lookaheads were not
2080 being skipped when checking for a possible empty match (negative lookaheads
2081 and both kinds of lookbehind were skipped).
2082
2083 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2084 inclusion of <windows.h> to before rather than after the definition of
2085 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2086
2087 17. Specifying a possessive quantifier with a specific limit for a Unicode
2088 character property caused pcre_compile() to compile bad code, which led at
2089 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2090 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2091 caused the error; without that there was no problem.
2092
2093 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2094
2095 19. Added --enable-pcretest-libreadline.
2096
2097 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2098 RE::GlobalReplace(). As a result, the number of replacements returned was
2099 double what it should be. I removed one of the increments, but Craig sent a
2100 later patch that removed the other one (the right fix) and added unit tests
2101 that check the return values (which was not done before).
2102
2103 21. Several CMake things:
2104
2105 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2106 the names libpcre and libpcreposix, not just pcre and pcreposix.
2107
2108 (2) The above change means that pcretest and pcregrep are now correctly
2109 linked with the newly-built libraries, not previously installed ones.
2110
2111 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2112
2113 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2114 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2115 UTF-8 newline character). The key issue is that the pattern starts .*;
2116 this means that the match must be either at the beginning, or after a
2117 newline. The bug was in the code for advancing after a failed match and
2118 checking that the new position followed a newline. It was not taking
2119 account of UTF-8 characters correctly.
2120
2121 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2122 character classes. PCRE was not treating the sequence [:...:] as a
2123 character class unless the ... were all letters. Perl, however, seems to
2124 allow any characters between [: and :], though of course it rejects as
2125 unknown any "names" that contain non-letters, because all the known class
2126 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2127 for example, whereas PCRE did not - it did not recognize a POSIX character
2128 class. This seemed a bit dangerous, so the code has been changed to be
2129 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2130 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2131 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2132 Perl does, and where it didn't before.
2133
2134 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2135 Windows environments %n is disabled by default.
2136
2137
2138 Version 7.4 21-Sep-07
2139 ---------------------
2140
2141 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2142 means that a class such as [\s] counted as "explicit reference to CR or
2143 LF". That isn't really right - the whole point of the change was to try to
2144 help when there was an actual mention of one of the two characters. So now
2145 the change happens only if \r or \n (or a literal CR or LF) character is
2146 encountered.
2147
2148 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2149 of both had grown to the point where there were only 3 bits left.
2150 Fortunately, there was spare space in the data structure, and so I have
2151 moved the internal flags into a new 16-bit field to free up more option
2152 bits.
2153
2154 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2155 but did not set the internal JCHANGED flag - either of these is enough to
2156 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2157 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2158 start sets both bits.
2159
2160 4. Added options (at build time, compile time, exec time) to change \R from
2161 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2162
2163 5. doc/pcresyntax.html was missing from the distribution.
2164
2165 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2166 compatibility, even though it is no longer used.
2167
2168 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2169 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2170 windows.h file is present (where different names are used). [This was
2171 reversed later after testing - see 16 below.]
2172
2173 8. Changed all #include <config.h> to #include "config.h". There were also
2174 some further <pcre.h> cases that I changed to "pcre.h".
2175
2176 9. When pcregrep was used with the --colour option, it missed the line ending
2177 sequence off the lines that it output.
2178
2179 10. It was pointed out to me that arrays of string pointers cause lots of
2180 relocations when a shared library is dynamically loaded. A technique of
2181 using a single long string with a table of offsets can drastically reduce
2182 these. I have refactored PCRE in four places to do this. The result is
2183 dramatic:
2184
2185 Originally: 290
2186 After changing UCP table: 187
2187 After changing error message table: 43
2188 After changing table of "verbs" 36
2189 After changing table of Posix names 22
2190
2191 Thanks to the folks working on Gregex for glib for this insight.
2192
2193 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2194 unicode-properties was also set.
2195
2196 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2197
2198 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2199 checked only for CRLF.
2200
2201 14. Added casts to pcretest.c to avoid compiler warnings.
2202
2203 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2204
2205 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2206 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2207 entirely. This removes changes made in 7 above.
2208
2209 17. The CMake files have been updated, and there is now more information about
2210 building with CMake in the NON-UNIX-USE document.
2211
2212
2213 Version 7.3 28-Aug-07
2214 ---------------------
2215
2216 1. In the rejigging of the build system that eventually resulted in 7.1, the
2217 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2218 brackets there is not right, since it causes compilers to look for an
2219 installed pcre.h, not the version that is in the source that is being
2220 compiled (which of course may be different). I have changed it back to:
2221
2222 #include "pcre.h"
2223
2224 I have a vague recollection that the change was concerned with compiling in
2225 different directories, but in the new build system, that is taken care of
2226 by the VPATH setting the Makefile.
2227
2228 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2229 when the subject happened to end in the byte 0x85 (e.g. if the last
2230 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2231 characters but of course it shouldn't be taken as a newline when it is part
2232 of another character. The bug was that, for an unlimited repeat of . in
2233 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2234 characters when looking for a newline.
2235
2236 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2237
2238 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2239 in debug output.
2240
2241 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2242 long printing in the pcrecpp unittest when running under MinGW.
2243
2244 6. ESC_K was left out of the EBCDIC table.
2245
2246 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2247 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2248 limit also applies to "virtual nesting" when a pattern is recursive, and in
2249 this case 1000 isn't so big. I have been able to remove this limit at the
2250 expense of backing off one optimization in certain circumstances. Normally,
2251 when pcre_exec() would call its internal match() function recursively and
2252 immediately return the result unconditionally, it uses a "tail recursion"
2253 feature to save stack. However, when a subpattern that can match an empty
2254 string has an unlimited repetition quantifier, it no longer makes this
2255 optimization. That gives it a stack frame in which to save the data for
2256 checking that an empty string has been matched. Previously this was taken
2257 from the 1000-entry workspace that had been reserved. So now there is no
2258 explicit limit, but more stack is used.
2259
2260 8. Applied Daniel's patches to solve problems with the import/export magic
2261 syntax that is required for Windows, and which was going wrong for the
2262 pcreposix and pcrecpp parts of the library. These were overlooked when this
2263 problem was solved for the main library.
2264
2265 9. There were some crude static tests to avoid integer overflow when computing
2266 the size of patterns that contain repeated groups with explicit upper
2267 limits. As the maximum quantifier is 65535, the maximum group length was
2268 set at 30,000 so that the product of these two numbers did not overflow a
2269 32-bit integer. However, it turns out that people want to use groups that
2270 are longer than 30,000 bytes (though not repeat them that many times).
2271 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2272 made it possible to implement the integer overflow checks in a much more
2273 dynamic way, which I have now done. The artificial limitation on group
2274 length has been removed - we now have only the limit on the total length of
2275 the compiled pattern, which depends on the LINK_SIZE setting.
2276
2277 10. Fixed a bug in the documentation for get/copy named substring when
2278 duplicate names are permitted. If none of the named substrings are set, the
2279 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2280 empty string.
2281
2282 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2283 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2284 because the ] is interpreted as the first data character and the
2285 terminating ] is not found. PCRE has been made compatible with Perl in this
2286 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2287 cause memory overwriting.
2288
2289 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2290 string has been matched (to stop an infinite loop). It was not recognizing
2291 a conditional subpattern that could match an empty string if that
2292 subpattern was within another subpattern. For example, it looped when
2293 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2294 condition was not nested. This bug has been fixed.
2295
2296 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2297 past the start of the subject in the presence of bytes with the top bit
2298 set, for example "\x8aBCD".
2299
2300 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2301 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2302
2303 14. Optimized (?!) to (*FAIL).
2304
2305 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2306 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2307 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2308 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2309 does: it's just the validity check that is more restrictive.
2310
2311 16. Inserted checks for integer overflows during escape sequence (backslash)
2312 processing, and also fixed erroneous offset values for syntax errors during
2313 backslash processing.
2314
2315 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2316 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2317
2318 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2319 caused an overrun.
2320
2321 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2322 something other than just ASCII characters) inside a group that had an
2323 unlimited repeat caused a loop at compile time (while checking to see
2324 whether the group could match an empty string).
2325
2326 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2327 [\P{Any}] did so. (Error in the code for printing property names.)
2328
2329 21. An orphan \E inside a character class could cause a crash.
2330
2331 22. A repeated capturing bracket such as (A)? could cause a wild memory
2332 reference during compilation.
2333
2334 23. There are several functions in pcre_compile() that scan along a compiled
2335 expression for various reasons (e.g. to see if it's fixed length for look
2336 behind). There were bugs in these functions when a repeated \p or \P was
2337 present in the pattern. These operators have additional parameters compared
2338 with \d, etc, and these were not being taken into account when moving along
2339 the compiled data. Specifically:
2340
2341 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2342 length.
2343
2344 (b) An item such as \pL+ within a repeated group could cause crashes or
2345 loops.
2346
2347 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2348 "reference to non-existent subpattern" error.
2349
2350 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2351
2352 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2353 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2354
2355 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2356
2357 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2358 character were causing crashes (broken optimization).
2359
2360 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2361 \p or \P) caused a compile-time loop.
2362
2363 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2364 break. For example, the unstudied pattern [\r\n]A does not match the string
2365 "\r\nA" because change 7.0/46 below moves the current point on by two
2366 characters after failing to match at the start. However, the pattern \nA
2367 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2368 the same is true. There doesn't seem any very clean way out of this, but
2369 what I have chosen to do makes the common cases work: PCRE now takes note
2370 of whether there can be an explicit match for \r or \n anywhere in the
2371 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2372 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2373 pattern has explicit CR or LF references.
2374
2375 29. Added (*CR) etc for changing newline setting at start of pattern.
2376
2377
2378 Version 7.2 19-Jun-07
2379 ---------------------
2380
2381 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2382 which is apparently normally available under Windows.
2383
2384 2. Re-jig the pcregrep tests with different newline settings in an attempt
2385 to make them independent of the local environment's newline setting.
2386
2387 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2388
2389 4. Some of the "internals" tests were previously cut out when the link size
2390 was not 2, because the output contained actual offsets. The recent new
2391 "Z" feature of pcretest means that these can be cut out, making the tests
2392 usable with all link sizes.
2393
2394 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2395 stack recursion. This gives a massive performance boost under BSD, but just
2396 a small improvement under Linux. However, it saves one field in the frame
2397 in all cases.
2398
2399 6. Added more features from the forthcoming Perl 5.10:
2400
2401 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2402 recursion call. It refers to the nth most recently opened parentheses.
2403
2404 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2405 to be opened parentheses.
2406
2407 (c) Conditions that refer to capturing parentheses can be specified
2408 relatively, for example, (?(-2)... or (?(+3)...
2409
2410 (d) \K resets the start of the current match so that everything before
2411 is not part of it.
2412
2413 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2414
2415 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2416 reference syntax.
2417
2418 (g) (?| introduces a group in which the numbering of parentheses in each
2419 alternative starts with the same number.
2420
2421 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2422
2423 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2424 PCRE_INFO_JCHANGED.
2425
2426 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2427 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2428 for detecting groups that can match an empty string.
2429
2430 9. A pattern with a very large number of alternatives (more than several
2431 hundred) was running out of internal workspace during the pre-compile
2432 phase, where pcre_compile() figures out how much memory will be needed. A
2433 bit of new cunning has reduced the workspace needed for groups with
2434 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2435 workspace instead of running out of the 4096 that are available.
2436
2437 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2438
2439 11. Applied patch from Google to remove an optimization that didn't quite work.
2440 The report of the bug said:
2441
2442 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2443 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2444 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2445
2446 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2447 it matched the wrong number of bytes.
2448
2449
2450 Version 7.1 24-Apr-07
2451 ---------------------
2452
2453 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2454 that is more "standard", making use of automake and other Autotools. There
2455 is some re-arrangement of the files and adjustment of comments consequent
2456 on this.
2457
2458 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2459 for recursive directory scanning broke on some systems because the files
2460 are not scanned in any specific order and on different systems the order
2461 was different. A call to "sort" has been inserted into RunGrepTest for the
2462 approprate test as a short-term fix. In the longer term there may be an
2463 alternative.
2464
2465 3. I had an email from Eric Raymond about problems translating some of PCRE's
2466 man pages to HTML (despite the fact that I distribute HTML pages, some
2467 people do their own conversions for various reasons). The problems
2468 concerned the use of low-level troff macros .br and .in. I have therefore
2469 removed all such uses from the man pages (some were redundant, some could
2470 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2471 HTML has been updated to handle .nf/.fi and to complain if it encounters
2472 .br or .in.
2473
2474 4. Updated comments in configure.ac that get placed in config.h.in and also
2475 arranged for config.h to be included in the distribution, with the name
2476 config.h.generic, for the benefit of those who have to compile without
2477 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2478
2479 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2480 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2481 makevp.bat for the current PCRE, using the additional files
2482 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2483
2484 6. A Windows user reported a minor discrepancy with test 2, which turned out
2485 to be caused by a trailing space on an input line that had got lost in his
2486 copy. The trailing space was an accident, so I've just removed it.
2487
2488 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2489 that is needed.
