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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1338 - (show annotations)
Fri Jun 14 07:54:36 2013 UTC (6 years, 4 months ago) by zherczeg
File size: 247572 byte(s)
Error occurred while calculating annotation data.
Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT executable memory. Patch inspired by Carsten Klein.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.34 xx-xxxx-201x
5 --------------------------
6
7 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8 executable memory. Patch inspired by Carsten Klein.
9
10
11 Version 8.33 28-May-2013
12 --------------------------
13
14 1. Added 'U' to some constants that are compared to unsigned integers, to
15 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
16 variables that are added to signed variables, to ensure the result is
17 signed and can be negated.
18
19 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
20 CMake config files.
21
22 3. Revise the creation of config.h.generic so that all boolean macros are
23 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
24 overriding via -D on the command line possible.
25
26 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
27 to unsigned int is reported to make a quite noticeable speed difference in
28 a specific Windows environment. Testing on Linux did also appear to show
29 some benefit (and it is clearly not harmful). Also fixed the definition of
30 Xop which should be unsigned.
31
32 5. Related to (4), changing the definition of the intermediate variable cc
33 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
34 improvements.
35
36 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
37 unnecessary spaces.
38
39 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
40 and later.
41
42 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
43
44 9. Optimizing fast_forward_start_bits in JIT.
45
46 10. Adding support for callouts in JIT, and fixing some issues revealed
47 during this work. Namely:
48
49 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
50
51 (b) Minimum length was not checked before the matching is started.
52
53 11. The value of capture_last that is passed to callouts was incorrect in some
54 cases when there was a capture on one path that was subsequently abandoned
55 after a backtrack. Also, the capture_last value is now reset after a
56 recursion, since all captures are also reset in this case.
57
58 12. The interpreter no longer returns the "too many substrings" error in the
59 case when an overflowing capture is in a branch that is subsequently
60 abandoned after a backtrack.
61
62 13. In the pathological case when an offset vector of size 2 is used, pcretest
63 now prints out the matched string after a yield of 0 or 1.
64
65 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
66 Only supported by the JIT compiler at the moment.
67
68 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
69
70 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
71 offset of the starting point of the matching process, provided the offsets
72 vector is large enough.
73
74 17. The \A escape now records a lookbehind value of 1, though its execution
75 does not actually inspect the previous character. This is to ensure that,
76 in partial multi-segment matching, at least one character from the old
77 segment is retained when a new segment is processed. Otherwise, if there
78 are no lookbehinds in the pattern, \A might match incorrectly at the start
79 of a new segment.
80
81 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
82
83 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
84 gives some modest performance improvement in 8-bit mode.
85
86 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
87 be expressed in certain programming languages using Universal Character
88 Names.
89
90 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
91 which points out that "non characters" are not "characters that may not
92 appear in Unicode strings" but rather "characters that are reserved for
93 internal use and have only local meaning".
94
95 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
96 there was a conditional group that depended on an assertion, if the
97 assertion was false, the callout that immediately followed the alternation
98 in the condition was skipped when pcre_exec() was used for matching.
99
100 23. Allow an explicit callout to be inserted before an assertion that is the
101 condition for a conditional group, for compatibility with automatic
102 callouts, which always insert a callout at this point.
103
104 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
105 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
106
107 25. (*PRUNE) is now supported by the JIT compiler.
108
109 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
110
111 27. Fix the case where there are two or more SKIPs with arguments that may be
112 ignored.
113
114 28. (*SKIP) is now supported by the JIT compiler.
115
116 29. (*THEN) is now supported by the JIT compiler.
117
118 30. Update RunTest with additional test selector options.
119
120 31. The way PCRE handles backtracking verbs has been changed in two ways.
121
122 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
123 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
124 backtracking. In some cases this makes it more Perl-compatible, but Perl's
125 rather obscure rules do not always do the same thing.
126
127 (2) Previously, backtracking verbs were confined within assertions. This is
128 no longer the case for positive assertions, except for (*ACCEPT). Again,
129 this sometimes improves Perl compatibility, and sometimes does not.
130
131 32. A number of tests that were in test 2 because Perl did things differently
132 have been moved to test 1, because either Perl or PCRE has changed, and
133 these tests are now compatible.
134
135 32. Backtracking control verbs are now handled in the same way in JIT and
136 interpreter.
137
138 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
139 contained a forward subroutine reference caused a compile error.
140
141 34. Auto-detect and optimize limited repetitions in JIT.
142
143 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
144 blocking (*UTF) etc.
145
146 36. In the interpreter, maximizing pattern repetitions for characters and
147 character types now use tail recursion, which reduces stack usage.
148
149 37. The value of the max lookbehind was not correctly preserved if a compiled
150 and saved regex was reloaded on a host of different endianness.
151
152 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
153 of the compiled pattern block, expand the flags field from 16 to 32 bits
154 because it was almost full.
155
156 39. Try madvise first before posix_madvise.
157
158 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
159 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
160 original change locked out all matches of empty strings. This has been
161 changed so that one match of an empty string per line is recognized.
162 Subsequent searches on the same line (for colouring or for --only-matching,
163 for example) do not recognize empty strings.
164
165 41. Applied a user patch to fix a number of spelling mistakes in comments.
166
167 42. Data lines longer than 65536 caused pcretest to crash.
168
169 43. Clarified the data type for length and startoffset arguments for pcre_exec
170 and pcre_dfa_exec in the function-specific man pages, where they were
171 explicitly stated to be in bytes, never having been updated. I also added
172 some clarification to the pcreapi man page.
173
174 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
175 a segmentation fault.
176
177
178 Version 8.32 30-November-2012
179 -----------------------------
180
181 1. Improved JIT compiler optimizations for first character search and single
182 character iterators.
183
184 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
185 Patch by Daniel Richard G.
186
187 3. Single character iterator optimizations in the JIT compiler.
188
189 4. Improved JIT compiler optimizations for character ranges.
190
191 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
192 Reported by Giuseppe D'Angelo.
193
194 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
195 of a line, was being set incorrectly in cases where .* appeared inside
196 atomic brackets at the start of a pattern, or where there was a subsequent
197 *PRUNE or *SKIP.
198
199 7. Improved instruction cache flush for POWER/PowerPC.
200 Patch by Daniel Richard G.
201
202 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
203 grep:
204
205 (a) There is now no limit to the number of patterns to be matched.
206
207 (b) An error is given if a pattern is too long.
208
209 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
210 are now supported.
211
212 (d) --exclude-from and --include-from (multiple use) have been added.
213
214 (e) Exclusions and inclusions now apply to all files and directories, not
215 just to those obtained from scanning a directory recursively.
216
217 (f) Multiple uses of -f and --file-list are now supported.
218
219 (g) In a Windows environment, the default for -d has been changed from
220 "read" (the GNU grep default) to "skip", because otherwise the presence
221 of a directory in the file list provokes an error.
222
223 (h) The documentation has been revised and clarified in places.
224
225 9. Improve the matching speed of capturing brackets.
226
227 10. Changed the meaning of \X so that it now matches a Unicode extended
228 grapheme cluster.
229
230 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
231 out POSIX threads when JIT support is configured.
232
233 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
234
235 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
236 structure could go wrong in environments where size_t is not the same size
237 as int.
238
239 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
240 set.
241
242 15. The EBCDIC support had decayed; later updates to the code had included
243 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
244 general tidy up of EBCDIC-related issues, and the documentation was also
245 not quite right. There is now a test that can be run on ASCII systems to
246 check some of the EBCDIC-related things (but is it not a full test).
247
248 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
249 in a small tidy to the code.
250
251 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
252
253 18. If the --only-matching (-o) option in pcregrep is specified multiple
254 times, each one causes appropriate output. For example, -o1 -o2 outputs the
255 substrings matched by the 1st and 2nd capturing parentheses. A separating
256 string can be specified by --om-separator (default empty).
257
258 19. Improving the first n character searches.
259
260 20. Turn case lists for horizontal and vertical white space into macros so that
261 they are defined only once.
262
263 21. This set of changes together give more compatible Unicode case-folding
264 behaviour for characters that have more than one other case when UCP
265 support is available.
266
267 (a) The Unicode property table now has offsets into a new table of sets of
268 three or more characters that are case-equivalent. The MultiStage2.py
269 script that generates these tables (the pcre_ucd.c file) now scans
270 CaseFolding.txt instead of UnicodeData.txt for character case
271 information.
272
273 (b) The code for adding characters or ranges of characters to a character
274 class has been abstracted into a generalized function that also handles
275 case-independence. In UTF-mode with UCP support, this uses the new data
276 to handle characters with more than one other case.
277
278 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
279 whose other case is greater than 256 are now correctly matched
280 caselessly. Previously, the high codepoint matched the low one, but not
281 vice versa.
282
283 (d) The processing of \h, \H, \v, and \ in character classes now makes use
284 of the new class addition function, using character lists defined as
285 macros alongside the case definitions of 20 above.
286
287 (e) Caseless back references now work with characters that have more than
288 one other case.
289
290 (f) General caseless matching of characters with more than one other case
291 is supported.
292
293 22. Unicode character properties were updated from Unicode 6.2.0
294
295 23. Improved CMake support under Windows. Patch by Daniel Richard G.
296
297 24. Add support for 32-bit character strings, and UTF-32
298
299 25. Major JIT compiler update (code refactoring and bugfixing).
300 Experimental Sparc 32 support is added.
301
302 26. Applied a modified version of Daniel Richard G's patch to create
303 pcre.h.generic and config.h.generic by "make" instead of in the
304 PrepareRelease script.
305
306 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
307 pcre_compile.c when checking for a zero character.
308
309 28. Introducing a native interface for JIT. Through this interface, the compiled
310 machine code can be directly executed. The purpose of this interface is to
311 provide fast pattern matching, so several sanity checks are not performed.
312 However, feature tests are still performed. The new interface provides
313 1.4x speedup compared to the old one.
314
315 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
316 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
317 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
318
319 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
320 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
321 the "old" RFC 2279). Instead, it ended up passing a negative length to
322 pcre_exec().
323
324 31. Add support for GCC's visibility feature to hide internal functions.
325
326 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
327 "unknown -C option" after outputting 0 or 1.
328
329 33. There is now support for generating a code coverage report for the test
330 suite in environments where gcc is the compiler and lcov is installed. This
331 is mainly for the benefit of the developers.
332
333 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
334 unaddressable using valgrind annotations, allowing valgrind to detect
335 invalid memory accesses. This is mainly for the benefit of the developers.
336
337 25. (*UTF) can now be used to start a pattern in any of the three libraries.
338
339 26. Give configure error if --enable-cpp but no C++ compiler found.
340
341
342 Version 8.31 06-July-2012
343 -------------------------
344
345 1. Fixing a wrong JIT test case and some compiler warnings.
346
347 2. Removed a bashism from the RunTest script.
348
349 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
350 to unsigned type, result still unsigned" that was given by an MS compiler
351 on encountering the code "-sizeof(xxx)".
352
353 4. Partial matching support is added to the JIT compiler.
354
355 5. Fixed several bugs concerned with partial matching of items that consist
356 of more than one character:
357
358 (a) /^(..)\1/ did not partially match "aba" because checking references was
359 done on an "all or nothing" basis. This also applied to repeated
360 references.
361
362 (b) \R did not give a hard partial match if \r was found at the end of the
363 subject.
364
365 (c) \X did not give a hard partial match after matching one or more
366 characters at the end of the subject.
367
368 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
369 a partial match for the string "\r".
370
371 (e) When newline was set to CRLF, the metacharacter "." did not recognize
372 a partial match for a CR character at the end of the subject string.
373
374 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
375 running pcretest, the text "(JIT)" added to the output whenever JIT is
376 actually used to run the match.
377
378 7. Individual JIT compile options can be set in pcretest by following -s+[+]
379 or /S+[+] with a digit between 1 and 7.
380
381 8. OP_NOT now supports any UTF character not just single-byte ones.
382
383 9. (*MARK) control verb is now supported by the JIT compiler.
384
385 10. The command "./RunTest list" lists the available tests without actually
386 running any of them. (Because I keep forgetting what they all are.)
387
388 11. Add PCRE_INFO_MAXLOOKBEHIND.
389
390 12. Applied a (slightly modified) user-supplied patch that improves performance
391 when the heap is used for recursion (compiled with --disable-stack-for-
392 recursion). Instead of malloc and free for each heap frame each time a
393 logical recursion happens, frames are retained on a chain and re-used where
394 possible. This sometimes gives as much as 30% improvement.
395
396 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
397 call.
398
399 14. As documented, (*COMMIT) is now confined to within a positive assertion.
400
401 15. It is now possible to link pcretest with libedit as an alternative to
402 libreadline.
403
404 16. (*COMMIT) control verb is now supported by the JIT compiler.
405
406 17. The Unicode data tables have been updated to Unicode 6.1.0.
407
408 18. Added --file-list option to pcregrep.
409
410 19. Added binary file support to pcregrep, including the -a, --binary-files,
411 -I, and --text options.
412
413 20. The madvise function is renamed for posix_madvise for QNX compatibility
414 reasons. Fixed by Giuseppe D'Angelo.
415
416 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
417
418 22. Changed the option for creating symbolic links for 16-bit man pages from
419 -s to -sf so that re-installing does not cause issues.
420
421 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
422
423 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
424 matches in certain environments (the workspace was not being correctly
425 retained). Also added to pcre_dfa_exec() a simple plausibility check on
426 some of the workspace data at the beginning of a restart.
427
428 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
429 was not doing so when it should - probably a typo introduced by SVN 528
430 (change 8.10/14).
431
432 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
433 \w+ when the character tables indicated that \x{c4} was a word character.
434 There were several related cases, all because the tests for doing a table
435 lookup were testing for characters less than 127 instead of 255.
436
437 27. If a pattern contains capturing parentheses that are not used in a match,
438 their slots in the ovector are set to -1. For those that are higher than
439 any matched groups, this happens at the end of processing. In the case when
440 there were back references that the ovector was too small to contain
441 (causing temporary malloc'd memory to be used during matching), and the
442 highest capturing number was not used, memory off the end of the ovector
443 was incorrectly being set to -1. (It was using the size of the temporary
444 memory instead of the true size.)
445
446 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
447 ovector size, it uses memory at the end of the block that it has got.
448
449 29. Check for an overlong MARK name and give an error at compile time. The
450 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
451
452 30. JIT compiler update.
453
454 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
455 Rill for the patch.
456
457 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
458
459 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
460
461 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
462 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
463 was enabled.
464
465 35. Improve JIT code generation for greedy plus quantifier.
466
467 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
468 1 to "aa" instead of to an empty string. The bug affected repeated groups
469 that could potentially match an empty string.
470
471 37. Optimizing single character iterators in JIT.
472
473 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
474 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
475 codepoints that are too big for the mode are faulted, and in a UTF mode,
476 disallowed codepoints are also faulted.
477
478 39. If PCRE was compiled with UTF support, in three places in the DFA
479 matcher there was code that should only have been obeyed in UTF mode, but
480 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
481 processing when bytes with values greater than 127 were present. In 16-bit
482 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
483 both cases the values are those that cannot be the first data item in a UTF
484 character. The three items that might have provoked this were recursions,
485 possessively repeated groups, and atomic groups.
486
487 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
488 and pcregrep, because some OS require shared objects to be explicitly
489 passed to ld, causing the link step to fail if they are not.
490
491 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
492 mode, patterns that started with \h* or \R* might be incorrectly matched.
493
494
495 Version 8.30 04-February-2012
496 -----------------------------
497
498 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
499 name is defined in ctype.h.
500
501 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
502 only in quite long subpatterns.
503
504 3. Removed the function pcre_info(), which has been obsolete and deprecated
505 since it was replaced by pcre_fullinfo() in February 2000.
506
507 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
508 match a (*MARK), and the match failed at the start of the subject, a
509 reference to memory before the start of the subject could occur. This bug
510 was introduced by fix 17 of release 8.21.
511
512 5. A reference to an unset group with zero minimum repetition was giving
513 totally wrong answers (in non-JavaScript-compatibility mode). For example,
514 /(another)?(\1?)test/ matched against "hello world test". This bug was
515 introduced in release 8.13.
516
517 6. Add support for 16-bit character strings (a large amount of work involving
518 many changes and refactorings).
519
520 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
521 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
522 from a file.
523
524 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
525 rounding is not applied in this particular case).
526
527 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
528 if they appear, or are escaped, in patterns.
529
530 10. Get rid of a number of -Wunused-but-set-variable warnings.
531
532 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
533 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
534 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
535 also returns the mark "x". This bug applied to capturing parentheses,
536 non-capturing parentheses, and atomic parentheses. It also applied to some
537 assertions.
538
539 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
540 information out of configure.ac instead of relying on pcre.h.generic, which
541 is not stored in the repository.
542
543 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
544 -lreadline.
545
546 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
547
548 15. Applied Graycode's patch to put the top-level frame on the stack rather
549 than the heap when not using the stack for recursion. This gives a
550 performance improvement in many cases when recursion is not deep.
551
552 16. Experimental code added to "pcretest -C" to output the stack frame size.
553
554
555 Version 8.21 12-Dec-2011
556 ------------------------
557
558 1. Updating the JIT compiler.
559
560 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
561 are added as well.
562
563 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
564 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
565 calling _pcre_jit_exec. Some extra comments are added.
566
567 4. (*MARK) settings inside atomic groups that do not contain any capturing
568 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
569 was introduced by change 18 for 8.20.
570
571 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
572 ECMA-262 standard.
573
574 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
575 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
576 This bug was probably introduced by change 9 of 8.13.
577
578 7. While fixing 6 above, I noticed that a number of other items were being
579 incorrectly rejected as "not fixed length". This arose partly because newer
580 opcodes had not been added to the fixed-length checking code. I have (a)
581 corrected the bug and added tests for these items, and (b) arranged for an
582 error to occur if an unknown opcode is encountered while checking for fixed
583 length instead of just assuming "not fixed length". The items that were
584 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
585 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
586 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
587
588 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
589 being incorrectly compiled and would have given unpredicatble results.
590
591 9. A possessively repeated subpattern with minimum repeat count greater than
592 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
593 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
594 the first (A) could occur when it should not.
595
596 10. Add a cast and remove a redundant test from the code.
597
598 11. JIT should use pcre_malloc/pcre_free for allocation.
599
600 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
601 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
602 is anything other than /usr, -L is still shown).
603
604 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
605
606 14. Perl does not support \N without a following name in a [] class; PCRE now
607 also gives an error.
608
609 15. If a forward reference was repeated with an upper limit of around 2000,
610 it caused the error "internal error: overran compiling workspace". The
611 maximum number of forward references (including repeats) was limited by the
612 internal workspace, and dependent on the LINK_SIZE. The code has been
613 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
614 the default depends on LINK_SIZE. There is a new upper limit (for safety)
615 of around 200,000 forward references. While doing this, I also speeded up
616 the filling in of repeated forward references.
617
618 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
619 incorrectly expecting the subject to contain another "a" after the start.
620
621 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
622 in the match, the SKIP should be ignored. This was not happening; instead
623 the SKIP was being treated as NOMATCH. For patterns such as
624 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
625 tested.
626
627 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
628 now much more compatible with Perl, in particular in cases where the result
629 is a non-match for a non-anchored pattern. For example, if
630 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
631 "m", where previously it did not return a name. A side effect of this
632 change is that for partial matches, the last encountered mark name is
633 returned, as for non matches. A number of tests that were previously not
634 Perl-compatible have been moved into the Perl-compatible test files. The
635 refactoring has had the pleasing side effect of removing one argument from
636 the match() function, thus reducing its stack requirements.
637
638 19. If the /S+ option was used in pcretest to study a pattern using JIT,
639 subsequent uses of /S (without +) incorrectly behaved like /S+.
640
641 21. Retrieve executable code size support for the JIT compiler and fixing
642 some warnings.
643
644 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
645 not work when the shorter character appeared right at the end of the
646 subject string.
647
648 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
649 systems.
650
651 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
652 output it when the /M option is used in pcretest.
653
654 25. The CheckMan script was not being included in the distribution. Also, added
655 an explicit "perl" to run Perl scripts from the PrepareRelease script
656 because this is reportedly needed in Windows.
657
658 26. If study data was being save in a file and studying had not found a set of
659 "starts with" bytes for the pattern, the data written to the file (though
660 never used) was taken from uninitialized memory and so caused valgrind to
661 complain.
662
663 27. Updated RunTest.bat as provided by Sheri Pierce.
664
665 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
666
667 29. Computation of memory usage for the table of capturing group names was
668 giving an unnecessarily large value.
669
670
671 Version 8.20 21-Oct-2011
672 ------------------------
673
674 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
675 a POSIX class. After further experiments with Perl, which convinced me that
676 Perl has bugs and confusions, a closing square bracket is no longer allowed
677 in a POSIX name. This bug also affected patterns with classes that started
678 with full stops.