2490
2491 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2492 as "const" (a) because they are and (b) because it helps the PHP
2493 maintainers who have recently made a script to detect big data structures
2494 in the php code that should be moved to the .rodata section. I remembered
2495 to update Builducptable as well, so it won't revert if ucptable.h is ever
2496 re-created.
2497
2498 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2499 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2500 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2501 support is not required. This saves 1.5-2K of code, which is important in
2502 some applications.
2503
2504 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2505 so as not to refer to the tables, even though these functions will never be
2506 called when UTF-8 support is disabled. Otherwise there are problems with a
2507 shared library.
2508
2509 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2510
2511 (a) It was defining its arguments as char * instead of void *.
2512
2513 (b) It was assuming that all moves were upwards in memory; this was true
2514 a long time ago when I wrote it, but is no longer the case.
2515
2516 The emulated memove() is provided for those environments that have neither
2517 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2518 is clearly not the case, as these two bugs were recently reported.
2519
2520 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2521 and Detrail to create the HTML documentation, the .txt form of the man
2522 pages, and it removes trailing spaces from listed files. It also creates
2523 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2524 case, it wraps all the #defines with #ifndefs. This script should be run
2525 before "make dist".
2526
2527 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2528 with Unicode property support.
2529
2530 (a) For a maximizing quantifier, if the two different cases of the
2531 character were of different lengths in their UTF-8 codings (there are
2532 some cases like this - I found 11), and the matching function had to
2533 back up over a mixture of the two cases, it incorrectly assumed they
2534 were both the same length.
2535
2536 (b) When PCRE was configured to use the heap rather than the stack for
2537 recursion during matching, it was not correctly preserving the data for
2538 the other case of a UTF-8 character when checking ahead for a match
2539 while processing a minimizing repeat. If the check also involved
2540 matching a wide character, but failed, corruption could cause an
2541 erroneous result when trying to check for a repeat of the original
2542 character.
2543
2544 13. Some tidying changes to the testing mechanism:
2545
2546 (a) The RunTest script now detects the internal link size and whether there
2547 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2548 values substituted by "configure". (The RunGrepTest script already did
2549 this for UTF-8.) The configure.ac script no longer substitutes the
2550 relevant variables.
2551
2552 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2553 with length and offset values. This means that the output is different
2554 for different internal link sizes. Test 2 is skipped for link sizes
2555 other than 2 because of this, bypassing the problem. Unfortunately,
2556 there was also a test in test 3 (the locale tests) that used /B and
2557 failed for link sizes other than 2. Rather than cut the whole test out,
2558 I have added a new /Z option to pcretest that replaces the length and
2559 offset values with spaces. This is now used to make test 3 independent
2560 of link size. (Test 2 will be tidied up later.)
2561
2562 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2563 segmentation fault instead of returning the appropriate error message.
2564
2565 15. In multiline mode when the newline sequence was set to "any", the pattern
2566 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2567 This doesn't seem right; it now treats the CRLF combination as the line
2568 ending, and so does not match in that case. It's only a pattern such as ^$
2569 that would hit this one: something like ^ABC$ would have failed after \r
2570 and then tried again after \r\n.
2571
2572 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2573 in an attempt to make files that differ only in their line terminators
2574 compare equal. This works on Linux.
2575
2576 17. Under certain error circumstances pcregrep might try to free random memory
2577 as it exited. This is now fixed, thanks to valgrind.
2578
2579 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2580 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2581 was because its rules for how to advance for /g after matching an empty
2582 string at the end of a line did not allow for this case. They now check for
2583 it specially.
2584
2585 20. pcretest is supposed to handle patterns and data of any length, by
2586 extending its buffers when necessary. It was getting this wrong when the
2587 buffer for a data line had to be extended.
2588
2589 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2590 CRLF as a newline sequence.
2591
2592 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2593 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2594 I have nevertheless tidied it up.
2595
2596 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2597
2598 24. Added a man page for pcre-config.
2599
2600
2601 Version 7.0 19-Dec-06
2602 ---------------------
2603
2604 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2605 moving to gcc 4.1.1.
2606
2607 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2608 sys/time.h, which is documented as needed for this function. It doesn't
2609 seem to matter on Linux, but it showed up on some releases of OS X.
2610
2611 3. It seems that there are systems where bytes whose values are greater than
2612 127 match isprint() in the "C" locale. The "C" locale should be the
2613 default when a C program starts up. In most systems, only ASCII printing
2614 characters match isprint(). This difference caused the output from pcretest
2615 to vary, making some of the tests fail. I have changed pcretest so that:
2616
2617 (a) When it is outputting text in the compiled version of a pattern, bytes
2618 other than 32-126 are always shown as hex escapes.
2619
2620 (b) When it is outputting text that is a matched part of a subject string,
2621 it does the same, unless a different locale has been set for the match
2622 (using the /L modifier). In this case, it uses isprint() to decide.
2623
2624 4. Fixed a major bug that caused incorrect computation of the amount of memory
2625 required for a compiled pattern when options that changed within the
2626 pattern affected the logic of the preliminary scan that determines the
2627 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2628 that the computed length was too small. The symptoms of this bug were
2629 either the PCRE error "internal error: code overflow" from pcre_compile(),
2630 or a glibc crash with a message such as "pcretest: free(): invalid next
2631 size (fast)". Examples of patterns that provoked this bug (shown in
2632 pcretest format) are:
2633
2634 /(?-x: )/x
2635 /(?x)(?-x: \s*#\s*)/
2636 /((?i)[\x{c0}])/8
2637 /(?i:[\x{c0}])/8
2638
2639 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2640 is now done differently.
2641
2642 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2643 wrapper classes; (b) implement a new function in the C++ scanner that is
2644 more efficient than the old way of doing things because it avoids levels of
2645 recursion in the regex matching; (c) add a paragraph to the documentation
2646 for the FullMatch() function.
2647
2648 6. The escape sequence \n was being treated as whatever was defined as
2649 "newline". Not only was this contrary to the documentation, which states
2650 that \n is character 10 (hex 0A), but it also went horribly wrong when
2651 "newline" was defined as CRLF. This has been fixed.
2652
2653 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2654 was being set to -1 for the "end of line" case (supposedly a value that no
2655 character can have). Though this value is never used (the check for end of
2656 line is "zero bytes in current character"), it caused compiler complaints.
2657 I've changed it to 0xffffffff.
2658
2659 8. In pcre_version.c, the version string was being built by a sequence of
2660 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2661 string (as it is for production releases) called a macro with an empty
2662 argument. The C standard says the result of this is undefined. The gcc
2663 compiler treats it as an empty string (which was what was wanted) but it is
2664 reported that Visual C gives an error. The source has been hacked around to
2665 avoid this problem.
2666
2667 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2668 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2669 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2670 of them did).
2671
2672 10. Originally, pcretest opened its input and output without "b"; then I was
2673 told that "b" was needed in some environments, so it was added for release
2674 5.0 to both the input and output. (It makes no difference on Unix-like
2675 systems.) Later I was told that it is wrong for the input on Windows. I've
2676 now abstracted the modes into two macros, to make it easier to fiddle with
2677 them, and removed "b" from the input mode under Windows.
2678
2679 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2680
2681 12. Added -help and --help to pcretest as an official way of being reminded
2682 of the options.
2683
2684 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2685 and pcrecpp.cc because they annoy compilers at high warning levels.
2686
2687 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2688
2689 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2690 scripts are not C programs :-) and which was not noticed because it works
2691 on Linux.
2692
2693 16. pcretest is supposed to handle any length of pattern and data line (as one
2694 line or as a continued sequence of lines) by extending its input buffer if
2695 necessary. This feature was broken for very long pattern lines, leading to
2696 a string of junk being passed to pcre_compile() if the pattern was longer
2697 than about 50K.
2698
2699 17. I have done a major re-factoring of the way pcre_compile() computes the
2700 amount of memory needed for a compiled pattern. Previously, there was code
2701 that made a preliminary scan of the pattern in order to do this. That was
2702 OK when PCRE was new, but as the facilities have expanded, it has become
2703 harder and harder to keep it in step with the real compile phase, and there
2704 have been a number of bugs (see for example, 4 above). I have now found a
2705 cunning way of running the real compile function in a "fake" mode that
2706 enables it to compute how much memory it would need, while actually only
2707 ever using a few hundred bytes of working memory and without too many
2708 tests of the mode. This should make future maintenance and development
2709 easier. A side effect of this work is that the limit of 200 on the nesting
2710 depth of parentheses has been removed (though this was never a serious
2711 limitation, I suspect). However, there is a downside: pcre_compile() now
2712 runs more slowly than before (30% or more, depending on the pattern). I
2713 hope this isn't a big issue. There is no effect on runtime performance.
2714
2715 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2716 newline (only possible for the last line of a file) and it was a
2717 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2718
2719 19. Added additional timing features to pcretest. (1) The -tm option now times
2720 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2721 separate command line item, by a number that specifies the number of
2722 repeats to use when timing. The default is 50000; this gives better
2723 precision, but takes uncomfortably long for very large patterns.
2724
2725 20. Extended pcre_study() to be more clever in cases where a branch of a
2726 subpattern has no definite first character. For example, (a*|b*)[cd] would
2727 previously give no result from pcre_study(). Now it recognizes that the
2728 first character must be a, b, c, or d.
2729
2730 21. There was an incorrect error "recursive call could loop indefinitely" if
2731 a subpattern (or the entire pattern) that was being tested for matching an
2732 empty string contained only one non-empty item after a nested subpattern.
2733 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2734 incorrectly, because the \d was being skipped in the check.
2735
2736 22. The pcretest program now has a new pattern option /B and a command line
2737 option -b, which is equivalent to adding /B to every pattern. This causes
2738 it to show the compiled bytecode, without the additional information that
2739 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2740 is the same as /B/I).
2741
2742 23. A new optimization is now able automatically to treat some sequences such
2743 as a*b as a*+b. More specifically, if something simple (such as a character
2744 or a simple class like \d) has an unlimited quantifier, and is followed by
2745 something that cannot possibly match the quantified thing, the quantifier
2746 is automatically "possessified".
2747
2748 24. A recursive reference to a subpattern whose number was greater than 39
2749 went wrong under certain circumstances in UTF-8 mode. This bug could also
2750 have affected the operation of pcre_study().
2751
2752 25. Realized that a little bit of performance could be had by replacing
2753 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2754
2755 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2756
2757 27. Possessive quantifiers such as a++ were previously implemented by turning
2758 them into atomic groups such as ($>a+). Now they have their own opcodes,
2759 which improves performance. This includes the automatically created ones
2760 from 23 above.
2761
2762 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2763 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2764 the first matched character to be a colon. This applied both to named and
2765 numbered groups.
2766
2767 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2768
2769 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2770 building PCRE on Windows easier, so I have included it in the distribution.
2771
2772 31. There is now a check in pcretest against a ridiculously large number being
2773 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2774 loop, the loop is abandoned.
2775
2776 32. Forward references to subpatterns in conditions such as (?(2)...) where
2777 subpattern 2 is defined later cause pcre_compile() to search forwards in
2778 the pattern for the relevant set of parentheses. This search went wrong
2779 when there were unescaped parentheses in a character class, parentheses
2780 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2781
2782 33. "Subroutine" calls and backreferences were previously restricted to
2783 referencing subpatterns earlier in the regex. This restriction has now
2784 been removed.
2785
2786 34. Added a number of extra features that are going to be in Perl 5.10. On the
2787 whole, these are just syntactic alternatives for features that PCRE had
2788 previously implemented using the Python syntax or my own invention. The
2789 other formats are all retained for compatibility.
2790
2791 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2792 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2793 also .NET compatible.
2794
2795 (b) A recursion or subroutine call to a named group can now be defined as
2796 (?&name) as well as (?P>name).
2797
2798 (c) A backreference to a named group can now be defined as \k<name> or
2799 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2800 5.10, are also .NET compatible.
2801
2802 (d) A conditional reference to a named group can now use the syntax
2803 (?(<name>) or (?('name') as well as (?(name).
2804
2805 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2806 groups (named and numbered) that are never evaluated inline, but can be
2807 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2808 is always false. There may be only one alternative in such a group.
2809
2810 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2811 as the simple (?(R). The condition is true only if the most recent
2812 recursion is that of the given number or name. It does not search out
2813 through the entire recursion stack.
2814
2815 (g) The escape \gN or \g{N} has been added, where N is a positive or
2816 negative number, specifying an absolute or relative reference.
2817
2818 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2819 some "unreachable code" warnings.
2820
2821 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2822 things, this adds five new scripts.
2823
2824 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2825 There were also incompatibilities regarding the handling of \Q..\E inside
2826 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2827 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2828
2829 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2830 matches an empty string, and forcibly breaks the loop. There were bugs in
2831 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2832 against aaaa the result was just "a" rather than "aaaa", for example. Two
2833 separate and independent bugs (that affected different cases) have been
2834 fixed.
2835
2836 39. Refactored the code to abolish the use of different opcodes for small
2837 capturing bracket numbers. This is a tidy that I avoided doing when I
2838 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2839 The new approach is not only tidier, it makes it possible to reduce the
2840 memory needed to fix the previous bug (38).
2841
2842 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2843 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2844 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2845 mode.
2846
2847 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2848 report.
2849
2850 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2851 copy construction and assignment in the C++ wrapper.
2852
2853 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2854 couple of bugs that could have given wrong results in the "--newline=crlf"
2855 case.
2856
2857 44. Added a number of casts and did some reorganization of signed/unsigned int
2858 variables following suggestions from Dair Grant. Also renamed the variable
2859 "this" as "item" because it is a C++ keyword.
2860
2861 45. Arranged for dftables to add
2862
2863 #include "pcre_internal.h"
2864
2865 to pcre_chartables.c because without it, gcc 4.x may remove the array
2866 definition from the final binary if PCRE is built into a static library and
2867 dead code stripping is activated.
2868
2869 46. For an unanchored pattern, if a match attempt fails at the start of a
2870 newline sequence, and the newline setting is CRLF or ANY, and the next two
2871 characters are CRLF, advance by two characters instead of one.
2872
2873
2874 Version 6.7 04-Jul-06
2875 ---------------------
2876
2877 1. In order to handle tests when input lines are enormously long, pcretest has
2878 been re-factored so that it automatically extends its buffers when
2879 necessary. The code is crude, but this _is_ just a test program. The
2880 default size has been increased from 32K to 50K.
2881
2882 2. The code in pcre_study() was using the value of the re argument before
2883 testing it for NULL. (Of course, in any sensible call of the function, it
2884 won't be NULL.)
2885
2886 3. The memmove() emulation function in pcre_internal.h, which is used on
2887 systems that lack both memmove() and bcopy() - that is, hardly ever -
2888 was missing a "static" storage class specifier.
2889
2890 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2891 containing an extended class (one that cannot be represented by a bitmap
2892 because it contains high-valued characters or Unicode property items, e.g.
2893 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2894 pattern, but PCRE should not loop if you do not (it no longer does).
2895 [Detail: two cases were found: (a) a repeated subpattern containing an
2896 extended class; (b) a recursive reference to a subpattern that followed a
2897 previous extended class. It wasn't skipping over the extended class
2898 correctly when UTF-8 mode was not set.]
2899
2900 5. A negated single-character class was not being recognized as fixed-length
2901 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2902 compile error "lookbehind assertion is not fixed length".
2903
2904 6. The RunPerlTest auxiliary script was showing an unexpected difference
2905 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2906 write a Perl script that can interpret lines of an input file either as
2907 byte characters or as UTF-8, which is what "perltest" was being required to
2908 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2909 can't do is switch easily at run time between having the "use utf8;" pragma
2910 or not. In the end, I fudged it by using the RunPerlTest script to insert
2911 "use utf8;" explicitly for the UTF-8 tests.
2912
2913 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2914 the end of the subject string, contrary to the documentation and to what
2915 Perl does. This was true of both matching functions. Now it matches only at
2916 the start of the subject and immediately after *internal* newlines.
2917
2918 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2919 a pointer to an int instead of a pointer to an unsigned long int. This
2920 caused problems on 64-bit systems.
2921
2922 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2923 instance of the 'standard' template library not being so standard".
2924
2925 10. There was no check on the number of named subpatterns nor the maximum
2926 length of a subpattern name. The product of these values is used to compute
2927 the size of the memory block for a compiled pattern. By supplying a very
2928 long subpattern name and a large number of named subpatterns, the size
2929 computation could be caused to overflow. This is now prevented by limiting
2930 the length of names to 32 characters, and the number of named subpatterns
2931 to 10,000.
2932
2933 11. Subpatterns that are repeated with specific counts have to be replicated in
2934 the compiled pattern. The size of memory for this was computed from the
2935 length of the subpattern and the repeat count. The latter is limited to
2936 65535, but there was no limit on the former, meaning that integer overflow
2937 could in principle occur. The compiled length of a repeated subpattern is
2938 now limited to 30,000 bytes in order to prevent this.
2939
2940 12. Added the optional facility to have named substrings with the same name.
2941
2942 13. Added the ability to use a named substring as a condition, using the
2943 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2944 are numbers (not recommended). Forward references are permitted.
2945
2946 14. Added forward references in named backreferences (if you see what I mean).
2947
2948 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2949 pattern could run off the end of the subject. For example, the pattern
2950 "(?s)(.{1,5})"8 did this with the subject "ab".
2951
2952 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2953 PCRE_CASELESS was set when matching characters that were quantified with ?
2954 or *.
2955
2956 17. A character class other than a single negated character that had a minimum
2957 but no maximum quantifier - for example [ab]{6,} - was not handled
2958 correctly by pce_dfa_exec(). It would match only one character.
2959
2960 18. A valid (though odd) pattern that looked like a POSIX character
2961 class but used an invalid character after [ (for example [[,abc,]]) caused
2962 pcre_compile() to give the error "Failed: internal error: code overflow" or
2963 in some cases to crash with a glibc free() error. This could even happen if
2964 the pattern terminated after [[ but there just happened to be a sequence of
2965 letters, a binary zero, and a closing ] in the memory that followed.
2966
2967 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2968 over the years. Originally (before any Unicode support), just the bottom 8
2969 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2970 output from "man perlunicode" includes this:
2971
2972 The regular expression compiler produces polymorphic opcodes. That
2973 is, the pattern adapts to the data and automatically switches to
2974 the Unicode character scheme when presented with Unicode data--or
2975 instead uses a traditional byte scheme when presented with byte
2976 data.
2977
2978 Sadly, a wide octal escape does not cause a switch, and in a string with
2979 no other multibyte characters, these octal escapes are treated as before.
2980 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2981 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2982 Unicode string.
2983
2984 I have not perpetrated such confusion in PCRE. Up till now, it took just
2985 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2986 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2987 translate to the appropriate multibyte character.
2988
2989 29. Applied some refactoring to reduce the number of warnings from Microsoft
2990 and Borland compilers. This has included removing the fudge introduced
2991 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2992 a warning about an unused variable.
2993
2994 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2995 characters since release 4.0, because Perl (from release 5.004) does not.
2996 [Or at least, is documented not to: some releases seem to be in conflict
2997 with the documentation.] However, when a pattern was studied with
2998 pcre_study() and all its branches started with \s, PCRE still included VT
2999 as a possible starting character. Of course, this did no harm; it just
3000 caused an unnecessary match attempt.
3001
3002 22. Removed a now-redundant internal flag bit that recorded the fact that case
3003 dependency changed within the pattern. This was once needed for "required
3004 byte" processing, but is no longer used. This recovers a now-scarce options
3005 bit. Also moved the least significant internal flag bit to the most-
3006 significant bit of the word, which was not previously used (hangover from
3007 the days when it was an int rather than a uint) to free up another bit for
3008 the future.
3009
3010 23. Added support for CRLF line endings as well as CR and LF. As well as the
3011 default being selectable at build time, it can now be changed at runtime
3012 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3013 specify that it is scanning data with non-default line endings.
3014
3015 24. Changed the definition of CXXLINK to make it agree with the definition of
3016 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3017
3018 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3019 recursions. This makes a big different to stack usage for some patterns.
3020
3021 26. If a subpattern containing a named recursion or subroutine reference such
3022 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3023 the space required for the compiled pattern went wrong and gave too small a
3024 value. Depending on the environment, this could lead to "Failed: internal
3025 error: code overflow at offset 49" or "glibc detected double free or
3026 corruption" errors.
3027
3028 27. Applied patches from Google (a) to support the new newline modes and (b) to
3029 advance over multibyte UTF-8 characters in GlobalReplace.
3030
3031 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3032 difference for some implementation of PCRE in some Windows version.
3033
3034 29. Added some extra testing facilities to pcretest:
3035
3036 \q<number> in a data line sets the "match limit" value
3037 \Q<number> in a data line sets the "match recursion limt" value
3038 -S <number> sets the stack size, where <number> is in megabytes
3039
3040 The -S option isn't available for Windows.
3041
3042
3043 Version 6.6 06-Feb-06
3044 ---------------------
3045
3046 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3047 in pcreposix.h. I have copied the definition from pcre.h.
3048
3049 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3050 because pcre.h is no longer a built file.
3051
3052 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3053 not normally included in the compiled code.
3054
3055
3056 Version 6.5 01-Feb-06
3057 ---------------------
3058
3059 1. When using the partial match feature with pcre_dfa_exec(), it was not
3060 anchoring the second and subsequent partial matches at the new starting
3061 point. This could lead to incorrect results. For example, with the pattern
3062 /1234/, partially matching against "123" and then "a4" gave a match.
3063
3064 2. Changes to pcregrep:
3065
3066 (a) All non-match returns from pcre_exec() were being treated as failures
3067 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3068 error message is output. Some extra information is given for the
3069 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3070 probably the only errors that are likely to be caused by users (by
3071 specifying a regex that has nested indefinite repeats, for instance).
3072 If there are more than 20 of these errors, pcregrep is abandoned.
3073
3074 (b) A binary zero was treated as data while matching, but terminated the
3075 output line if it was written out. This has been fixed: binary zeroes
3076 are now no different to any other data bytes.
3077
3078 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3079 used to set a locale for matching. The --locale=xxxx long option has
3080 been added (no short equivalent) to specify a locale explicitly on the
3081 pcregrep command, overriding the environment variables.
3082
3083 (d) When -B was used with -n, some line numbers in the output were one less
3084 than they should have been.
3085
3086 (e) Added the -o (--only-matching) option.
3087
3088 (f) If -A or -C was used with -c (count only), some lines of context were
3089 accidentally printed for the final match.
3090
3091 (g) Added the -H (--with-filename) option.
3092
3093 (h) The combination of options -rh failed to suppress file names for files
3094 that were found from directory arguments.
3095
3096 (i) Added the -D (--devices) and -d (--directories) options.
3097
3098 (j) Added the -F (--fixed-strings) option.
3099
3100 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3101
3102 (l) Added the --colo(u)r option.
3103
3104 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3105 is not present by default.
3106
3107 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3108 items such as (?R) or (?1), when the recursion could match a number of
3109 alternatives. If it matched one of the alternatives, but subsequently,
3110 outside the recursion, there was a failure, the code tried to back up into
3111 the recursion. However, because of the way PCRE is implemented, this is not
3112 possible, and the result was an incorrect result from the match.
3113
3114 In order to prevent this happening, the specification of recursion has
3115 been changed so that all such subpatterns are automatically treated as
3116 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3117
3118 4. I had overlooked the fact that, in some locales, there are characters for
3119 which isalpha() is true but neither isupper() nor islower() are true. In
3120 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3121 and ordfeminine) are like this. This affected the treatment of \w and \W
3122 when they appeared in character classes, but not when they appeared outside
3123 a character class. The bit map for "word" characters is now created
3124 separately from the results of isalnum() instead of just taking it from the
3125 upper, lower, and digit maps. (Plus the underscore character, of course.)
3126
3127 5. The above bug also affected the handling of POSIX character classes such as
3128 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3129 permanent tables. Instead, the bit maps for such a class were previously
3130 created as the appropriate unions of the upper, lower, and digit bitmaps.
3131 Now they are created by subtraction from the [[:word:]] class, which has
3132 its own bitmap.
3133
3134 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3135 It is created by subtracting the vertical space characters (\x09, \x0a,
3136 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3137 subtraction was done in the overall bitmap for a character class, meaning
3138 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3139 be recognized. This bug has been fixed.
3140
3141 7. Patches from the folks at Google:
3142
3143 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3144 real life, but is still worth protecting against".
3145
3146 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3147 regular expressions".
3148
3149 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3150 have it.
3151
3152 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3153 "configure" and the latter not, in order to fix a problem somebody had
3154 with compiling the Arg class on HP-UX.
3155
3156 (e) Improve the error-handling of the C++ wrapper a little bit.
3157
3158 (f) New tests for checking recursion limiting.
3159
3160 8. The pcre_memmove() function, which is used only if the environment does not
3161 have a standard memmove() function (and is therefore rarely compiled),
3162 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3163 returning a result (though PCRE never actually uses the result).
3164
3165 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3166 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3167 returned instead of calling malloc() with an overflowing number that would
3168 most likely cause subsequent chaos.
3169
3170 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3171
3172 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3173 with this option is matched, the nmatch and pmatch options of regexec() are
3174 ignored.
3175
3176 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3177 provided in case anyone wants to the the POSIX interface with UTF-8
3178 strings.
3179
3180 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3181 C++ linking (needed for some HP-UX environments).
3182
3183 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3184 (unused parameter) and in the pcre_printint() function (omitted "default"
3185 switch label when the default is to do nothing).
3186
3187 15. Added some code to make it possible, when PCRE is compiled as a C++
3188 library, to replace subject pointers for pcre_exec() with a smart pointer
3189 class, thus making it possible to process discontinuous strings.