679
680 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
681 captured substring, but while checking the failing first alternative,
682 substring 1 is temporarily captured. If the output vector supplied to
683 pcre_exec() was not big enough for this capture, the yield of the function
684 was still zero ("insufficient space for captured substrings"). This cannot
685 be totally fixed without adding another stack variable, which seems a lot
686 of expense for a edge case. However, I have improved the situation in cases
687 such as /(a)(b)x|abc/ matched against "abc", where the return code
688 indicates that fewer than the maximum number of slots in the ovector have
689 been set.
690
691 3. Related to (2) above: when there are more back references in a pattern than
692 slots in the output vector, pcre_exec() uses temporary memory during
693 matching, and copies in the captures as far as possible afterwards. It was
694 using the entire output vector, but this conflicts with the specification
695 that only 2/3 is used for passing back captured substrings. Now it uses
696 only the first 2/3, for compatibility. This is, of course, another edge
697 case.
698
699 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
700 main code base, and can be used by building with --enable-jit. When this is
701 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
702 runtime --no-jit option is given.
703
704 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
705 ovector, the return from the function was zero, implying that there were
706 other matches that did not fit. The correct "exactly full" value is now
707 returned.
708
709 6. If a subpattern that was called recursively or as a subroutine contained
710 (*PRUNE) or any other control that caused it to give a non-standard return,
711 invalid errors such as "Error -26 (nested recursion at the same subject
712 position)" or even infinite loops could occur.
713
714 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
715 computing the minimum length on reaching *ACCEPT, and so ended up with the
716 wrong value of 1 rather than 0. Further investigation indicates that
717 computing a minimum subject length in the presence of *ACCEPT is difficult
718 (think back references, subroutine calls), and so I have changed the code
719 so that no minimum is registered for a pattern that contains *ACCEPT.
720
721 8. If (*THEN) was present in the first (true) branch of a conditional group,
722 it was not handled as intended. [But see 16 below.]
723
724 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
725 Sheri Pierce.
726
727 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
728 the first byte in a match must be "a".
729
730 11. Change 17 for 8.13 increased the recursion depth for patterns like
731 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
732 pattern contains any instances of (*THEN). If it does not, the old
733 optimizations are restored. It would be nice to do this on a per-group
734 basis, but at the moment that is not feasible.
735
736 12. In some environments, the output of pcretest -C is CRLF terminated. This
737 broke RunTest's code that checks for the link size. A single white space
738 character after the value is now allowed for.
739
740 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
741 For "fr", it uses the Windows-specific input and output files.
742
743 14. If (*THEN) appeared in a group that was called recursively or as a
744 subroutine, it did not work as intended. [But see next item.]
745
746 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
747 pattern fragments (but not containing any | characters). If A and B are
748 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
749 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
750 D. In other words, Perl considers parentheses that do not contain any |
751 characters to be part of a surrounding alternative, whereas PCRE was
752 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
753 differently. PCRE now behaves in the same way as Perl, except in the case
754 of subroutine/recursion calls such as (?1) which have in any case always
755 been different (but PCRE had them first :-).
756
757 16. Related to 15 above: Perl does not treat the | in a conditional group as
758 creating alternatives. Such a group is treated in the same way as an
759 ordinary group without any | characters when processing (*THEN). PCRE has
760 been changed to match Perl's behaviour.
761
762 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
763 RunGrepTest script failed.
764
765 18. Change 22 for version 13 caused atomic groups to use more stack. This is
766 inevitable for groups that contain captures, but it can lead to a lot of
767 stack use in large patterns. The old behaviour has been restored for atomic
768 groups that do not contain any capturing parentheses.
769
770 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
771 suppress the check for a minimum subject length at run time. (If it was
772 given to pcre_exec() or pcre_dfa_exec() it did work.)
773
774 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
775 fail to work when decoding hex characters in data strings in EBCDIC
776 environments.
777
778 21. It appears that in at least one Mac OS environment, the isxdigit() function
779 is implemented as a macro that evaluates to its argument more than once,
780 contravening the C 90 Standard (I haven't checked a later standard). There
781 was an instance in pcretest which caused it to go wrong when processing
782 \x{...} escapes in subject strings. The has been rewritten to avoid using
783 things like p++ in the argument of isxdigit().
784
785
786 Version 8.13 16-Aug-2011
787 ------------------------
788
789 1. The Unicode data tables have been updated to Unicode 6.0.0.
790
791 2. Two minor typos in pcre_internal.h have been fixed.
792
793 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
794 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
795 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
796
797 4. There were a number of related bugs in the code for matching backrefences
798 caselessly in UTF-8 mode when codes for the characters concerned were
799 different numbers of bytes. For example, U+023A and U+2C65 are an upper
800 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
801 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
802 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
803 2-byte code at the end of the subject (it thought there wasn't enough data
804 left).
805
806 5. Comprehensive information about what went wrong is now returned by
807 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
808 as the output vector has at least 2 elements. The offset of the start of
809 the failing character and a reason code are placed in the vector.
810
811 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
812 now returned is for the first byte of the failing character, instead of the
813 last byte inspected. This is an incompatible change, but I hope it is small
814 enough not to be a problem. It makes the returned offset consistent with
815 pcre_exec() and pcre_dfa_exec().
816
817 7. pcretest now gives a text phrase as well as the error number when
818 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
819 failure, the offset and reason code are output.
820
821 8. When \R was used with a maximizing quantifier it failed to skip backwards
822 over a \r\n pair if the subsequent match failed. Instead, it just skipped
823 back over a single character (\n). This seems wrong (because it treated the
824 two characters as a single entity when going forwards), conflicts with the
825 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
826 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
827 has been changed.
828
829 9. Some internal refactoring has changed the processing so that the handling
830 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
831 time (the PCRE_DOTALL option was changed this way some time ago: version
832 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
833 which was always a bit of a fudge. It also means that there is one less
834 argument for the match() function, which reduces its stack requirements
835 slightly. This change also fixes an incompatibility with Perl: the pattern
836 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
837
838 10. More internal refactoring has drastically reduced the number of recursive
839 calls to match() for possessively repeated groups such as (abc)++ when
840 using pcre_exec().
841
842 11. While implementing 10, a number of bugs in the handling of groups were
843 discovered and fixed:
844
845 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
846 (a|)*(?1) gave a compile-time internal error.
847 ((a|)+)+ did not notice that the outer group could match an empty string.
848 (^a|^)+ was not marked as anchored.
849 (.*a|.*)+ was not marked as matching at start or after a newline.
850
851 12. Yet more internal refactoring has removed another argument from the match()
852 function. Special calls to this function are now indicated by setting a
853 value in a variable in the "match data" data block.
854
855 13. Be more explicit in pcre_study() instead of relying on "default" for
856 opcodes that mean there is no starting character; this means that when new
857 ones are added and accidentally left out of pcre_study(), testing should
858 pick them up.
859
860 14. The -s option of pcretest has been documented for ages as being an old
861 synonym of -m (show memory usage). I have changed it to mean "force study
862 for every regex", that is, assume /S for every regex. This is similar to -i
863 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
864 using it. It makes it easier to run collections of tests with and without
865 study enabled, and thereby test pcre_study() more easily. All the standard
866 tests are now run with and without -s (but some patterns can be marked as
867 "never study" - see 20 below).
868
869 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
870 restoration of the capturing data to the outer values was not happening
871 correctly.
872
873 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
874 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
875 pattern had matched an empty string, and so incorrectly returned a no
876 match.
877
878 17. There was optimizing code for the last branch of non-capturing parentheses,
879 and also for the obeyed branch of a conditional subexpression, which used
880 tail recursion to cut down on stack usage. Unfortunately, now that there is
881 the possibility of (*THEN) occurring in these branches, tail recursion is
882 no longer possible because the return has to be checked for (*THEN). These
883 two optimizations have therefore been removed. [But see 8.20/11 above.]
884
885 18. If a pattern containing \R was studied, it was assumed that \R always
886 matched two bytes, thus causing the minimum subject length to be
887 incorrectly computed because \R can also match just one byte.
888
889 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
890 was incorrectly computed.
891
892 20. If /S is present twice on a test pattern in pcretest input, it now
893 *disables* studying, thereby overriding the use of -s on the command line
894 (see 14 above). This is necessary for one or two tests to keep the output
895 identical in both cases.
896
897 21. When (*ACCEPT) was used in an assertion that matched an empty string and
898 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
899
900 22. When an atomic group that contained a capturing parenthesis was
901 successfully matched, but the branch in which it appeared failed, the
902 capturing was not being forgotten if a higher numbered group was later
903 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
904 group 1 to "a", when in fact it should be unset. This applied to multi-
905 branched capturing and non-capturing groups, repeated or not, and also to
906 positive assertions (capturing in negative assertions does not happen
907 in PCRE) and also to nested atomic groups.
908
909 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
910 subject after a captured substring, to make it easier to tell which of a
911 number of identical substrings has been captured.
912
913 24. The way atomic groups are processed by pcre_exec() has been changed so that
914 if they are repeated, backtracking one repetition now resets captured
915 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
916 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
917 "aaa". Previously, it would have been "a". As part of this code
918 refactoring, the way recursive calls are handled has also been changed.
919
920 25. If an assertion condition captured any substrings, they were not passed
921 back unless some other capturing happened later. For example, if
922 (?(?=(a))a) was matched against "a", no capturing was returned.
923
924 26. When studying a pattern that contained subroutine calls or assertions,
925 the code for finding the minimum length of a possible match was handling
926 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
927 group 1 called group 2 while simultaneously a separate group 2 called group
928 1). A stack overflow occurred in this case. I have fixed this by limiting
929 the recursion depth to 10.
930
931 27. Updated RunTest.bat in the distribution to the version supplied by Tom
932 Fortmann. This supports explicit test numbers on the command line, and has
933 argument validation and error reporting.
934
935 28. An instance of \X with an unlimited repeat could fail if at any point the
936 first character it looked at was a mark character.
937
938 29. Some minor code refactoring concerning Unicode properties and scripts
939 should reduce the stack requirement of match() slightly.
940
941 30. Added the '=' option to pcretest to check the setting of unused capturing
942 slots at the end of the pattern, which are documented as being -1, but are
943 not included in the return count.
944
945 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
946 compiled something random. Now it gives a compile-time error (as does
947 Perl).
948
949 32. A *MARK encountered during the processing of a positive assertion is now
950 recorded and passed back (compatible with Perl).
951
952 33. If --only-matching or --colour was set on a pcregrep call whose pattern
953 had alternative anchored branches, the search for a second match in a line
954 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
955 matched the line "0102" twice. The same bug affected patterns that started
956 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
957 twice.
958
959 34. Previously, PCRE did not allow quantification of assertions. However, Perl
960 does, and because of capturing effects, quantifying parenthesized
961 assertions may at times be useful. Quantifiers are now allowed for
962 parenthesized assertions.
963
964 35. A minor code tidy in pcre_compile() when checking options for \R usage.
965
966 36. \g was being checked for fancy things in a character class, when it should
967 just be a literal "g".
968
969 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
970 appearance of a nested POSIX class supersedes an apparent external class.
971 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
972 unescaped square brackets may also appear as part of class names. For
973 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
974 more like Perl. (But see 8.20/1 above.)
975
976 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
977 was because it thought it was \N{name}, which is not supported).
978
979 39. Add minix to OS list not supporting the -S option in pcretest.
980
981 40. PCRE tries to detect cases of infinite recursion at compile time, but it
982 cannot analyze patterns in sufficient detail to catch mutual recursions
983 such as ((?1))((?2)). There is now a runtime test that gives an error if a
984 subgroup is called recursively as a subpattern for a second time at the
985 same position in the subject string. In previous releases this might have
986 been caught by the recursion limit, or it might have run out of stack.
987
988 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
989 happen only once. PCRE was, however incorrectly giving a compile time error
990 "recursive call could loop indefinitely" because it cannot analyze the
991 pattern in sufficient detail. The compile time test no longer happens when
992 PCRE is compiling a conditional subpattern, but actual runaway loops are
993 now caught at runtime (see 40 above).
994
995 42. It seems that Perl allows any characters other than a closing parenthesis
996 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
997 has been changed to be the same.
998
999 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1000 as not to get warnings when autogen.sh is called. Also changed
1001 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1002
1003 44. To help people who use pcregrep to scan files containing exceedingly long
1004 lines, the following changes have been made:
1005
1006 (a) The default value of the buffer size parameter has been increased from
1007 8K to 20K. (The actual buffer used is three times this size.)
1008
1009 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1010 PCRE is built.
1011
1012 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1013 to be set at run time.
1014
1015 (d) Numerical values in pcregrep options can be followed by K or M, for
1016 example --buffer-size=50K.
1017
1018 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1019 given and the return code is set to 2.
1020
1021 45. Add a pointer to the latest mark to the callout data block.
1022
1023 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1024 partial match of an empty string instead of no match. This was specific to
1025 the use of ".".
1026
1027 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1028 complete match instead of a partial match. This bug was dependent on both
1029 the PCRE_UTF8 and PCRE_DOTALL options being set.
1030
1031 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1032 starting byte set, because \b was not being ignored.
1033
1034
1035 Version 8.12 15-Jan-2011
1036 ------------------------
1037
1038 1. Fixed some typos in the markup of the man pages, and wrote a script that
1039 checks for such things as part of the documentation building process.
1040
1041 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1042 --match-limit and --recursion-limit options (added for 8.11). In
1043 particular, this made one of the standard tests fail. (The integer value
1044 went into the wrong half of a long int.)
1045
1046 3. If the --colour option was given to pcregrep with -v (invert match), it
1047 did strange things, either producing crazy output, or crashing. It should,
1048 of course, ignore a request for colour when reporting lines that do not
1049 match.
1050
1051 4. Another pcregrep bug caused similar problems if --colour was specified with
1052 -M (multiline) and the pattern match finished with a line ending.
1053
1054 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1055 matched in multiline mode, the following line was shown as part of the
1056 match. This seems wrong, so I have changed it.
1057
1058 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1059 the check for further matches in the same line (so they could be coloured)
1060 to overrun the end of the current line. If another match was found, it was
1061 incorrectly shown (and then shown again when found in the next line).
1062
1063 7. If pcregrep was compiled under Windows, there was a reference to the
1064 function pcregrep_exit() before it was defined. I am assuming this was
1065 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1066 reported by a user. I've moved the definition above the reference.
1067
1068
1069 Version 8.11 10-Dec-2010
1070 ------------------------
1071
1072 1. (*THEN) was not working properly if there were untried alternatives prior
1073 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1074 backtracked to try for "b" instead of moving to the next alternative branch
1075 at the same level (in this case, to look for "c"). The Perl documentation
1076 is clear that when (*THEN) is backtracked onto, it goes to the "next
1077 alternative in the innermost enclosing group".
1078
1079 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1080 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1081 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1082 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1083 (*THEN).
1084
1085 3. If \s appeared in a character class, it removed the VT character from
1086 the class, even if it had been included by some previous item, for example
1087 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1088 of \s, but is part of the POSIX "space" class.)
1089
1090 4. A partial match never returns an empty string (because you can always
1091 match an empty string at the end of the subject); however the checking for
1092 an empty string was starting at the "start of match" point. This has been
1093 changed to the "earliest inspected character" point, because the returned
1094 data for a partial match starts at this character. This means that, for
1095 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1096 (previously it gave "no match").
1097
1098 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1099 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1100 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1101 has an implication that the given string is incomplete (because a partial
1102 match is preferred over a full match). For this reason, these items now
1103 give a partial match in this situation. [Aside: previously, the one case
1104 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1105 match rather than a full match, which was wrong by the old rules, but is
1106 now correct.]
1107
1108 6. There was a bug in the handling of #-introduced comments, recognized when
1109 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1110 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1111 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1112 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1113 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1114 places in pcre_compile().
1115
1116 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1117 comments when looking ahead for named forward references to subpatterns,
1118 the only newline sequence it recognized was NL. It now handles newlines
1119 according to the set newline convention.
1120
1121 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1122 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1123 cater for a lack of strerror(). These oversights have been fixed.
1124
1125 9. Added --match-limit and --recursion-limit to pcregrep.
1126
1127 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1128
1129 11. When the -o option was used, pcregrep was setting a return code of 1, even
1130 when matches were found, and --line-buffered was not being honoured.
1131
1132 12. Added an optional parentheses number to the -o and --only-matching options
1133 of pcregrep.
1134
1135 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1136 can match an empty string. The code to do it in pcretest and pcredemo
1137 needed fixing:
1138
1139 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1140 only one byte after an empty string match just before CRLF (this case
1141 just got forgotten; "any" and "anycrlf" were OK).
1142
1143 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1144 mode when an empty string match preceded an ASCII character followed by
1145 a non-ASCII character. (The code for advancing by one character rather
1146 than one byte was nonsense.)
1147
1148 (c) The pcredemo.c sample program did not have any code at all to handle
1149 the cases when CRLF is a valid newline sequence.
1150
1151 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1152 as a starting offset was within the subject string. There is now a new
1153 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1154 negative or greater than the length of the string. In order to test this,
1155 pcretest is extended to allow the setting of negative starting offsets.
1156
1157 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1158 starting offset points to the beginning of a UTF-8 character was
1159 unnecessarily clumsy. I tidied it up.
1160
1161 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1162 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1163
1164 17. Nobody had reported that the --include_dir option, which was added in
1165 release 7.7 should have been called --include-dir (hyphen, not underscore)
1166 for compatibility with GNU grep. I have changed it to --include-dir, but
1167 left --include_dir as an undocumented synonym, and the same for
1168 --exclude-dir, though that is not available in GNU grep, at least as of
1169 release 2.5.4.
1170
1171 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1172 characters from a string of bytes) have been redefined so as not to use
1173 loops, in order to improve performance in some environments. At the same
1174 time, I abstracted some of the common code into auxiliary macros to save
1175 repetition (this should not affect the compiled code).
1176
1177 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1178 compile-time error is now given if \c is not followed by an ASCII
1179 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1180 different, and any byte value is allowed.)
1181
1182 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1183 START_OPTIMIZE option, which is now allowed at compile time - but just
1184 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1185 to pcregrep and other applications that have no direct access to PCRE
1186 options. The new /Y option in pcretest sets this option when calling
1187 pcre_compile().
1188
1189 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1190 back references. Groups containing recursive back references were forced to
1191 be atomic by that change, but in the case of named groups, the amount of
1192 memory required was incorrectly computed, leading to "Failed: internal
1193 error: code overflow". This has been fixed.
1194
1195 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1196 pcretest.c, to avoid build problems in some Borland environments.
1197
1198
1199 Version 8.10 25-Jun-2010
1200 ------------------------
1201
1202 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1203 THEN.
1204
1205 2. (*ACCEPT) was not working when inside an atomic group.
1206
1207 3. Inside a character class, \B is treated as a literal by default, but
1208 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1209 causes the error). The code is unchanged, but I tidied the documentation.
1210
1211 4. Inside a character class, PCRE always treated \R and \X as literals,
1212 whereas Perl faults them if its -w option is set. I have changed PCRE so
1213 that it faults them when PCRE_EXTRA is set.
1214
1215 5. Added support for \N, which always matches any character other than
1216 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1217
1218 6. When compiling pcregrep with newer versions of gcc which may have
1219 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1220 declared with attribute warn_unused_result" were given. Just casting the
1221 result to (void) does not stop the warnings; a more elaborate fudge is
1222 needed. I've used a macro to implement this.
1223
1224 7. Minor change to pcretest.c to avoid a compiler warning.
1225
1226 8. Added four artifical Unicode properties to help with an option to make
1227 \s etc use properties (see next item). The new properties are: Xan
1228 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1229
1230 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1231 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1232 this option. Modified pcretest to add /W to test this facility. Added
1233 REG_UCP to make it available via the POSIX interface.
1234
1235 10. Added --line-buffered to pcregrep.
1236
1237 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1238 studied, and the match started with a letter with a code point greater than
1239 127 whose first byte was different to the first byte of the other case of
1240 the letter, the other case of this starting letter was not recognized
1241 (#976).
1242
1243 12. If a pattern that was studied started with a repeated Unicode property
1244 test, for example, \p{Nd}+, there was the theoretical possibility of
1245 setting up an incorrect bitmap of starting bytes, but fortunately it could
1246 not have actually happened in practice until change 8 above was made (it
1247 added property types that matched character-matching opcodes).
1248
1249 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1250 possible starting bytes for non-anchored patterns.
1251
1252 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1253 \R, and also a number of cases that involve Unicode properties, both
1254 explicit and implicit when PCRE_UCP is set.
1255
1256 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1257 input, it could crash or give wrong results if characters with values
1258 greater than 0xc0 were present in the subject string. (Detail: it assumed
1259 UTF-8 input when processing these items.)
1260
1261 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1262 size_t is 64-bit (#991).
1263
1264 17. Added a check for running out of memory when PCRE is compiled with
1265 --disable-stack-for-recursion (#990).
1266
1267 18. If the last data line in a file for pcretest does not have a newline on
1268 the end, a newline was missing in the output.