3190
3191 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3192 much the same function. They were added by different people who were trying
3193 to make PCRE easy to compile on non-Unix systems. It has been suggested
3194 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3195 for compiling on Windows systems. I have therefore replaced it with
3196 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3197 defaults to "extern" for C or "extern C" for C++, which works fine on
3198 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3199 SCOPE with something explicit in config.h. In addition:
3200
3201 (a) pcreposix.h still had just "extern" instead of either of these macros;
3202 I have replaced it with PCRE_DATA_SCOPE.
3203
3204 (b) Functions such as _pcre_xclass(), which are internal to the library,
3205 but external in the C sense, all had PCRE_EXPORT in their definitions.
3206 This is apparently wrong for the Windows case, so I have removed it.
3207 (It makes no difference on Unix-like systems.)
3208
3209 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3210 of recursive calls to match(). This is different to MATCH_LIMIT because
3211 that limits the total number of calls to match(), not all of which increase
3212 the depth of recursion. Limiting the recursion depth limits the amount of
3213 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3214 when PCRE is compiled, and changed at run time. A patch from Google adds
3215 this functionality to the C++ interface.
3216
3217 18. Changes to the handling of Unicode character properties:
3218
3219 (a) Updated the table to Unicode 4.1.0.
3220
3221 (b) Recognize characters that are not in the table as "Cn" (undefined).
3222
3223 (c) I revised the way the table is implemented to a much improved format
3224 which includes recognition of ranges. It now supports the ranges that
3225 are defined in UnicodeData.txt, and it also amalgamates other
3226 characters into ranges. This has reduced the number of entries in the
3227 table from around 16,000 to around 3,000, thus reducing its size
3228 considerably. I realized I did not need to use a tree structure after
3229 all - a binary chop search is just as efficient. Having reduced the
3230 number of entries, I extended their size from 6 bytes to 8 bytes to
3231 allow for more data.
3232
3233 (d) Added support for Unicode script names via properties such as \p{Han}.
3234
3235 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3236 matching that character.
3237
3238 20. When matching a repeated Unicode property with a minimum greater than zero,
3239 (for example \pL{2,}), PCRE could look past the end of the subject if it
3240 reached it while seeking the minimum number of characters. This could
3241 happen only if some of the characters were more than one byte long, because
3242 there is a check for at least the minimum number of bytes.
3243
3244 21. Refactored the implementation of \p and \P so as to be more general, to
3245 allow for more different types of property in future. This has changed the
3246 compiled form incompatibly. Anybody with saved compiled patterns that use
3247 \p or \P will have to recompile them.
3248
3249 22. Added "Any" and "L&" to the supported property types.
3250
3251 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3252 but give a compile time error if the value is greater than 0xff.
3253
3254 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3255 accidentally not being installed or uninstalled.
3256
3257 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3258 made were to insert the current release number. This seemed silly, because
3259 it made things harder for people building PCRE on systems that don't run
3260 "configure". I have turned pcre.h into a distributed file, no longer built
3261 by "configure", with the version identification directly included. There is
3262 no longer a pcre.h.in file.
3263
3264 However, this change necessitated a change to the pcre-config script as
3265 well. It is built from pcre-config.in, and one of the substitutions was the
3266 release number. I have updated configure.ac so that ./configure now finds
3267 the release number by grepping pcre.h.
3268
3269 26. Added the ability to run the tests under valgrind.
3270
3271
3272 Version 6.4 05-Sep-05
3273 ---------------------
3274
3275 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3276 "--" to be printed when multiple files were scanned, even when none of the
3277 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3278 consider it to be a bug, and have restored the previous behaviour.
3279
3280 2. A couple of code tidies to get rid of compiler warnings.
3281
3282 3. The pcretest program used to cheat by referring to symbols in the library
3283 whose names begin with _pcre_. These are internal symbols that are not
3284 really supposed to be visible externally, and in some environments it is
3285 possible to suppress them. The cheating is now confined to including
3286 certain files from the library's source, which is a bit cleaner.
3287
3288 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3289 file's purpose clearer.
3290
3291 5. Reorganized pcre_ucp_findchar().
3292
3293
3294 Version 6.3 15-Aug-05
3295 ---------------------
3296
3297 1. The file libpcre.pc.in did not have general read permission in the tarball.
3298
3299 2. There were some problems when building without C++ support:
3300
3301 (a) If C++ support was not built, "make install" and "make test" still
3302 tried to test it.
3303
3304 (b) There were problems when the value of CXX was explicitly set. Some
3305 changes have been made to try to fix these, and ...
3306
3307 (c) --disable-cpp can now be used to explicitly disable C++ support.
3308
3309 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3310 backslash in a target when C++ was disabled. This confuses some
3311 versions of "make", apparently. Using an intermediate variable solves
3312 this. (Same for CPP_LOBJ.)
3313
3314 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3315 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3316 necessary on certain architectures.
3317
3318 4. Added a setting of -export-symbols-regex to the link command to remove
3319 those symbols that are exported in the C sense, but actually are local
3320 within the library, and not documented. Their names all begin with
3321 "_pcre_". This is not a perfect job, because (a) we have to except some
3322 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3323 available (and never for static libraries). I have made a note to try to
3324 find a way round (a) in the future.
3325
3326
3327 Version 6.2 01-Aug-05
3328 ---------------------
3329
3330 1. There was no test for integer overflow of quantifier values. A construction
3331 such as {1111111111111111} would give undefined results. What is worse, if
3332 a minimum quantifier for a parenthesized subpattern overflowed and became
3333 negative, the calculation of the memory size went wrong. This could have
3334 led to memory overwriting.
3335
3336 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3337
3338 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3339 operating environments where this matters.
3340
3341 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3342 PCRE options from within the C++ wrapper.
3343
3344 5. Named capturing subpatterns were not being correctly counted when a pattern
3345 was compiled. This caused two problems: (a) If there were more than 100
3346 such subpatterns, the calculation of the memory needed for the whole
3347 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3348 back references of the form \12, where the number was greater than 9, were
3349 not recognized as back references, even though there were sufficient
3350 previous subpatterns.
3351
3352 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3353 versions of gcc, e.g. 2.95.4.
3354
3355
3356 Version 6.1 21-Jun-05
3357 ---------------------
3358
3359 1. There was one reference to the variable "posix" in pcretest.c that was not
3360 surrounded by "#if !defined NOPOSIX".
3361
3362 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3363 the cross-check on the old pcre_info() function, for the benefit of the
3364 cut-down version of PCRE that is currently imported into Exim.
3365
3366 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3367 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3368 patterns that start (?i) but I don't think that matters. The use of (?i) is
3369 just an example; this all applies to the other options as well.
3370
3371 4. Since libtool seems to echo the compile commands it is issuing, the output
3372 from "make" can be reduced a bit by putting "@" in front of each libtool
3373 compile command.
3374
3375 5. Patch from the folks at Google for configure.in to be a bit more thorough
3376 in checking for a suitable C++ installation before trying to compile the
3377 C++ stuff. This should fix a reported problem when a compiler was present,
3378 but no suitable headers.
3379
3380 6. The man pages all had just "PCRE" as their title. I have changed them to
3381 be the relevant file name. I have also arranged that these names are
3382 retained in the file doc/pcre.txt, which is a concatenation in text format
3383 of all the man pages except the little individual ones for each function.
3384
3385 7. The NON-UNIX-USE file had not been updated for the different set of source
3386 files that come with release 6. I also added a few comments about the C++
3387 wrapper.
3388
3389
3390 Version 6.0 07-Jun-05
3391 ---------------------
3392
3393 1. Some minor internal re-organization to help with my DFA experiments.
3394
3395 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3396 didn't matter for the library itself when fully configured, but did matter
3397 when compiling without UCP support, or within Exim, where the ucp files are
3398 not imported.
3399
3400 3. Refactoring of the library code to split up the various functions into
3401 different source modules. The addition of the new DFA matching code (see
3402 below) to a single monolithic source would have made it really too
3403 unwieldy, quite apart from causing all the code to be include in a
3404 statically linked application, when only some functions are used. This is
3405 relevant even without the DFA addition now that patterns can be compiled in
3406 one application and matched in another.
3407
3408 The downside of splitting up is that there have to be some external
3409 functions and data tables that are used internally in different modules of
3410 the library but which are not part of the API. These have all had their
3411 names changed to start with "_pcre_" so that they are unlikely to clash
3412 with other external names.
3413
3414 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3415 a different (DFA) algorithm. Although it is slower than the original
3416 function, it does have some advantages for certain types of matching
3417 problem.
3418
3419 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3420 including restarting after a partial match.
3421
3422 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3423 defined when compiling for Windows was sent to me. I have put it into the
3424 code, though I have no means of testing or verifying it.
3425
3426 7. Added the pcre_refcount() auxiliary function.
3427
3428 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3429 match before or at the first newline in the subject string. In pcretest,
3430 the /f option on a pattern can be used to set this.
3431
3432 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3433 would behave wrongly. This has been present in PCRE since release 4.0.
3434
3435 10. A number of changes to the pcregrep command:
3436
3437 (a) Refactored how -x works; insert ^(...)$ instead of setting
3438 PCRE_ANCHORED and checking the length, in preparation for adding
3439 something similar for -w.
3440
3441 (b) Added the -w (match as a word) option.
3442
3443 (c) Refactored the way lines are read and buffered so as to have more
3444 than one at a time available.
3445
3446 (d) Implemented a pcregrep test script.
3447
3448 (e) Added the -M (multiline match) option. This allows patterns to match
3449 over several lines of the subject. The buffering ensures that at least
3450 8K, or the rest of the document (whichever is the shorter) is available
3451 for matching (and similarly the previous 8K for lookbehind assertions).
3452
3453 (f) Changed the --help output so that it now says
3454
3455 -w, --word-regex(p)
3456
3457 instead of two lines, one with "regex" and the other with "regexp"
3458 because that confused at least one person since the short forms are the
3459 same. (This required a bit of code, as the output is generated
3460 automatically from a table. It wasn't just a text change.)
3461
3462 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3463 option but starts with a hyphen. Could be a pattern or a path name
3464 starting with a hyphen, for instance.
3465
3466 (h) "-" can be given as a file name to represent stdin.
3467
3468 (i) When file names are being printed, "(standard input)" is used for
3469 the standard input, for compatibility with GNU grep. Previously
3470 "<stdin>" was used.
3471
3472 (j) The option --label=xxx can be used to supply a name to be used for
3473 stdin when file names are being printed. There is no short form.
3474
3475 (k) Re-factored the options decoding logic because we are going to add
3476 two more options that take data. Such options can now be given in four
3477 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3478
3479 (l) Added the -A, -B, and -C options for requesting that lines of context
3480 around matches be printed.
3481
3482 (m) Added the -L option to print the names of files that do not contain
3483 any matching lines, that is, the complement of -l.
3484
3485 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3486 continue to scan other files.
3487
3488 (o) The -s option was incorrectly implemented. For compatibility with other
3489 greps, it now suppresses the error message for a non-existent or non-
3490 accessible file (but not the return code). There is a new option called
3491 -q that suppresses the output of matching lines, which was what -s was
3492 previously doing.
3493
3494 (p) Added --include and --exclude options to specify files for inclusion
3495 and exclusion when recursing.
3496
3497 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3498 Hopefully, it now does.
3499
3500 12. Missing cast in pcre_study().
3501
3502 13. Added an "uninstall" target to the makefile.
3503
3504 14. Replaced "extern" in the function prototypes in Makefile.in with
3505 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3506 world, but is set differently for Windows.
3507
3508 15. Added a second compiling function called pcre_compile2(). The only
3509 difference is that it has an extra argument, which is a pointer to an
3510 integer error code. When there is a compile-time failure, this is set
3511 non-zero, in addition to the error test pointer being set to point to an
3512 error message. The new argument may be NULL if no error number is required
3513 (but then you may as well call pcre_compile(), which is now just a
3514 wrapper). This facility is provided because some applications need a
3515 numeric error indication, but it has also enabled me to tidy up the way
3516 compile-time errors are handled in the POSIX wrapper.
3517
3518 16. Added VPATH=.libs to the makefile; this should help when building with one
3519 prefix path and installing with another. (Or so I'm told by someone who
3520 knows more about this stuff than I do.)
3521
3522 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3523 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3524 match everything, including newlines. This is not POSIX-compatible, but
3525 somebody wanted the feature. From pcretest it can be activated by using
3526 both the P and the s flags.
3527
3528 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3529
3530 19. libpcre.pc was being incorrectly installed as executable.
3531
3532 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3533 it now also looks for '\r' so that it will work unmodified on Windows.
3534
3535 21. Added Google's contributed C++ wrapper to the distribution.
3536
3537 22. Added some untidy missing memory free() calls in pcretest, to keep
3538 Electric Fence happy when testing.
3539
3540
3541
3542 Version 5.0 13-Sep-04
3543 ---------------------
3544
3545 1. Internal change: literal characters are no longer packed up into items
3546 containing multiple characters in a single byte-string. Each character
3547 is now matched using a separate opcode. However, there may be more than one
3548 byte in the character in UTF-8 mode.
3549
3550 2. The pcre_callout_block structure has two new fields: pattern_position and
3551 next_item_length. These contain the offset in the pattern to the next match
3552 item, and its length, respectively.
3553
3554 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3555 insertion of callouts before each pattern item. Added the /C option to
3556 pcretest to make use of this.
3557
3558 4. On the advice of a Windows user, the lines
3559
3560 #if defined(_WIN32) || defined(WIN32)
3561 _setmode( _fileno( stdout ), 0x8000 );
3562 #endif /* defined(_WIN32) || defined(WIN32) */
3563
3564 have been added to the source of pcretest. This apparently does useful
3565 magic in relation to line terminators.
3566
3567 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3568 for the benefit of those environments where the "b" makes a difference.
3569
3570 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3571 to know about it. I have put a hack into configure.in that adds in code
3572 to set GCC=yes if CC=icc. This seems to end up at a point in the
3573 generated configure script that is early enough to affect the setting of
3574 compiler options, which is what is needed, but I have no means of testing
3575 whether it really works. (The user who reported this had patched the
3576 generated configure script, which of course I cannot do.)
3577
3578 LATER: After change 22 below (new libtool files), the configure script
3579 seems to know about icc (and also ecc). Therefore, I have commented out
3580 this hack in configure.in.
3581
3582 7. Added support for pkg-config (2 patches were sent in).
3583
3584 8. Negated POSIX character classes that used a combination of internal tables
3585 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3586 [[:^ascii]]. Typically, they would match almost any characters. The other
3587 POSIX classes were not broken in this way.
3588
3589 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3590 to find the match, as PCRE was deluded into thinking that the match had to
3591 start at the start point or following a newline. The same bug applied to
3592 patterns with negative forward assertions or any backward assertions
3593 preceding ".*" at the start, unless the pattern required a fixed first
3594 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3595
3596 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3597 starting at the last subject character, bytes beyond the end of the subject
3598 string were read.
3599
3600 11. Renamed the variable "class" as "classbits" to make life easier for C++
3601 users. (Previously there was a macro definition, but it apparently wasn't
3602 enough.)
3603
3604 12. Added the new field "tables" to the extra data so that tables can be passed
3605 in at exec time, or the internal tables can be re-selected. This allows
3606 a compiled regex to be saved and re-used at a later time by a different
3607 program that might have everything at different addresses.
3608
3609 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3610 -R library as well as a -L library.
3611
3612 14. The debugging options of pcretest (-d on the command line or D on a
3613 pattern) showed incorrect output for anything following an extended class
3614 that contained multibyte characters and which was followed by a quantifier.
3615
3616 15. Added optional support for general category Unicode character properties
3617 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3618 support. It adds about 90K to the size of the library. The meanings of the
3619 inbuilt class escapes such as \d and \s have NOT been changed.
3620
3621 16. Updated pcredemo.c to include calls to free() to release the memory for the
3622 compiled pattern.
3623
3624 17. The generated file chartables.c was being created in the source directory
3625 instead of in the building directory. This caused the build to fail if the
3626 source directory was different from the building directory, and was
3627 read-only.
3628
3629 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3630 file. No doubt somebody will tell me if they don't make sense... Also added
3631 Dan Mooney's comments about building on OpenVMS.
3632
3633 19. Added support for partial matching via the PCRE_PARTIAL option for
3634 pcre_exec() and the \P data escape in pcretest.
3635
3636 20. Extended pcretest with 3 new pattern features:
3637
3638 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3639 write the compiled pattern to the file whose name is "rest-of-line".
3640 This is a straight binary dump of the data, with the saved pointer to
3641 the character tables forced to be NULL. The study data, if any, is
3642 written too. After writing, pcretest reads a new pattern.
3643
3644 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3645 compiled pattern from the given file. There must not be any
3646 occurrences of "<" in the file name (pretty unlikely); if there are,
3647 pcretest will instead treat the initial "<" as a pattern delimiter.
3648 After reading in the pattern, pcretest goes on to read data lines as
3649 usual.
3650
3651 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3652 and 16-bit fields in a compiled pattern, to simulate a pattern that
3653 was compiled on a host of opposite endianness.
3654
3655 21. The pcre-exec() function can now cope with patterns that were compiled on
3656 hosts of opposite endianness, with this restriction:
3657
3658 As for any compiled expression that is saved and used later, the tables
3659 pointer field cannot be preserved; the extra_data field in the arguments
3660 to pcre_exec() should be used to pass in a tables address if a value
3661 other than the default internal tables were used at compile time.
3662
3663 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3664 now diagnosed as an error. Previously, most of the time, a negative number
3665 would have been treated as zero, but if in addition "ovector" was passed as
3666 NULL, a crash could occur.
3667
3668 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3669 new versions from the libtool 1.5 distribution (the last one is a copy of
3670 a file called libtool.m4). This seems to have fixed the need to patch
3671 "configure" to support Darwin 1.3 (which I used to do). However, I still
3672 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3673 workstation).
3674
3675 24. Changed the PCRE licence to be the more standard "BSD" licence.
3676
3677
3678 Version 4.5 01-Dec-03
3679 ---------------------
3680
3681 1. There has been some re-arrangement of the code for the match() function so
3682 that it can be compiled in a version that does not call itself recursively.
3683 Instead, it keeps those local variables that need separate instances for
3684 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3685 needs to "recurse". Keeping track of where control must go is done by means
3686 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3687 hide most of the details from the main code, and operates only if
3688 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3689 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3690 operating.
3691
3692 To make it easier for callers to provide specially tailored get/free
3693 functions for this usage, two new functions, pcre_stack_malloc, and
3694 pcre_stack_free, are used. They are always called in strict stacking order,
3695 and the size of block requested is always the same.
3696
3697 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3698 PCRE has been compiled to use the stack or the heap for recursion. The
3699 -C option of pcretest uses this to show which version is compiled.
3700
3701 A new data escape \S, is added to pcretest; it causes the amounts of store
3702 obtained and freed by both kinds of malloc/free at match time to be added
3703 to the output.
3704
3705 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3706 what's available on my current Linux desktop machine.
3707
3708 3. When matching a UTF-8 string, the test for a valid string at the start has
3709 been extended. If start_offset is not zero, PCRE now checks that it points
3710 to a byte that is the start of a UTF-8 character. If not, it returns
3711 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3712 this is necessary because there may be backward assertions in the pattern.
3713 When matching the same subject several times, it may save resources to use
3714 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3715
3716 4. The code for checking the validity of UTF-8 strings has been tightened so
3717 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3718 containing "overlong sequences".
3719
3720 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3721 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3722 should have been "&", but it just so happened that all the cases this let
3723 through by mistake were picked up later in the function.
3724
3725 6. I had used a variable called "isblank" - this is a C99 function, causing
3726 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3727
3728 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3729 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3730 so the version is shown; (c) stop "make test" echoing "./RunTest".
3731
3732 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3733
3734 9. The prototype for memmove() for systems that don't have it was using
3735 size_t, but the inclusion of the header that defines size_t was later. I've
3736 moved the #includes for the C headers earlier to avoid this.
3737
3738 10. Added some adjustments to the code to make it easier to compiler on certain
3739 special systems:
3740
3741 (a) Some "const" qualifiers were missing.
3742 (b) Added the macro EXPORT before all exported functions; by default this
3743 is defined to be empty.
3744 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3745 that it reads its output file name as an argument instead of writing
3746 to the standard output and assuming this can be redirected.
3747
3748 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3749 class containing characters with values greater than 255, PCRE compilation
3750 went into a loop.
3751
3752 12. A recursive reference to a subpattern that was within another subpattern
3753 that had a minimum quantifier of zero caused PCRE to crash. For example,
3754 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3755 recursion. If the recursively-called subpattern itself had a zero repeat,
3756 that was OK.
3757
3758 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3759 buffer into which it was copied (for escape processing) was still set at
3760 1024, so long lines caused crashes.
3761
3762 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3763 "internal error: code overflow...". This applied to any character class
3764 that was followed by a possessive quantifier.
3765
3766 15. Modified the Makefile to add libpcre.la as a prerequisite for
3767 libpcreposix.la because I was told this is needed for a parallel build to
3768 work.
3769
3770 16. If a pattern that contained .* following optional items at the start was
3771 studied, the wrong optimizing data was generated, leading to matching
3772 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3773 matching string must start with a or b or c. The correct conclusion for
3774 this pattern is that a match can start with any character.
3775
3776
3777 Version 4.4 13-Aug-03
3778 ---------------------
3779
3780 1. In UTF-8 mode, a character class containing characters with values between
3781 127 and 255 was not handled correctly if the compiled pattern was studied.
3782 In fixing this, I have also improved the studying algorithm for such
3783 classes (slightly).
3784
3785 2. Three internal functions had redundant arguments passed to them. Removal
3786 might give a very teeny performance improvement.
3787
3788 3. Documentation bug: the value of the capture_top field in a callout is *one
3789 more than* the number of the hightest numbered captured substring.
3790
3791 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3792 in incorrectly linking with a previously installed version. They now link
3793 explicitly with libpcre.la.
3794
3795 5. configure.in no longer needs to recognize Cygwin specially.
3796
3797 6. A problem in pcre.in for Windows platforms is fixed.
3798
3799 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3800 pcretest, it used to include the size of the study block as part of its
3801 output. Unfortunately, the structure contains a field that has a different
3802 size on different hardware architectures. This meant that the tests that
3803 showed this size failed. As the block is currently always of a fixed size,
3804 this information isn't actually particularly useful in pcretest output, so
3805 I have just removed it.
3806
3807 8. Three pre-processor statements accidentally did not start in column 1.
3808 Sadly, there are *still* compilers around that complain, even though
3809 standard C has not required this for well over a decade. Sigh.
3810
3811 9. In pcretest, the code for checking callouts passed small integers in the
3812 callout_data field, which is a void * field. However, some picky compilers
3813 complained about the casts involved for this on 64-bit systems. Now
3814 pcretest passes the address of the small integer instead, which should get
3815 rid of the warnings.
3816
3817 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3818 both compile and run time, and gives an error if an invalid UTF-8 sequence
3819 is found. There is a option for disabling this check in cases where the
3820 string is known to be correct and/or the maximum performance is wanted.
3821
3822 11. In response to a bug report, I changed one line in Makefile.in from
3823
3824 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3825 to
3826 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3827
3828 to look similar to other lines, but I have no way of telling whether this
3829 is the right thing to do, as I do not use Windows. No doubt I'll get told
3830 if it's wrong...
3831
3832
3833 Version 4.3 21-May-03
3834 ---------------------
3835
3836 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3837 Makefile.
3838
3839 2. Some refactoring to improve the quality of the code:
3840
3841 (i) The utf8_table... variables are now declared "const".
3842
3843 (ii) The code for \cx, which used the "case flipping" table to upper case
3844 lower case letters, now just substracts 32. This is ASCII-specific,
3845 but the whole concept of \cx is ASCII-specific, so it seems
3846 reasonable.
3847
3848 (iii) PCRE was using its character types table to recognize decimal and
3849 hexadecimal digits in the pattern. This is silly, because it handles
3850 only 0-9, a-f, and A-F, but the character types table is locale-
3851 specific, which means strange things might happen. A private
3852 table is now used for this - though it costs 256 bytes, a table is
3853 much faster than multiple explicit tests. Of course, the standard
3854 character types table is still used for matching digits in subject
3855 strings against \d.
3856
3857 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3858 ending in _t are). So I've renamed it as ESC_tee.
3859
3860 3. The first argument for regexec() in the POSIX wrapper should have been
3861 defined as "const".
3862
3863 4. Changed pcretest to use malloc() for its buffers so that they can be
3864 Electric Fenced for debugging.
3865
3866 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3867 to read one or more bytes before the start of the subject string. Often this
3868 had no effect on PCRE's behaviour, but in some circumstances it could
3869 provoke a segmentation fault.
3870
3871 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3872 to try to read one or more bytes before the start of the subject string.
3873
3874 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3875 UTF-8 support could misbehave in various ways if the subject string
3876 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3877 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3878 back over UTF-8 characters.)
3879
3880
3881 Version 4.2 14-Apr-03
3882 ---------------------
3883
3884 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3885
3886 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3887 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3888 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3889 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3890 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3891 and BUILD_EXEEXT
3892 Note: automatic setting of the BUILD variables is not yet working
3893 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3894 compile-time but not at link-time
3895 [LINK]: use for linking executables only
3896 make different versions for Windows and non-Windows
3897 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3898 libraries
3899 [LINK_FOR_BUILD]: new variable
3900 [OBJEXT]: use throughout
3901 [EXEEXT]: use throughout
3902 <winshared>: new target
3903 <wininstall>: new target
3904 <dftables.o>: use native compiler
3905 <dftables>: use native linker
3906 <install>: handle Windows platform correctly
3907 <clean>: ditto
3908 <check>: ditto
3909 copy DLL to top builddir before testing
3910
3911 As part of these changes, -no-undefined was removed again. This was reported
3912 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3913 in any case.
3914
3915 3. Some tidies to get rid of compiler warnings:
3916
3917 . In the match_data structure, match_limit was an unsigned long int, whereas
3918 match_call_count was an int. I've made them both unsigned long ints.
3919
3920 . In pcretest the fact that a const uschar * doesn't automatically cast to
3921 a void * provoked a warning.