1269
1270 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1271 less than 128) in its various bitmaps. However, there is a facility for
1272 generating tables according to the current locale when PCRE is compiled. It
1273 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1274 characters, are recognized by isspace() and therefore were getting set in
1275 these tables, and indeed these tables seem to approximate to ISO 8859. This
1276 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1277 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1278 which of course cannot start UTF-8 characters. I have changed the code so
1279 that only real ASCII characters (less than 128) and the correct starting
1280 bytes for UTF-8 encodings are set for characters greater than 127 when in
1281 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1282 altogether.)
1283
1284 20. Added the /T option to pcretest so as to be able to run tests with non-
1285 standard character tables, thus making it possible to include the tests
1286 used for 19 above in the standard set of tests.
1287
1288 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1289 reference to a subpattern the other side of a comment that contains an
1290 opening parenthesis caused either an internal compiling error, or a
1291 reference to the wrong subpattern.
1292
1293
1294 Version 8.02 19-Mar-2010
1295 ------------------------
1296
1297 1. The Unicode data tables have been updated to Unicode 5.2.0.
1298
1299 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1300 configured.
1301
1302 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1303 original author of that file, following a query about its status.
1304
1305 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1306 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1307
1308 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1309 quantifier applied to a forward-referencing subroutine call, could compile
1310 incorrect code or give the error "internal error: previously-checked
1311 referenced subpattern not found".
1312
1313 6. Both MS Visual Studio and Symbian OS have problems with initializing
1314 variables to point to external functions. For these systems, therefore,
1315 pcre_malloc etc. are now initialized to local functions that call the
1316 relevant global functions.
1317
1318 7. There were two entries missing in the vectors called coptable and poptable
1319 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1320 I've fixed the data, and added a kludgy way of testing at compile time that
1321 the lengths are correct (equal to the number of opcodes).
1322
1323 8. Following on from 7, I added a similar kludge to check the length of the
1324 eint vector in pcreposix.c.
1325
1326 9. Error texts for pcre_compile() are held as one long string to avoid too
1327 much relocation at load time. To find a text, the string is searched,
1328 counting zeros. There was no check for running off the end of the string,
1329 which could happen if a new error number was added without updating the
1330 string.
1331
1332 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1333
1334 11. \K was not working if it appeared in an atomic group or in a group that
1335 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1336 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1337 the assertion is positive, but not if it is negative.
1338
1339 12. Change 11 fortuitously reduced the size of the stack frame used in the
1340 "match()" function of pcre_exec.c by one pointer. Forthcoming
1341 implementation of support for (*MARK) will need an extra pointer on the
1342 stack; I have reserved it now, so that the stack frame size does not
1343 decrease.
1344
1345 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1346 item in branch that calls a recursion is a subroutine call - as in the
1347 second branch in the above example - was incorrectly given the compile-
1348 time error "recursive call could loop indefinitely" because pcre_compile()
1349 was not correctly checking the subroutine for matching a non-empty string.
1350
1351 14. The checks for overrunning compiling workspace could trigger after an
1352 overrun had occurred. This is a "should never occur" error, but it can be
1353 triggered by pathological patterns such as hundreds of nested parentheses.
1354 The checks now trigger 100 bytes before the end of the workspace.
1355
1356 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1357
1358
1359 Version 8.01 19-Jan-2010
1360 ------------------------
1361
1362 1. If a pattern contained a conditional subpattern with only one branch (in
1363 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1364 computed the wrong minimum data length (which is of course zero for such
1365 subpatterns). This could cause incorrect "no match" results.
1366
1367 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1368 the pattern is reset in the first branch, pcre_compile() failed with
1369 "internal error: code overflow at offset...". This happened only when
1370 the reset was to the original external option setting. (An optimization
1371 abstracts leading options settings into an external setting, which was the
1372 cause of this.)
1373
1374 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1375 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1376 assertion pattern did not match (meaning that the assertion was true), it
1377 was incorrectly treated as false if the SKIP had been reached during the
1378 matching. This also applied to assertions used as conditions.
1379
1380 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1381 assertion subpattern, including such a pattern used as a condition,
1382 unpredictable results occurred, instead of the error return
1383 PCRE_ERROR_DFA_UITEM.
1384
1385 5. The C++ GlobalReplace function was not working like Perl for the special
1386 situation when an empty string is matched. It now does the fancy magic
1387 stuff that is necessary.
1388
1389 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1390 removed. (These were left over from very, very early versions of PCRE.)
1391
1392 7. Some cosmetic changes to the code to make life easier when compiling it
1393 as part of something else:
1394
1395 (a) Change DEBUG to PCRE_DEBUG.
1396
1397 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1398 called "current" as "current_branch", to prevent a collision with the
1399 Linux macro when compiled as a kernel module.
1400
1401 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1402 prevent a collision with the Linux macro when compiled as a kernel
1403 module.
1404
1405 8. In pcre_compile() there are some checks for integer overflows that used to
1406 cast potentially large values to (double). This has been changed to that
1407 when building, a check for int64_t is made, and if it is found, it is used
1408 instead, thus avoiding the use of floating point arithmetic. (There is no
1409 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1410 double.
1411
1412 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1413 2005 (difference between two addresses compared to an unsigned value).
1414
1415 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1416 custom one, because of the following reported problem in Windows:
1417
1418 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1419 under Win32.
1420 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1421 therefore missing the function definition.
1422 - The compiler thus generates a "C" signature for the test function.
1423 - The linker fails to find the "C" function.
1424 - PCRE fails to configure if asked to do so against libbz2.
1425
1426 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1427 messages were output:
1428
1429 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1430 rerunning libtoolize, to keep the correct libtool macros in-tree.
1431 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1432
1433 I have done both of these things.
1434
1435 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1436 most of the time, it *can* run out if it is given a pattern that contains a
1437 runaway infinite recursion. I updated the discussion in the pcrestack man
1438 page.
1439
1440 13. Now that we have gone to the x.xx style of version numbers, the minor
1441 version may start with zero. Using 08 or 09 is a bad idea because users
1442 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1443 interpreted as invalid octal numbers. I've updated the previous comment in
1444 configure.ac, and also added a check that gives an error if 08 or 09 are
1445 used.
1446
1447 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1448 causing partial matching to fail when the end of the subject matched \W
1449 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1450
1451 15. There were some discrepancies between the declarations in pcre_internal.h
1452 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1453 their definitions. The declarations used "const uschar *" and the
1454 definitions used USPTR. Even though USPTR is normally defined as "const
1455 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1456 reported that: "This difference in casting confuses some C++ compilers, for
1457 example, SunCC recognizes above declarations as different functions and
1458 generates broken code for hbpcre." I have changed the declarations to use
1459 USPTR.
1460
1461 16. GNU libtool is named differently on some systems. The autogen.sh script now
1462 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1463 (FreeBSD).
1464
1465 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1466 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1467 comment: "Figure out how to create a longlong from a string: strtoll and
1468 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1469 instance, but it only takes 2 args instead of 3!"
1470
1471 18. A subtle bug concerned with back references has been fixed by a change of
1472 specification, with a corresponding code fix. A pattern such as
1473 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1474 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1475 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1476 same bug. Such groups have to be quantified to be useful, or contained
1477 inside another quantified group. (If there's no repetition, the reference
1478 can never match.) The problem arises because, having left the group and
1479 moved on to the rest of the pattern, a later failure that backtracks into
1480 the group uses the captured value from the final iteration of the group
1481 rather than the correct earlier one. I have fixed this in PCRE by forcing
1482 any group that contains a reference to itself to be an atomic group; that
1483 is, there cannot be any backtracking into it once it has completed. This is
1484 similar to recursive and subroutine calls.
1485
1486
1487 Version 8.00 19-Oct-09
1488 ----------------------
1489
1490 1. The table for translating pcre_compile() error codes into POSIX error codes
1491 was out-of-date, and there was no check on the pcre_compile() error code
1492 being within the table. This could lead to an OK return being given in
1493 error.
1494
1495 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1496 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1497 in a Windows environment.
1498
1499 3. The pcregrep --count option prints the count for each file even when it is
1500 zero, as does GNU grep. However, pcregrep was also printing all files when
1501 --files-with-matches was added. Now, when both options are given, it prints
1502 counts only for those files that have at least one match. (GNU grep just
1503 prints the file name in this circumstance, but including the count seems
1504 more useful - otherwise, why use --count?) Also ensured that the
1505 combination -clh just lists non-zero counts, with no names.
1506
1507 4. The long form of the pcregrep -F option was incorrectly implemented as
1508 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1509 but it seems right to fix it, and I didn't think it was worth preserving
1510 the old behaviour.
1511
1512 5. The command line items --regex=pattern and --regexp=pattern were not
1513 recognized by pcregrep, which required --regex pattern or --regexp pattern
1514 (with a space rather than an '='). The man page documented the '=' forms,
1515 which are compatible with GNU grep; these now work.
1516
1517 6. No libpcreposix.pc file was created for pkg-config; there was just
1518 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1519
1520 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1521 when UCP support is not needed, by modifying the Python script that
1522 generates it from Unicode data files. This should not matter if the module
1523 is correctly used as a library, but I received one complaint about 50K of
1524 unwanted data. My guess is that the person linked everything into his
1525 program rather than using a library. Anyway, it does no harm.
1526
1527 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1528 was a minimum greater than 1 for a wide character in a possessive
1529 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1530 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1531 character. Chaos in the form of incorrect output or a compiling loop could
1532 result.
1533
1534 9. The restrictions on what a pattern can contain when partial matching is
1535 requested for pcre_exec() have been removed. All patterns can now be
1536 partially matched by this function. In addition, if there are at least two
1537 slots in the offset vector, the offset of the earliest inspected character
1538 for the match and the offset of the end of the subject are set in them when
1539 PCRE_ERROR_PARTIAL is returned.
1540
1541 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1542 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1543 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1544 and may be more useful for multi-segment matching.
1545
1546 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1547 used to be given if ever the end of the subject was reached; now it is
1548 given only if matching could not proceed because another character was
1549 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1550 string "Z", which now yields "no match" instead of "partial match". In the
1551 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1552 final character ended with (*FAIL).
1553
1554 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1555 if the pattern had a "must contain" character that was already found in the
1556 earlier partial match, unless partial matching was again requested. For
1557 example, with the pattern /dog.(body)?/, the "must contain" character is
1558 "g". If the first part-match was for the string "dog", restarting with
1559 "sbody" failed. This bug has been fixed.
1560
1561 13. The string returned by pcre_dfa_exec() after a partial match has been
1562 changed so that it starts at the first inspected character rather than the
1563 first character of the match. This makes a difference only if the pattern
1564 starts with a lookbehind assertion or \b or \B (\K is not supported by
1565 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1566 matching functions compatible, and I think it's the right thing to do.