3922
3923 . Turning on some more compiler warnings threw up some "shadow" variables
3924 and a few more missing casts.
3925
3926 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3927 option, a class that contained a single character with a value between 128
3928 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3929
3930 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3931 option, a class that contained several characters, but with at least one
3932 whose value was between 128 and 255 caused PCRE to crash.
3933
3934
3935 Version 4.1 12-Mar-03
3936 ---------------------
3937
3938 1. Compiling with gcc -pedantic found a couple of places where casts were
3939 needed, and a string in dftables.c that was longer than standard compilers are
3940 required to support.
3941
3942 2. Compiling with Sun's compiler found a few more places where the code could
3943 be tidied up in order to avoid warnings.
3944
3945 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3946 first of these names is deprecated in the latest Autoconf in favour of the name
3947 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3948 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3949 analogy I have changed it to CFLAGS_FOR_BUILD.
3950
3951 4. Added -no-undefined to the linking command in the Makefile, because this is
3952 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3953 linking step for the pcreposix library.
3954
3955 5. PCRE was failing to diagnose the case of two named groups with the same
3956 name.
3957
3958 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3959 literal character that is needed in the subject for a match, and scans along to
3960 ensure that it is present before embarking on the full matching process. This
3961 saves time in cases of nested unlimited repeats that are never going to match.
3962 Problem: the scan can take a lot of time if the subject is very long (e.g.
3963 megabytes), thus penalizing straightforward matches. It is now done only if the
3964 amount of subject to be scanned is less than 1000 bytes.
3965
3966 7. A lesser problem with the same optimization is that it was recording the
3967 first character of an anchored pattern as "needed", thus provoking a search
3968 right along the subject, even when the first match of the pattern was going to
3969 fail. The "needed" character is now not set for anchored patterns, unless it
3970 follows something in the pattern that is of non-fixed length. Thus, it still
3971 fulfils its original purpose of finding quick non-matches in cases of nested
3972 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3973
3974
3975 Version 4.0 17-Feb-03
3976 ---------------------
3977
3978 1. If a comment in an extended regex that started immediately after a meta-item
3979 extended to the end of string, PCRE compiled incorrect data. This could lead to
3980 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3981
3982 2. Moved to autoconf 2.53 and libtool 1.4.2.
3983
3984 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3985 the special perltest8 script is no longer needed - all the tests can be run
3986 from a single perltest script.
3987
3988 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3989 by \s. It has now been removed in PCRE. This means it isn't recognized as
3990 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3991 class [:space:] *does* include VT, thereby creating a mess.
3992
3993 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3994 space and tab.
3995
3996 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3997 its new features into the main test script, reducing the number of scripts.
3998
3999 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4000 were backward compatible, and made the (?i) apply to the whole pattern, as if
4001 /i were given. Now it behaves more logically, and applies the option setting
4002 only to what follows. PCRE has been changed to follow suit. However, if it
4003 finds options settings right at the start of the pattern, it extracts them into
4004 the global options, as before. Thus, they show up in the info data.
4005
4006 8. Added support for the \Q...\E escape sequence. Characters in between are
4007 treated as literals. This is slightly different from Perl in that $ and @ are
4008 also handled as literals inside the quotes. In Perl, they will cause variable
4009 interpolation. Note the following examples:
4010
4011 Pattern PCRE matches Perl matches
4012
4013 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4014 \Qabc\$xyz\E abc\$xyz abc\$xyz
4015 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4016
4017 For compatibility with Perl, \Q...\E sequences are recognized inside character
4018 classes as well as outside them.
4019
4020 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4021 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4022 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4023 signed/unsigned warnings.
4024
4025 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4026 option for pcretest, so I've replaced it by a simple function that does just
4027 that job.
4028
4029 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4030 "pcregrep -".
4031
4032 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4033 Java package. This provides some syntactic sugar for simple cases of what my
4034 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4035 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4036 item, you can use this simplified notation. Note that only makes sense with
4037 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4038 greediness, whatever the setting of the PCRE_UNGREEDY option.
4039
4040 13. A change of greediness default within a pattern was not taking effect at
4041 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4042 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4043 was abstracted outside.
4044
4045 14. PCRE now supports the \G assertion. It is true when the current matching
4046 position is at the start point of the match. This differs from \A when the
4047 starting offset is non-zero. Used with the /g option of pcretest (or similar
4048 code), it works in the same way as it does for Perl's /g option. If all
4049 alternatives of a regex begin with \G, the expression is anchored to the start
4050 match position, and the "anchored" flag is set in the compiled expression.
4051
4052 15. Some bugs concerning the handling of certain option changes within patterns
4053 have been fixed. These applied to options other than (?ims). For example,
4054 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4055 been the other way round. Some of this was related to change 7 above.
4056
4057 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4058 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4059 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4060 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4061
4062 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4063 mode. Unlike ".", it always matches newline, whatever the setting of
4064 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4065 assertions. Perl allows it, but it doesn't (in general) work because it can't
4066 calculate the length of the lookbehind. At least, that's the case for Perl
4067 5.8.0 - I've been told they are going to document that it doesn't work in
4068 future.
4069
4070 18. Added an error diagnosis for escapes that PCRE does not support: these are
4071 \L, \l, \N, \P, \p, \U, \u, and \X.
4072
4073 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4074 reading past the end of the pattern in cases such as /[abcd/.
4075
4076 20. PCRE was getting more memory than necessary for patterns with classes that
4077 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4078
4079 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4080 compiling PCRE for use with Virtual Pascal.
4081
4082 22. Small fix to the Makefile to make it work properly if the build is done
4083 outside the source tree.
4084
4085 23. Added a new extension: a condition to go with recursion. If a conditional
4086 subpattern starts with (?(R) the "true" branch is used if recursion has
4087 happened, whereas the "false" branch is used only at the top level.
4088
4089 24. When there was a very long string of literal characters (over 255 bytes
4090 without UTF support, over 250 bytes with UTF support), the computation of how
4091 much memory was required could be incorrect, leading to segfaults or other
4092 strange effects.
4093
4094 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4095 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4096 there was a subsequent back reference to those brackets. This meant that, for
4097 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4098 possible to check for precisely this case. All we can do is abandon the
4099 optimization if .* occurs inside capturing brackets when there are any back
4100 references whatsoever. (See below for a better fix that came later.)
4101
4102 26. The handling of the optimization for finding the first character of a
4103 non-anchored pattern, and for finding a character that is required later in the
4104 match were failing in some cases. This didn't break the matching; it just
4105 failed to optimize when it could. The way this is done has been re-implemented.
4106
4107 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4108
4109 28. Added a new feature that provides some of the functionality that Perl
4110 provides with (?{...}). The facility is termed a "callout". The way it is done
4111 in PCRE is for the caller to provide an optional function, by setting
4112 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4113 global variable. By default it is unset, which disables all calling out. To get
4114 the function called, the regex must include (?C) at appropriate points. This
4115 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4116 This provides a means of identifying different callout points. When PCRE
4117 reaches such a point in the regex, if pcre_callout has been set, the external
4118 function is called. It is provided with data in a structure called
4119 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4120 matching continues; if it returns a non-zero value, the match at the current
4121 point fails. However, backtracking will occur if possible. [This was changed
4122 later and other features added - see item 49 below.]
4123
4124 29. pcretest is upgraded to test the callout functionality. It provides a
4125 callout function that displays information. By default, it shows the start of
4126 the match and the current position in the text. There are some new data escapes
4127 to vary what happens:
4128
4129 \C+ in addition, show current contents of captured substrings
4130 \C- do not supply a callout function
4131 \C!n return 1 when callout number n is reached
4132 \C!n!m return 1 when callout number n is reached for the mth time
4133
4134 30. If pcregrep was called with the -l option and just a single file name, it
4135 output "<stdin>" if a match was found, instead of the file name.
4136
4137 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4138 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4139 pcre_exec(). This saves a malloc/free per call. The default value of
4140 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4141 when configuring.
4142
4143 32. The default maximum size of a compiled pattern is 64K. There have been a
4144 few cases of people hitting this limit. The code now uses macros to handle the
4145 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4146 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4147 configuring. Tests 2 and 5 work only with 2-byte links because they output
4148 debugging information about compiled patterns.
4149
4150 33. Internal code re-arrangements:
4151
4152 (a) Moved the debugging function for printing out a compiled regex into
4153 its own source file (printint.c) and used #include to pull it into
4154 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4155 separate copies.
4156
4157 (b) Defined the list of op-code names for debugging as a macro in
4158 internal.h so that it is next to the definition of the opcodes.
4159
4160 (c) Defined a table of op-code lengths for simpler skipping along compiled
4161 code. This is again a macro in internal.h so that it is next to the
4162 definition of the opcodes.
4163
4164 34. Added support for recursive calls to individual subpatterns, along the
4165 lines of Robin Houston's patch (but implemented somewhat differently).
4166
4167 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4168 allow it to read and process whole directories in Win32. This code was
4169 contributed by Lionel Fourquaux; it has not been tested by me.
4170
4171 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4172 used to name a group. Names consist of alphanumerics and underscores, and must
4173 be unique. Back references use the syntax (?P=name) and recursive calls use
4174 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4175 numbers. The function pcre_fullinfo() can be used after compilation to extract
4176 a name/number map. There are three relevant calls:
4177
4178 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4179 PCRE_INFO_NAMECOUNT yields the number of entries
4180 PCRE_INFO_NAMETABLE yields a pointer to the map.
4181
4182 The map is a vector of fixed-size entries. The size of each entry depends on
4183 the length of the longest name used. The first two bytes of each entry are the
4184 group number, most significant byte first. There follows the corresponding
4185 name, zero terminated. The names are in alphabetical order.
4186
4187 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4188 case instead of 255. Making it the same both with and without UTF-8 support
4189 means that the same test output works with both.
4190
4191 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4192 calling malloc() with a zero argument.
4193
4194 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4195 optimization. I've improved things by keeping a bitmap of backreferences with
4196 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4197 fact referenced, the optimization can be applied. It is unlikely that a
4198 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4199 the match to follow \n) will appear inside brackets with a number greater than
4200 31, but if it does, any back reference > 31 suppresses the optimization.
4201
4202 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4203 of disabling numbered capturing parentheses. Any opening parenthesis that is
4204 not followed by ? behaves as if it were followed by ?: but named parentheses
4205 can still be used for capturing (and they will acquire numbers in the usual
4206 way).
4207
4208 41. Redesigned the return codes from the match() function into yes/no/error so
4209 that errors can be passed back from deep inside the nested calls. A malloc
4210 failure while inside a recursive subpattern call now causes the
4211 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4212
4213 42. It is now possible to set a limit on the number of times the match()
4214 function is called in a call to pcre_exec(). This facility makes it possible to
4215 limit the amount of recursion and backtracking, though not in a directly
4216 obvious way, because the match() function is used in a number of different
4217 circumstances. The count starts from zero for each position in the subject
4218 string (for non-anchored patterns). The default limit is, for compatibility, a
4219 large number, namely 10 000 000. You can change this in two ways:
4220
4221 (a) When configuring PCRE before making, you can use --with-match-limit=n
4222 to set a default value for the compiled library.
4223
4224 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4225 a different value is set. See 45 below.
4226
4227 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4228
4229 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4230 of things that can be changed at compile time. The first argument specifies
4231 what is wanted and the second points to where the information is to be placed.
4232 The current list of available information is:
4233
4234 PCRE_CONFIG_UTF8
4235
4236 The output is an integer that is set to one if UTF-8 support is available;
4237 otherwise it is set to zero.
4238
4239 PCRE_CONFIG_NEWLINE
4240
4241 The output is an integer that it set to the value of the code that is used for
4242 newline. It is either LF (10) or CR (13).
4243
4244 PCRE_CONFIG_LINK_SIZE
4245
4246 The output is an integer that contains the number of bytes used for internal
4247 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4248
4249 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4250
4251 The output is an integer that contains the threshold above which the POSIX
4252 interface uses malloc() for output vectors. See item 31 above.
4253
4254 PCRE_CONFIG_MATCH_LIMIT
4255
4256 The output is an unsigned integer that contains the default limit of the number
4257 of match() calls in a pcre_exec() execution. See 42 above.
4258
4259 44. pcretest has been upgraded by the addition of the -C option. This causes it
4260 to extract all the available output from the new pcre_config() function, and to
4261 output it. The program then exits immediately.
4262
4263 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4264 order to support additional features. One way would have been to define
4265 pcre_exec2() (for example) with extra arguments, but this would not have been
4266 extensible, and would also have required all calls to the original function to
4267 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4268 is used for passing in "extra" data from pcre_study().
4269
4270 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4271 contains the following fields:
4272
4273 flags a bitmap indicating which of the following fields are set
4274 study_data opaque data from pcre_study()
4275 match_limit a way of specifying a limit on match() calls for a specific
4276 call to pcre_exec()
4277 callout_data data for callouts (see 49 below)
4278
4279 The flag bits are also defined in pcre.h, and are
4280
4281 PCRE_EXTRA_STUDY_DATA
4282 PCRE_EXTRA_MATCH_LIMIT
4283 PCRE_EXTRA_CALLOUT_DATA
4284
4285 The pcre_study() function now returns one of these new pcre_extra blocks, with
4286 the actual study data pointed to by the study_data field, and the
4287 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4288 before. That is, this change is entirely upwards-compatible and requires no
4289 change to existing code.