1567
1568 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1569 so that the demonstration program is easily available in environments where
1570 PCRE has not been installed from source.
1571
1572 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1573 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1574 library.
1575
1576 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1577 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1578 is not the first non-POSIX option to be added. Clearly some people find
1579 these options useful.
1580
1581 17. If a caller to the POSIX matching function regexec() passes a non-zero
1582 value for nmatch with a NULL value for pmatch, the value of
1583 nmatch is forced to zero.
1584
1585 18. RunGrepTest did not have a test for the availability of the -u option of
1586 the diff command, as RunTest does. It now checks in the same way as
1587 RunTest, and also checks for the -b option.
1588
1589 19. If an odd number of negated classes containing just a single character
1590 interposed, within parentheses, between a forward reference to a named
1591 subpattern and the definition of the subpattern, compilation crashed with
1592 an internal error, complaining that it could not find the referenced
1593 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1594 [The bug was that it was starting one character too far in when skipping
1595 over the character class, thus treating the ] as data rather than
1596 terminating the class. This meant it could skip too much.]
1597
1598 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1599 /g option in pcretest when the pattern contains \K, which makes it possible
1600 to have an empty string match not at the start, even when the pattern is
1601 anchored. Updated pcretest and pcredemo to use this option.
1602
1603 21. If the maximum number of capturing subpatterns in a recursion was greater
1604 than the maximum at the outer level, the higher number was returned, but
1605 with unset values at the outer level. The correct (outer level) value is
1606 now given.
1607
1608 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1609 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1610 make it do so. The string so far is captured, making this feature
1611 compatible with Perl.
1612
1613 23. The tests have been re-organized, adding tests 11 and 12, to make it
1614 possible to check the Perl 5.10 features against Perl 5.10.
1615
1616 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1617 pattern matches a fixed length string. PCRE did not allow this; now it
1618 does. Neither allows recursion.
1619
1620 25. I finally figured out how to implement a request to provide the minimum
1621 length of subject string that was needed in order to match a given pattern.
1622 (It was back references and recursion that I had previously got hung up
1623 on.) This code has now been added to pcre_study(); it finds a lower bound
1624 to the length of subject needed. It is not necessarily the greatest lower
1625 bound, but using it to avoid searching strings that are too short does give
1626 some useful speed-ups. The value is available to calling programs via
1627 pcre_fullinfo().
1628
1629 26. While implementing 25, I discovered to my embarrassment that pcretest had
1630 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1631 study optimizations had never been tested with that matching function.
1632 Oops. What is worse, even when it was passed study data, there was a bug in
1633 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1634 were also very few tests of studied patterns with pcre_dfa_exec().
1635
1636 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1637 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1638 on the other side of the coin, they are no longer allowed to have different
1639 names, because these cannot be distinguished in PCRE, and this has caused
1640 confusion. (This is a difference from Perl.)
1641
1642 28. When duplicate subpattern names are present (necessarily with different
1643 numbers, as required by 27 above), and a test is made by name in a
1644 conditional pattern, either for a subpattern having been matched, or for
1645 recursion in such a pattern, all the associated numbered subpatterns are
1646 tested, and the overall condition is true if the condition is true for any
1647 one of them. This is the way Perl works, and is also more like the way
1648 testing by number works.
1649
1650
1651 Version 7.9 11-Apr-09
1652 ---------------------
1653
1654 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1655 (pcretest), all targets were linked against these libraries. This included
1656 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1657 libraries. This caused unwanted dependencies to be created. This problem
1658 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1659 pcretest is linked with readline.
1660
1661 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1662 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1663 moved outside it again, because FALSE and TRUE are already defined in AIX,
1664 but BOOL is not.
1665
1666 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1667 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1668
1669 4. The pcregrep documentation said spaces were inserted as well as colons (or
1670 hyphens) following file names and line numbers when outputting matching
1671 lines. This is not true; no spaces are inserted. I have also clarified the
1672 wording for the --colour (or --color) option.
1673
1674 5. In pcregrep, when --colour was used with -o, the list of matching strings
1675 was not coloured; this is different to GNU grep, so I have changed it to be
1676 the same.
1677
1678 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1679 each matching line was coloured. Now it goes on to look for further matches
1680 of any of the test patterns, which is the same behaviour as GNU grep.
1681
1682 7. A pattern that could match an empty string could cause pcregrep to loop; it
1683 doesn't make sense to accept an empty string match in pcregrep, so I have
1684 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1685 seems to be how GNU grep behaves. [But see later change 40 for release
1686 8.33.]
1687
1688 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1689 start or after a newline", because the conditional assertion was not being
1690 correctly handled. The rule now is that both the assertion and what follows
1691 in the first alternative must satisfy the test.
1692
1693 9. If auto-callout was enabled in a pattern with a conditional group whose
1694 condition was an assertion, PCRE could crash during matching, both with
1695 pcre_exec() and pcre_dfa_exec().
1696
1697 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1698 used for matching.
1699
1700 11. Unicode property support in character classes was not working for
1701 characters (bytes) greater than 127 when not in UTF-8 mode.
1702
1703 12. Added the -M command line option to pcretest.
1704
1705 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1706
1707 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1708
1709 16. Added comments and documentation about mis-use of no_arg in the C++
1710 wrapper.
1711
1712 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1713 from Martin Jerabek that uses macro names for all relevant character and
1714 string constants.
1715
1716 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1717 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1718 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1719 these, but not everybody uses configure.
1720
1721 19. A conditional group that had only one branch was not being correctly
1722 recognized as an item that could match an empty string. This meant that an
1723 enclosing group might also not be so recognized, causing infinite looping
1724 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1725 with the subject "ab", where knowledge that the repeated group can match
1726 nothing is needed in order to break the loop.
1727
1728 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1729 exec(), but without supplying a callout function, matching went wrong.
1730
1731 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1732 leak if the size of the offset vector was greater than 30. When the vector
1733 is smaller, the saved offsets during recursion go onto a local stack
1734 vector, but for larger vectors malloc() is used. It was failing to free
1735 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1736 error, in fact).
1737
1738 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1739 heapframe that is used only when UTF-8 support is enabled. This caused no
1740 problem, but was untidy.
1741
1742 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1743 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1744 included within another project.
1745
1746 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1747 slightly modified by me:
1748
1749 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1750 not building pcregrep.
1751
1752 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1753 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1754
1755 25. Forward references, both numeric and by name, in patterns that made use of
1756 duplicate group numbers, could behave incorrectly or give incorrect errors,
1757 because when scanning forward to find the reference group, PCRE was not
1758 taking into account the duplicate group numbers. A pattern such as
1759 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1760
1761 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1762 the feature of a custom pointer more persuasive (as requested by a user).
1763
1764 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1765 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1766 pre-defined.
1767
1768 28. Added support for (*UTF8) at the start of a pattern.
1769
1770 29. Arrange for flags added by the "release type" setting in CMake to be shown
1771 in the configuration summary.
1772
1773
1774 Version 7.8 05-Sep-08
1775 ---------------------
1776
1777 1. Replaced UCP searching code with optimized version as implemented for Ad
1778 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1779 stage table and inline lookup instead of a function, giving speed ups of 2
1780 to 5 times on some simple patterns that I tested. Permission was given to
1781 distribute the MultiStage2.py script that generates the tables (it's not in
1782 the tarball, but is in the Subversion repository).
1783
1784 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1785 scripts.
1786
1787 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1788 a group with a zero qualifier. The result of the study could be incorrect,
1789 or the function might crash, depending on the pattern.
1790
1791 4. Caseless matching was not working for non-ASCII characters in back
1792 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1793 It now works when Unicode Property Support is available.
1794
1795 5. In pcretest, an escape such as \x{de} in the data was always generating
1796 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1797 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1798 truncation.
1799
1800 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1801
1802 7. Added two (int) casts to pcregrep when printing the difference of two
1803 pointers, in case they are 64-bit values.
1804
1805 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1806 test 2 if it fails.
1807
1808 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1809 and a #define of that name to empty if it is not externally set. This is to
1810 allow users of MSVC to set it if necessary.
1811
1812 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1813 the convenience functions in the pcre_get.c source file.
1814
1815 11. An option change at the start of a pattern that had top-level alternatives
1816 could cause overwriting and/or a crash. This command provoked a crash in
1817 some environments:
1818
1819 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1820
1821 This potential security problem was recorded as CVE-2008-2371.
1822
1823 12. For a pattern where the match had to start at the beginning or immediately
1824 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1825 pcre_dfa_exec() could read past the end of the passed subject if there was
1826 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1827 pcretest so that it places the subject at the end of its malloc-ed buffer.
1828
1829 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1830 exec() might read past the end of the data buffer in UTF-8 mode.
1831
1832 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1833 the data contained the byte 0x85 as part of a UTF-8 character within its
1834 first line. This applied both to normal and DFA matching.
1835
1836 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1837 /^[^d]*?$/8 failed to match "abc".
1838
1839 16. Added a missing copyright notice to pcrecpp_internal.h.
1840
1841 17. Make it more clear in the documentation that values returned from
1842 pcre_exec() in ovector are byte offsets, not character counts.
1843
1844 18. Tidied a few places to stop certain compilers from issuing warnings.
1845
1846 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1847 supplied by Stefan Weber. I made a further small update for 7.8 because
1848 there is a change of source arrangements: the pcre_searchfuncs.c module is
1849 replaced by pcre_ucd.c.
1850
1851
1852 Version 7.7 07-May-08
1853 ---------------------
1854
1855 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1856 a string to a long long, pretend we don't even have a long long." This is
1857 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1858
1859 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1860 pre-7.6 versions, which defined a global no_arg variable instead of putting
1861 it in the RE class. (See also #8 below.)
1862
1863 3. Remove a line of dead code, identified by coverity and reported by Nuno
1864 Lopes.
1865
1866 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1867
1868 (1) The include/exclude patterns were being applied to the whole pathnames
1869 of files, instead of just to the final components.
1870
1871 (2) If there was more than one level of directory, the subdirectories were
1872 skipped unless they satisfied the include/exclude conditions. This is
1873 inconsistent with GNU grep (and could even be seen as contrary to the
1874 pcregrep specification - which I improved to make it absolutely clear).
1875 The action now is always to scan all levels of directory, and just
1876 apply the include/exclude patterns to regular files.
1877
1878 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1879 --exclude_dir in the tests to avoid scanning .svn directories.
1880
1881 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1882 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1883 doesn't support NULs in patterns.
1884
1885 7. Added some missing "const"s to declarations of static tables in
1886 pcre_compile.c and pcre_dfa_exec.c.
1887
1888 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1889 caused by fix #2 above. (Subsequently also a second patch to fix the
1890 first patch. And a third patch - this was a messy problem.)
1891
1892 9. Applied Craig's patch to remove the use of push_back().
1893
1894 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1895 matching function regexec().
1896
1897 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1898 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1899 references. PCRE supports relative numbers with this syntax (I don't think
1900 Oniguruma does).
1901
1902 12. Previously, a group with a zero repeat such as (...){0} was completely
1903 omitted from the compiled regex. However, this means that if the group
1904 was called as a subroutine from elsewhere in the pattern, things went wrong
1905 (an internal error was given). Such groups are now left in the compiled
1906 pattern, with a new opcode that causes them to be skipped at execution
1907 time.
1908
1909 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1910 to the way PCRE behaves:
1911
1912 (a) A lone ] character is dis-allowed (Perl treats it as data).
1913
1914 (b) A back reference to an unmatched subpattern matches an empty string
1915 (Perl fails the current match path).
1916
1917 (c) A data ] in a character class must be notated as \] because if the
1918 first data character in a class is ], it defines an empty class. (In
1919 Perl it is not possible to have an empty class.) The empty class []
1920 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1921 The negative empty class [^] matches any one character, independently
1922 of the DOTALL setting.
1923
1924 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1925 non-existent subpattern following a character class starting with ']' and
1926 containing () gave an internal compiling error instead of "reference to
1927 non-existent subpattern". Fortunately, when the pattern did exist, the
1928 compiled code was correct. (When scanning forwards to check for the
1929 existence of the subpattern, it was treating the data ']' as terminating
1930 the class, so got the count wrong. When actually compiling, the reference
1931 was subsequently set up correctly.)
1932
1933 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1934 it was being rejected as not supported by pcre_dfa_exec(), even though
1935 other assertions are supported. I have made pcre_dfa_exec() support
1936 (*FAIL).
1937
1938 16. The implementation of 13c above involved the invention of a new opcode,
1939 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1940 cannot be changed at match time, I realized I could make a small
1941 improvement to matching performance by compiling OP_ALLANY instead of
1942 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1943 on the OP_ANY path.
1944
1945 17. Compiling pcretest on Windows with readline support failed without the
1946 following two fixes: (1) Make the unistd.h include conditional on
1947 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1948
1949 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1950 ncurses library to be included for pcretest when ReadLine support is
1951 requested, but also to allow for it to be overridden. This patch came from
1952 Daniel Bergström.
1953
1954 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1955 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1956 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1957 spotting this.
1958
1959
1960 Version 7.6 28-Jan-08
1961 ---------------------
1962
1963 1. A character class containing a very large number of characters with
1964 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1965 overflow.
1966
1967 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1968 HAVE_LONG_LONG is not defined.
1969
1970 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1971 bring them up to date and include new features. This patch includes:
1972
1973 - Fixed PH's badly added libz and libbz2 support.
1974 - Fixed a problem with static linking.
1975 - Added pcredemo. [But later removed - see 7 below.]
1976 - Fixed dftables problem and added an option.
1977 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1978 HAVE_LONG_LONG.
1979 - Added readline support for pcretest.
1980 - Added an listing of the option settings after cmake has run.
1981
1982 4. A user submitted a patch to Makefile that makes it easy to create
1983 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1984 Makefile.am that cause it to include this special target, without
1985 affecting anything else. Note that the same mingw target plus all
1986 the other distribution libraries and programs are now supported
1987 when configuring with CMake (see 6 below) instead of with
1988 Configure/Make.
1989
1990 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1991 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1992 exported in the Windows port". It has not yet been confirmed that the patch
1993 solves the problem, but it does no harm.
1994
1995 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1996 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1997 with CMake, and also correct the comment about stack recursion.
1998
1999 7. Remove the automatic building of pcredemo from the ./configure system and
2000 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2001 of a program that users should build themselves after PCRE is installed, so
2002 building it automatically is not really right. What is more, it gave
2003 trouble in some build environments.
2004
2005 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2006
2007
2008 Version 7.5 10-Jan-08
2009 ---------------------
2010
2011 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2012 values in parens when parsing an RE using the C++ wrapper."
2013
2014 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2015 Characters greater than 255 were excluded from the class instead of being
2016 included.
2017
2018 3. The same bug as (2) above applied to negated POSIX classes such as
2019 [:^space:].
2020
2021 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2022 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2023 I have changed it.
2024
2025 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2026 first named subpattern) and a construct such as (?&a) would reference the
2027 first named subpattern whose name started with "a" (in other words, the
2028 length check was missing). Both these problems are fixed. "Subpattern name
2029 expected" is now given for (?&) (a zero-length name), and this patch also
2030 makes it give the same error for \k'' (previously it complained that that
2031 was a reference to a non-existent subpattern).
2032
2033 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2034 this is right because (?- can be followed by option settings as well as by
2035 digits. I have, however, made the messages clearer.
2036
2037 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2038 than the number used in the conditional) now cause a compile-time error.
2039 This is actually not compatible with Perl, which accepts such patterns, but
2040 treats the conditional as always being FALSE (as PCRE used to), but it
2041 seems to me that giving a diagnostic is better.
2042
2043 8. Change "alphameric" to the more common word "alphanumeric" in comments
2044 and messages.
2045
2046 9. Fix two occurrences of "backslash" in comments that should have been
2047 "backspace".
2048
2049 10. Remove two redundant lines of code that can never be obeyed (their function
2050 was moved elsewhere).
2051
2052 11. The program that makes PCRE's Unicode character property table had a bug
2053 which caused it to generate incorrect table entries for sequences of
2054 characters that have the same character type, but are in different scripts.
2055 It amalgamated them into a single range, with the script of the first of
2056 them. In other words, some characters were in the wrong script. There were
2057 thirteen such cases, affecting characters in the following ranges:
2058
2059 U+002b0 - U+002c1
2060 U+0060c - U+0060d
2061 U+0061e - U+00612
2062 U+0064b - U+0065e
2063 U+0074d - U+0076d
2064 U+01800 - U+01805
2065 U+01d00 - U+01d77
2066 U+01d9b - U+01dbf
2067 U+0200b - U+0200f
2068 U+030fc - U+030fe
2069 U+03260 - U+0327f
2070 U+0fb46 - U+0fbb1
2071 U+10450 - U+1049d
2072
2073 12. The -o option (show only the matching part of a line) for pcregrep was not
2074 compatible with GNU grep in that, if there was more than one match in a
2075 line, it showed only the first of them. It now behaves in the same way as
2076 GNU grep.
2077
2078 13. If the -o and -v options were combined for pcregrep, it printed a blank
2079 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2080 does the same. The return code can be used to tell if there were any
2081 non-matching lines.
2082
2083 14. Added --file-offsets and --line-offsets to pcregrep.
2084
2085 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2086 infinitely looping recursion. The bug was that positive lookaheads were not
2087 being skipped when checking for a possible empty match (negative lookaheads
2088 and both kinds of lookbehind were skipped).
2089
2090 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2091 inclusion of <windows.h> to before rather than after the definition of
2092 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2093
2094 17. Specifying a possessive quantifier with a specific limit for a Unicode
2095 character property caused pcre_compile() to compile bad code, which led at
2096 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2097 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2098 caused the error; without that there was no problem.
2099
2100 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2101
2102 19. Added --enable-pcretest-libreadline.
2103
2104 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2105 RE::GlobalReplace(). As a result, the number of replacements returned was
2106 double what it should be. I removed one of the increments, but Craig sent a
2107 later patch that removed the other one (the right fix) and added unit tests
2108 that check the return values (which was not done before).
2109
2110 21. Several CMake things:
2111
2112 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2113 the names libpcre and libpcreposix, not just pcre and pcreposix.
2114
2115 (2) The above change means that pcretest and pcregrep are now correctly
2116 linked with the newly-built libraries, not previously installed ones.
2117
2118 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2119
2120 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2121 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2122 UTF-8 newline character). The key issue is that the pattern starts .*;
2123 this means that the match must be either at the beginning, or after a
2124 newline. The bug was in the code for advancing after a failed match and
2125 checking that the new position followed a newline. It was not taking
2126 account of UTF-8 characters correctly.
2127
2128 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2129 character classes. PCRE was not treating the sequence [:...:] as a
2130 character class unless the ... were all letters. Perl, however, seems to
2131 allow any characters between [: and :], though of course it rejects as
2132 unknown any "names" that contain non-letters, because all the known class
2133 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2134 for example, whereas PCRE did not - it did not recognize a POSIX character
2135 class. This seemed a bit dangerous, so the code has been changed to be
2136 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2137 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2138 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2139 Perl does, and where it didn't before.
2140
2141 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2142 Windows environments %n is disabled by default.
2143
2144
2145 Version 7.4 21-Sep-07
2146 ---------------------
2147
2148 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2149 means that a class such as [\s] counted as "explicit reference to CR or
2150 LF". That isn't really right - the whole point of the change was to try to
2151 help when there was an actual mention of one of the two characters. So now
2152 the change happens only if \r or \n (or a literal CR or LF) character is
2153 encountered.
2154
2155 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2156 of both had grown to the point where there were only 3 bits left.
2157 Fortunately, there was spare space in the data structure, and so I have
2158 moved the internal flags into a new 16-bit field to free up more option
2159 bits.
2160
2161 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2162 but did not set the internal JCHANGED flag - either of these is enough to
2163 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2164 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2165 start sets both bits.
2166
2167 4. Added options (at build time, compile time, exec time) to change \R from
2168 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2169
2170 5. doc/pcresyntax.html was missing from the distribution.
2171
2172 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2173 compatibility, even though it is no longer used.
2174
2175 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2176 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2177 windows.h file is present (where different names are used). [This was
2178 reversed later after testing - see 16 below.]
2179
2180 8. Changed all #include <config.h> to #include "config.h". There were also
2181 some further <pcre.h> cases that I changed to "pcre.h".
2182
2183 9. When pcregrep was used with the --colour option, it missed the line ending
2184 sequence off the lines that it output.
2185
2186 10. It was pointed out to me that arrays of string pointers cause lots of
2187 relocations when a shared library is dynamically loaded. A technique of
2188 using a single long string with a table of offsets can drastically reduce
2189 these. I have refactored PCRE in four places to do this. The result is
2190 dramatic:
2191
2192 Originally: 290
2193 After changing UCP table: 187
2194 After changing error message table: 43
2195 After changing table of "verbs" 36
2196 After changing table of Posix names 22
2197
2198 Thanks to the folks working on Gregex for glib for this insight.
2199
2200 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2201 unicode-properties was also set.