4290
4291 If you want to pass in additional data to pcre_exec(), you can either place it
4292 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4293 block.
4294
4295 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4296 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4297 times with different match limits, until it finds the minimum value needed for
4298 pcre_exec() to complete. The value is then output. This can be instructive; for
4299 most simple matches the number is quite small, but for pathological cases it
4300 gets very large very quickly.
4301
4302 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4303 returns the size of the data block pointed to by the study_data field in a
4304 pcre_extra block, that is, the value that was passed as the argument to
4305 pcre_malloc() when PCRE was getting memory in which to place the information
4306 created by pcre_study(). The fourth argument should point to a size_t variable.
4307 pcretest has been extended so that this information is shown after a successful
4308 pcre_study() call when information about the compiled regex is being displayed.
4309
4310 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4311 because what follows is always an absolute path. (Later: it turns out that this
4312 is more than cosmetic for MinGW, because it doesn't like empty path
4313 components.)
4314
4315 49. Some changes have been made to the callout feature (see 28 above):
4316
4317 (i) A callout function now has three choices for what it returns:
4318
4319 0 => success, carry on matching
4320 > 0 => failure at this point, but backtrack if possible
4321 < 0 => serious error, return this value from pcre_exec()
4322
4323 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4324 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4325 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4326 use by callout functions. It will never be used by PCRE itself.
4327
4328 (ii) The pcre_extra structure (see 45 above) has a void * field called
4329 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4330 pcre_callout_block structure has a field of the same name. The contents of
4331 the field passed in the pcre_extra structure are passed to the callout
4332 function in the corresponding field in the callout block. This makes it
4333 easier to use the same callout-containing regex from multiple threads. For
4334 testing, the pcretest program has a new data escape
4335
4336 \C*n pass the number n (may be negative) as callout_data
4337
4338 If the callout function in pcretest receives a non-zero value as
4339 callout_data, it returns that value.
4340
4341 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4342 there were some redundant $(CFLAGS) in commands that are now specified as
4343 $(LINK), which already includes $(CFLAGS).
4344
4345 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4346 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4347 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4348 one-byte characters throughout. Note that case-insensitive matching applies
4349 only to characters whose values are less than 256. PCRE doesn't support the
4350 notion of cases for higher-valued characters.
4351
4352 (i) A character class whose characters are all within 0-255 is handled as
4353 a bit map, and the map is inverted for negative classes. Previously, a
4354 character > 255 always failed to match such a class; however it should
4355 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4356
4357 (ii) A negated character class with a single character < 255 is coded as
4358 "not this character" (OP_NOT). This wasn't working properly when the test
4359 character was multibyte, either singly or repeated.
4360
4361 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4362 mode, for example: \x{100}{2,3}.
4363
4364 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4365 singly or repeated) now correctly test multibyte characters. However,
4366 PCRE doesn't recognize any characters with values greater than 255 as
4367 digits, spaces, or word characters. Such characters always match \D, \S,
4368 and \W, and never match \d, \s, or \w.
4369
4370 (v) Classes may now contain characters and character ranges with values
4371 greater than 255. For example: [ab\x{100}-\x{400}].
4372
4373 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4374 PCRE in UTF-8 mode.
4375
4376 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4377 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4378 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4379 value.)
4380
4381 53. The single man page has become too large. I have therefore split it up into
4382 a number of separate man pages. These also give rise to individual HTML pages;
4383 these are now put in a separate directory, and there is an index.html page that
4384 lists them all. Some hyperlinking between the pages has been installed.
4385
4386 54. Added convenience functions for handling named capturing parentheses.
4387
4388 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4389 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4390 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4391 are faulted.
4392
4393 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4394 calling configure. These values are used when compiling the dftables.c program
4395 which is run to generate the source of the default character tables. They
4396 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4397 you will need to set these values.
4398
4399 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4400
4401
4402 Version 3.9 02-Jan-02
4403 ---------------------
4404
4405 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4406
4407 2. If --disable-static was given, the building process failed when trying to
4408 build pcretest and pcregrep. (For some reason it was using libtool to compile
4409 them, which is not right, as they aren't part of the library.)
4410
4411
4412 Version 3.8 18-Dec-01
4413 ---------------------
4414
4415 1. The experimental UTF-8 code was completely screwed up. It was packing the
4416 bytes in the wrong order. How dumb can you get?
4417
4418
4419 Version 3.7 29-Oct-01
4420 ---------------------
4421
4422 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4423 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4424 this didn't happen under Solaris 8, where I normally test things.
4425
4426 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4427 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4428 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4429 this fix an hour or so after the initial 3.7 release.)
4430
4431
4432 Version 3.6 23-Oct-01
4433 ---------------------
4434
4435 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4436 offsets passed as NULL with zero offset count.
4437
4438 2. The config.guess and config.sub files had not been updated when I moved to
4439 the latest autoconf.
4440
4441
4442 Version 3.5 15-Aug-01
4443 ---------------------
4444
4445 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4446 had been forgotten.
4447
4448 2. By using declared but undefined structures, we can avoid using "void"
4449 definitions in pcre.h while keeping the internal definitions of the structures
4450 private.
4451
4452 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4453 user point of view, this means that both static and shared libraries are built
4454 by default, but this can be individually controlled. More of the work of
4455 handling this static/shared cases is now inside libtool instead of PCRE's make
4456 file.
4457
4458 4. The pcretest utility is now installed along with pcregrep because it is
4459 useful for users (to test regexs) and by doing this, it automatically gets
4460 relinked by libtool. The documentation has been turned into a man page, so
4461 there are now .1, .txt, and .html versions in /doc.
4462
4463 5. Upgrades to pcregrep:
4464 (i) Added long-form option names like gnu grep.
4465 (ii) Added --help to list all options with an explanatory phrase.
4466 (iii) Added -r, --recursive to recurse into sub-directories.
4467 (iv) Added -f, --file to read patterns from a file.
4468
4469 6. pcre_exec() was referring to its "code" argument before testing that
4470 argument for NULL (and giving an error if it was NULL).
4471
4472 7. Upgraded Makefile.in to allow for compiling in a different directory from
4473 the source directory.
4474
4475 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4476 options bits, the pointer it was passed was to an int instead of to an unsigned
4477 long int. This mattered only on 64-bit systems.
4478
4479 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4480 generated) instead of pcre.in, which it its source. Also made the same change
4481 in several of the .c files.
4482
4483 10. A new release of gcc defines printf() as a macro, which broke pcretest
4484 because it had an ifdef in the middle of a string argument for printf(). Fixed
4485 by using separate calls to printf().
4486
4487 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4488 script, to force use of CR or LF instead of \n in the source. On non-Unix
4489 systems, the value can be set in config.h.
4490
4491 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4492 absolute limit. Changed the text of the error message to make this clear, and
4493 likewise updated the man page.
4494
4495 13. The limit of 99 on the number of capturing subpatterns has been removed.
4496 The new limit is 65535, which I hope will not be a "real" limit.
4497
4498
4499 Version 3.4 22-Aug-00
4500 ---------------------
4501
4502 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4503
4504 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4505
4506
4507 Version 3.3 01-Aug-00
4508 ---------------------
4509
4510 1. If an octal character was given, but the value was greater than \377, it
4511 was not getting masked to the least significant bits, as documented. This could
4512 lead to crashes in some systems.
4513
4514 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4515 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4516
4517 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4518 These just pass their arguments on to (pcre_free)(), but they are provided
4519 because some uses of PCRE bind it to non-C systems that can call its functions,
4520 but cannot call free() or pcre_free() directly.
4521
4522 4. Add "make test" as a synonym for "make check". Corrected some comments in
4523 the Makefile.
4524
4525 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4526 Makefile.
4527
4528 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4529 command called pgrep for grepping around the active processes.
4530
4531 7. Added the beginnings of support for UTF-8 character strings.
4532
4533 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4534 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4535 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4536 out for the ar command.)
4537
4538
4539 Version 3.2 12-May-00
4540 ---------------------
4541
4542 This is purely a bug fixing release.
4543
4544 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4545 of ZA. This was just one example of several cases that could provoke this bug,
4546 which was introduced by change 9 of version 2.00. The code for breaking
4547 infinite loops after an iteration that matches an empty string was't working
4548 correctly.
4549
4550 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4551 when matched against abbab (for example). After matching an empty string, it
4552 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4553 caused it to match further down the string than it should.
4554
4555 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4556 was there because it doesn't seem to be needed, and it causes trouble on some
4557 systems, as it is not a Standard C header. It has been removed.
4558
4559 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4560 were reported on the Macintosh. The changes were from
4561
4562 while ((c = *(++ptr)) != 0 && c != '\n');
4563 to
4564 while ((c = *(++ptr)) != 0 && c != '\n') ;
4565
4566 Totally extraordinary, but if that's what it takes...
4567
4568 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4569 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4570 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4571 assumes the way PCRE uses memmove() (always moving upwards).
4572
4573 6. PCRE is being used in one environment where strchr() is not available. There
4574 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4575 faster code anyway.
4576
4577
4578 Version 3.1 09-Feb-00
4579 ---------------------
4580
4581 The only change in this release is the fixing of some bugs in Makefile.in for
4582 the "install" target:
4583
4584 (1) It was failing to install pcreposix.h.
4585
4586 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4587
4588
4589 Version 3.0 01-Feb-00
4590 ---------------------
4591
4592 1. Add support for the /+ modifier to perltest (to output $` like it does in
4593 pcretest).
4594
4595 2. Add support for the /g modifier to perltest.
4596
4597 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4598 matches null strings.
4599
4600 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4601 pattern. Perl treats empty patterns specially - it reuses the most recent
4602 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4603 effect.
4604
4605 5. The POSIX interface was broken in that it was just handing over the POSIX
4606 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4607 required a bigger vector, with some working space on the end. This means that
4608 the POSIX wrapper now has to get and free some memory, and copy the results.
4609
4610 6. Added some simple autoconf support, placing the test data and the
4611 documentation in separate directories, re-organizing some of the
4612 information files, and making it build pcre-config (a GNU standard). Also added
4613 libtool support for building PCRE as a shared library, which is now the
4614 default.
4615
4616 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4617 09 are not valid octal constants. Single digits will be used for minor values
4618 less than 10.
4619
4620 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4621 existing programs that set these in the POSIX interface can use PCRE without
4622 modification.
4623
4624 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4625 return all that pcre_info() returns, plus additional data. The pcre_info()
4626 function is retained for compatibility, but is considered to be obsolete.
4627
4628 10. Added experimental recursion feature (?R) to handle one common case that
4629 Perl 5.6 will be able to do with (?p{...}).
4630
4631 11. Added support for POSIX character classes like [:alpha:], which Perl is
4632 adopting.
4633
4634
4635 Version 2.08 31-Aug-99
4636 ----------------------
4637
4638 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4639 trying to match at the startoffset position, but instead was moving forward to
4640 the next newline as if a previous match had failed.
4641
4642 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4643 and could get into a loop if a null string was matched other than at the start
4644 of the subject.
4645
4646 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4647 be distinguished at compile time, and for completeness also added PCRE_DATE.
4648
4649 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4650 in GnuWin32 environments.
4651
4652
4653 Version 2.07 29-Jul-99
4654 ----------------------
4655
4656 1. The documentation is now supplied in plain text form and HTML as well as in
4657 the form of man page sources.
4658
4659 2. C++ compilers don't like assigning (void *) values to other pointer types.
4660 In particular this affects malloc(). Although there is no problem in Standard
4661 C, I've put in casts to keep C++ compilers happy.
4662
4663 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4664 should be (const char *).
4665
4666 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4667 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4668 However, I haven't made this a standard facility. The documentation doesn't
4669 mention it, and the Makefile doesn't support it.
4670
4671 5. The Makefile now contains an "install" target, with editable destinations at
4672 the top of the file. The pcretest program is not installed.
4673
4674 6. pgrep -V now gives the PCRE version number and date.
4675
4676 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4677 causing the entire string to be ignored, instead of just the last character.
4678
4679 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4680 non-matching string, it can take a very, very long time, even for strings of
4681 quite modest length, because of the nested recursion. PCRE now does better in
4682 some of these cases. It does this by remembering the last required literal
4683 character in the pattern, and pre-searching the subject to ensure it is present
4684 before running the real match. In other words, it applies a heuristic to detect
4685 some types of certain failure quickly, and in the above example, if presented
4686 with a string that has no trailing " it gives "no match" very quickly.
4687
4688 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4689 other alternatives are tried instead.
4690
4691
4692 Version 2.06 09-Jun-99
4693 ----------------------
4694
4695 1. Change pcretest's output for amount of store used to show just the code
4696 space, because the remainder (the data block) varies in size between 32-bit and
4697 64-bit systems.
4698
4699 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4700 start matching at. This allows lookbehinds to work when searching for multiple
4701 occurrences in a string.
4702
4703 3. Added additional options to pcretest for testing multiple occurrences:
4704
4705 /+ outputs the rest of the string that follows a match
4706 /g loops for multiple occurrences, using the new startoffset argument
4707 /G loops for multiple occurrences by passing an incremented pointer
4708
4709 4. PCRE wasn't doing the "first character" optimization for patterns starting
4710 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4711 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4712 the letter 'x'. On long subject strings, this gives a significant speed-up.
4713
4714
4715 Version 2.05 21-Apr-99
4716 ----------------------
4717
4718 1. Changed the type of magic_number from int to long int so that it works
4719 properly on 16-bit systems.
4720
4721 2. Fixed a bug which caused patterns starting with .* not to work correctly
4722 when the subject string contained newline characters. PCRE was assuming
4723 anchoring for such patterns in all cases, which is not correct because .* will
4724 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4725 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4726 must be retried after every newline in the subject.
4727
4728
4729 Version 2.04 18-Feb-99
4730 ----------------------
4731
4732 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4733 computation of the store needed to hold the pattern was incorrect (too large).
4734 If such patterns were nested a few deep, this could multiply and become a real
4735 problem.
4736
4737 2. Added /M option to pcretest to show the memory requirement of a specific
4738 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4739
4740 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4741 compiled in such a way that the backtracking after subsequent failure was
4742 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4743 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4744
4745
4746 Version 2.03 02-Feb-99
4747 ----------------------
4748
4749 1. Fixed typo and small mistake in man page.
4750
4751 2. Added 4th condition (GPL supersedes if conflict) and created separate
4752 LICENCE file containing the conditions.
4753
4754 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4755 Perl, that is the internal \ allows the delimiter to be included in the
4756 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4757 the final delimiter, add \ to the end of the pattern (to test the error).
4758
4759 4. Added the convenience functions for extracting substrings after a successful
4760 match. Updated pcretest to make it able to test these functions.
4761
4762
4763 Version 2.02 14-Jan-99
4764 ----------------------
4765
4766 1. Initialized the working variables associated with each extraction so that
4767 their saving and restoring doesn't refer to uninitialized store.
4768
4769 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4770 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4771 fix the problem.
4772
4773 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4774 calls, and wasn't printing the correct value for compiling calls. Increased the
4775 default value of LOOPREPEAT, and the number of significant figures in the
4776 times.
4777
4778 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4779
4780 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4781 a building problem on Windows NT with a FAT file system.
4782
4783
4784 Version 2.01 21-Oct-98
4785 ----------------------
4786
4787 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4788 to character tables built by pcre_maketables() in the current locale. If NULL
4789 is passed, the default tables are used.
4790
4791
4792 Version 2.00 24-Sep-98
4793 ----------------------
4794
4795 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4796 it any more.
4797
4798 2. Allow quantification of (?>) groups, and make it work correctly.
4799
4800 3. The first character computation wasn't working for (?>) groups.
4801
4802 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4803 end of the subject) and add 5.005's \z, which really does match only at the
4804 very end of the subject.
4805
4806 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4807
4808 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4809 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4810 localized options. All options to pcre_study() were also removed.
4811
4812 7. Add other new features from 5.005:
4813
4814 $(?<= positive lookbehind
4815 $(?<! negative lookbehind
4816 (?imsx-imsx) added the unsetting capability
4817 such a setting is global if at outer level; local otherwise
4818 (?imsx-imsx:) non-capturing groups with option setting
4819 (?(cond)re|re) conditional pattern matching
4820
4821 A backreference to itself in a repeated group matches the previous
4822 captured string.
4823
4824 8. General tidying up of studying (both automatic and via "study")
4825 consequential on the addition of new assertions.
4826
4827 9. As in 5.005, unlimited repeated groups that could match an empty substring
4828 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4829 runtime if any iteration does actually match an empty substring.
4830
4831 10. Include the RunTest script in the distribution.
4832
4833 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4834 discrepancies, some of which were old and were also with respect to 5.004. They
4835 have now been fixed.
4836
4837
4838 Version 1.09 28-Apr-98
4839 ----------------------
4840
4841 1. A negated single character class followed by a quantifier with a minimum
4842 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4843 program crashes, or just wrong answers. This did not apply to negated classes
4844 containing more than one character, or to minima other than one.
4845
4846
4847 Version 1.08 27-Mar-98
4848 ----------------------
4849
4850 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4851
4852 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4853 latter must appear before anything that relies on it in the pattern.
4854
4855
4856 Version 1.07 16-Feb-98
4857 ----------------------
4858
4859 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4860 repeat of a potentially empty string).
4861
4862
4863 Version 1.06 23-Jan-98
4864 ----------------------
4865
4866 1. Added Markus Oberhumer's little patches for C++.
4867
4868 2. Literal strings longer than 255 characters were broken.
4869
4870
4871 Version 1.05 23-Dec-97
4872 ----------------------
4873
4874 1. Negated character classes containing more than one character were failing if
4875 PCRE_CASELESS was set at run time.
4876
4877
4878 Version 1.04 19-Dec-97
4879 ----------------------
4880
4881 1. Corrected the man page, where some "const" qualifiers had been omitted.
4882
4883 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4884 input syntax.
4885
4886 3. Fixed memory leak which occurred when a regex with back references was
4887 matched with an offsets vector that wasn't big enough. The temporary memory
4888 that is used in this case wasn't being freed if the match failed.
4889
4890 4. Tidied pcretest to ensure it frees memory that it gets.
4891
4892 5. Temporary memory was being obtained in the case where the passed offsets
4893 vector was exactly big enough.
4894
4895 6. Corrected definition of offsetof() from change 5 below.
4896
4897 7. I had screwed up change 6 below and broken the rules for the use of
4898 setjmp(). Now fixed.
4899
4900
4901 Version 1.03 18-Dec-97
4902 ----------------------
4903
4904 1. A erroneous regex with a missing opening parenthesis was correctly
4905 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4906 on some systems.
4907
4908 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4909 it was reported that one broken compiler failed on the former because "code" is
4910 also an independent variable.
4911
4912 3. The erroneous regex a[]b caused an array overrun reference.
4913
4914 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4915 fail on data ending with that character. (It was going on too far, and checking
4916 the next character, typically a binary zero.) This was specific to the
4917 optimized code for single-character negative classes.
4918
4919 5. Added a contributed patch from the TIN world which does the following:
4920
4921 + Add an undef for memmove, in case the the system defines a macro for it.
4922
4923 + Add a definition of offsetof(), in case there isn't one. (I don't know
4924 the reason behind this - offsetof() is part of the ANSI standard - but
4925 it does no harm).
4926
4927 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4928 most of the places where whitespace preceded '#'. I have given up and
4929 allowed the remaining 2 cases to be at the margin.
4930
4931 + Rename some variables in pcre to eliminate shadowing. This seems very
4932 pedantic, but does no harm, of course.
4933
4934 6. Moved the call to setjmp() into its own function, to get rid of warnings
4935 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4936
4937 7. Constructs such as \d{8,} were compiling into the equivalent of
4938 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4939 outcome, but in this particular case used more store than had been allocated,
4940 which caused the bug to be discovered because it threw up an internal error.
4941
4942 8. The debugging code in both pcre and pcretest for outputting the compiled
4943 form of a regex was going wrong in the case of back references followed by
4944 curly-bracketed repeats.
4945
4946
4947 Version 1.02 12-Dec-97
4948 ----------------------
4949
4950 1. Typos in pcre.3 and comments in the source fixed.
4951
4952 2. Applied a contributed patch to get rid of places where it used to remove
4953 'const' from variables, and fixed some signed/unsigned and uninitialized
4954 variable warnings.
4955
4956 3. Added the "runtest" target to Makefile.
4957
4958 4. Set default compiler flag to -O2 rather than just -O.
4959
4960
4961 Version 1.01 19-Nov-97
4962 ----------------------
4963
4964 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4965 like /([ab]*)*/, that is, for classes with more than one character in them.
4966
4967 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4968 as /((?>a*))*/ (a PCRE_EXTRA facility).
4969
4970
4971 Version 1.00 18-Nov-97
4972 ----------------------
4973
4974 1. Added compile-time macros to support systems such as SunOS4 which don't have
4975 memmove() or strerror() but have other things that can be used instead.
4976
4977 2. Arranged that "make clean" removes the executables.
4978
4979
4980 Version 0.99 27-Oct-97
4981 ----------------------
4982
4983 1. Fixed bug in code for optimizing classes with only one character. It was
4984 initializing a 32-byte map regardless, which could cause it to run off the end
4985 of the memory it had got.
4986
4987 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4988
4989
4990 Version 0.98 22-Oct-97
4991 ----------------------
4992
4993 1. Fixed bug in code for handling temporary memory usage when there are more
4994 back references than supplied space in the ovector. This could cause segfaults.
4995
4996
4997 Version 0.97 21-Oct-97
4998 ----------------------
4999
5000 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5001
5002 2. Optimized negated single characters not to use a bit map.
5003
5004 3. Brought error texts together as macro definitions; clarified some of them;
5005 fixed one that was wrong - it said "range out of order" when it meant "invalid
5006 escape sequence".
5007
5008 4. Changed some char * arguments to const char *.
5009
5010 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5011
5012 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5013 pcretest.
5014
5015
5016 Version 0.96 16-Oct-97
5017 ----------------------
5018
5019 1. Added a simple "pgrep" utility to the distribution.
5020
5021 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5022 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5023 where "ddd" means "one or more decimal digits".
5024
5025 3. Fixed serious bug. If a pattern had a back reference, but the call to
5026 pcre_exec() didn't supply a large enough ovector to record the related
5027 identifying subpattern, the match always failed. PCRE now remembers the number
5028 of the largest back reference, and gets some temporary memory in which to save
5029 the offsets during matching if necessary, in order to ensure that
5030 backreferences always work.
5031
5032 4. Increased the compatibility with Perl in a number of ways:
5033
5034 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5035 to request this handling. The option can be set at compile or exec time.
5036
5037 (b) $ matches before a terminating newline by default; an option
5038 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5039 mode). The option can be set at compile or exec time.
5040
5041 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5042 the same as Perl's. If the decimal number it represents is less than 10
5043 or there aren't that many previous left capturing parentheses, an octal
5044 escape is read. Inside a character class, it's always an octal escape,
5045 even if it is a single digit.
5046
5047 (d) An escaped but undefined alphabetic character is taken as a literal,
5048 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5049 escapes.
5050
5051 (e) {0} is now permitted. (The previous item is removed from the compiled
5052 pattern).
5053
5054 5. Changed all the names of code files so that the basic parts are no longer
5055 than 10 characters, and abolished the teeny "globals.c" file.
5056
5057 6. Changed the handling of character classes; they are now done with a 32-byte
5058 bit map always.
5059
5060 7. Added the -d and /D options to pcretest to make it possible to look at the
5061 internals of compilation without having to recompile pcre.
5062
5063
5064 Version 0.95 23-Sep-97
5065 ----------------------
5066
5067 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5068 \x20 at the start of a run of normal characters. These were being treated as
5069 real characters, instead of the source characters being re-checked.
5070
5071
5072 Version 0.94 18-Sep-97
5073 ----------------------
5074
5075 1. The functions are now thread-safe, with the caveat that the global variables
5076 containing pointers to malloc() and free() or alternative functions are the
5077 same for all threads.
5078
5079 2. Get pcre_study() to generate a bitmap of initial characters for non-
5080 anchored patterns when this is possible, and use it if passed to pcre_exec().
5081
5082
5083 Version 0.93 15-Sep-97
5084 ----------------------
5085
5086 1. /(b)|(:+)/ was computing an incorrect first character.
5087
5088 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5089 but not actually doing anything yet.
5090
5091 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5092 as Perl does (e.g. [-az] or [az-]).
5093
5094 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5095 all possible positions.
5096
5097 5. Split up into different modules to avoid including unneeded functions in a
5098 compiled binary. However, compile and exec are still in one module. The "study"
5099 function is split off.
5100
5101 6. The character tables are now in a separate module whose source is generated
5102 by an auxiliary program - but can then be edited by hand if required. There are
5103 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5104 toupper() in the code.
5105
5106 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5107 make them global. Abolish the function for setting them, as the caller can now
5108 set them directly.
5109
5110
5111 Version 0.92 11-Sep-97
5112 ----------------------
5113
5114 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5115 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5116
5117 2. Caseless matching was not working in character classes if the characters in
5118 the pattern were in upper case.
5119
5120 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5121
5122 4. Make PCRE_ANCHORED public and accept as a compile option.
5123
5124 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5125 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5126 pass them.
5127
5128 6. Give an error if bad option bits passed at compile or run time.
5129
5130 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5131 pcretest to cause it to pass that flag.
5132
5133 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5134 options, and the first character, if set.
5135
5136 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5137
5138
5139 Version 0.91 10-Sep-97
5140 ----------------------
5141
5142 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5143 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5144
5145 2. PCRE was looping on encountering an indefinitely repeated back reference to
5146 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5147 Perl does - treats the match as successful.
5148
5149 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5