2202
2203 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2204
2205 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2206 checked only for CRLF.
2207
2208 14. Added casts to pcretest.c to avoid compiler warnings.
2209
2210 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2211
2212 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2213 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2214 entirely. This removes changes made in 7 above.
2215
2216 17. The CMake files have been updated, and there is now more information about
2217 building with CMake in the NON-UNIX-USE document.
2218
2219
2220 Version 7.3 28-Aug-07
2221 ---------------------
2222
2223 1. In the rejigging of the build system that eventually resulted in 7.1, the
2224 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2225 brackets there is not right, since it causes compilers to look for an
2226 installed pcre.h, not the version that is in the source that is being
2227 compiled (which of course may be different). I have changed it back to:
2228
2229 #include "pcre.h"
2230
2231 I have a vague recollection that the change was concerned with compiling in
2232 different directories, but in the new build system, that is taken care of
2233 by the VPATH setting the Makefile.
2234
2235 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2236 when the subject happened to end in the byte 0x85 (e.g. if the last
2237 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2238 characters but of course it shouldn't be taken as a newline when it is part
2239 of another character. The bug was that, for an unlimited repeat of . in
2240 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2241 characters when looking for a newline.
2242
2243 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2244
2245 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2246 in debug output.
2247
2248 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2249 long printing in the pcrecpp unittest when running under MinGW.
2250
2251 6. ESC_K was left out of the EBCDIC table.
2252
2253 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2254 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2255 limit also applies to "virtual nesting" when a pattern is recursive, and in
2256 this case 1000 isn't so big. I have been able to remove this limit at the
2257 expense of backing off one optimization in certain circumstances. Normally,
2258 when pcre_exec() would call its internal match() function recursively and
2259 immediately return the result unconditionally, it uses a "tail recursion"
2260 feature to save stack. However, when a subpattern that can match an empty
2261 string has an unlimited repetition quantifier, it no longer makes this
2262 optimization. That gives it a stack frame in which to save the data for
2263 checking that an empty string has been matched. Previously this was taken
2264 from the 1000-entry workspace that had been reserved. So now there is no
2265 explicit limit, but more stack is used.
2266
2267 8. Applied Daniel's patches to solve problems with the import/export magic
2268 syntax that is required for Windows, and which was going wrong for the
2269 pcreposix and pcrecpp parts of the library. These were overlooked when this
2270 problem was solved for the main library.
2271
2272 9. There were some crude static tests to avoid integer overflow when computing
2273 the size of patterns that contain repeated groups with explicit upper
2274 limits. As the maximum quantifier is 65535, the maximum group length was
2275 set at 30,000 so that the product of these two numbers did not overflow a
2276 32-bit integer. However, it turns out that people want to use groups that
2277 are longer than 30,000 bytes (though not repeat them that many times).
2278 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2279 made it possible to implement the integer overflow checks in a much more
2280 dynamic way, which I have now done. The artificial limitation on group
2281 length has been removed - we now have only the limit on the total length of
2282 the compiled pattern, which depends on the LINK_SIZE setting.
2283
2284 10. Fixed a bug in the documentation for get/copy named substring when
2285 duplicate names are permitted. If none of the named substrings are set, the
2286 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2287 empty string.
2288
2289 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2290 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2291 because the ] is interpreted as the first data character and the
2292 terminating ] is not found. PCRE has been made compatible with Perl in this
2293 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2294 cause memory overwriting.
2295
2296 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2297 string has been matched (to stop an infinite loop). It was not recognizing
2298 a conditional subpattern that could match an empty string if that
2299 subpattern was within another subpattern. For example, it looped when
2300 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2301 condition was not nested. This bug has been fixed.
2302
2303 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2304 past the start of the subject in the presence of bytes with the top bit
2305 set, for example "\x8aBCD".
2306
2307 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2308 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2309
2310 14. Optimized (?!) to (*FAIL).
2311
2312 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2313 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2314 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2315 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2316 does: it's just the validity check that is more restrictive.
2317
2318 16. Inserted checks for integer overflows during escape sequence (backslash)
2319 processing, and also fixed erroneous offset values for syntax errors during
2320 backslash processing.
2321
2322 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2323 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2324
2325 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2326 caused an overrun.
2327
2328 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2329 something other than just ASCII characters) inside a group that had an
2330 unlimited repeat caused a loop at compile time (while checking to see
2331 whether the group could match an empty string).
2332
2333 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2334 [\P{Any}] did so. (Error in the code for printing property names.)
2335
2336 21. An orphan \E inside a character class could cause a crash.
2337
2338 22. A repeated capturing bracket such as (A)? could cause a wild memory
2339 reference during compilation.
2340
2341 23. There are several functions in pcre_compile() that scan along a compiled
2342 expression for various reasons (e.g. to see if it's fixed length for look
2343 behind). There were bugs in these functions when a repeated \p or \P was
2344 present in the pattern. These operators have additional parameters compared
2345 with \d, etc, and these were not being taken into account when moving along
2346 the compiled data. Specifically:
2347
2348 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2349 length.
2350
2351 (b) An item such as \pL+ within a repeated group could cause crashes or
2352 loops.
2353
2354 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2355 "reference to non-existent subpattern" error.
2356
2357 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2358
2359 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2360 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2361
2362 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2363
2364 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2365 character were causing crashes (broken optimization).
2366
2367 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2368 \p or \P) caused a compile-time loop.
2369
2370 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2371 break. For example, the unstudied pattern [\r\n]A does not match the string
2372 "\r\nA" because change 7.0/46 below moves the current point on by two
2373 characters after failing to match at the start. However, the pattern \nA
2374 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2375 the same is true. There doesn't seem any very clean way out of this, but
2376 what I have chosen to do makes the common cases work: PCRE now takes note
2377 of whether there can be an explicit match for \r or \n anywhere in the
2378 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2379 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2380 pattern has explicit CR or LF references.
2381
2382 29. Added (*CR) etc for changing newline setting at start of pattern.
2383
2384
2385 Version 7.2 19-Jun-07
2386 ---------------------
2387
2388 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2389 which is apparently normally available under Windows.
2390
2391 2. Re-jig the pcregrep tests with different newline settings in an attempt
2392 to make them independent of the local environment's newline setting.
2393
2394 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2395
2396 4. Some of the "internals" tests were previously cut out when the link size
2397 was not 2, because the output contained actual offsets. The recent new
2398 "Z" feature of pcretest means that these can be cut out, making the tests
2399 usable with all link sizes.
2400
2401 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2402 stack recursion. This gives a massive performance boost under BSD, but just
2403 a small improvement under Linux. However, it saves one field in the frame
2404 in all cases.
2405
2406 6. Added more features from the forthcoming Perl 5.10:
2407
2408 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2409 recursion call. It refers to the nth most recently opened parentheses.
2410
2411 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2412 to be opened parentheses.
2413
2414 (c) Conditions that refer to capturing parentheses can be specified
2415 relatively, for example, (?(-2)... or (?(+3)...
2416
2417 (d) \K resets the start of the current match so that everything before
2418 is not part of it.
2419
2420 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2421
2422 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2423 reference syntax.
2424
2425 (g) (?| introduces a group in which the numbering of parentheses in each
2426 alternative starts with the same number.
2427
2428 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2429
2430 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2431 PCRE_INFO_JCHANGED.
2432
2433 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2434 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2435 for detecting groups that can match an empty string.
2436
2437 9. A pattern with a very large number of alternatives (more than several
2438 hundred) was running out of internal workspace during the pre-compile
2439 phase, where pcre_compile() figures out how much memory will be needed. A
2440 bit of new cunning has reduced the workspace needed for groups with
2441 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2442 workspace instead of running out of the 4096 that are available.
2443
2444 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2445
2446 11. Applied patch from Google to remove an optimization that didn't quite work.
2447 The report of the bug said:
2448
2449 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2450 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2451 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2452
2453 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2454 it matched the wrong number of bytes.
2455
2456
2457 Version 7.1 24-Apr-07
2458 ---------------------
2459
2460 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2461 that is more "standard", making use of automake and other Autotools. There
2462 is some re-arrangement of the files and adjustment of comments consequent
2463 on this.
2464
2465 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2466 for recursive directory scanning broke on some systems because the files
2467 are not scanned in any specific order and on different systems the order
2468 was different. A call to "sort" has been inserted into RunGrepTest for the
2469 approprate test as a short-term fix. In the longer term there may be an
2470 alternative.
2471
2472 3. I had an email from Eric Raymond about problems translating some of PCRE's
2473 man pages to HTML (despite the fact that I distribute HTML pages, some
2474 people do their own conversions for various reasons). The problems
2475 concerned the use of low-level troff macros .br and .in. I have therefore
2476 removed all such uses from the man pages (some were redundant, some could
2477 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2478 HTML has been updated to handle .nf/.fi and to complain if it encounters
2479 .br or .in.
2480
2481 4. Updated comments in configure.ac that get placed in config.h.in and also
2482 arranged for config.h to be included in the distribution, with the name
2483 config.h.generic, for the benefit of those who have to compile without
2484 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2485
2486 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2487 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2488 makevp.bat for the current PCRE, using the additional files
2489 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2490
2491 6. A Windows user reported a minor discrepancy with test 2, which turned out
2492 to be caused by a trailing space on an input line that had got lost in his
2493 copy. The trailing space was an accident, so I've just removed it.
2494
2495 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2496 that is needed.
2497
2498 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2499 as "const" (a) because they are and (b) because it helps the PHP
2500 maintainers who have recently made a script to detect big data structures
2501 in the php code that should be moved to the .rodata section. I remembered
2502 to update Builducptable as well, so it won't revert if ucptable.h is ever
2503 re-created.
2504
2505 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2506 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2507 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2508 support is not required. This saves 1.5-2K of code, which is important in
2509 some applications.
2510
2511 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2512 so as not to refer to the tables, even though these functions will never be
2513 called when UTF-8 support is disabled. Otherwise there are problems with a
2514 shared library.
2515
2516 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2517
2518 (a) It was defining its arguments as char * instead of void *.
2519
2520 (b) It was assuming that all moves were upwards in memory; this was true
2521 a long time ago when I wrote it, but is no longer the case.
2522
2523 The emulated memove() is provided for those environments that have neither
2524 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2525 is clearly not the case, as these two bugs were recently reported.
2526
2527 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2528 and Detrail to create the HTML documentation, the .txt form of the man
2529 pages, and it removes trailing spaces from listed files. It also creates
2530 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2531 case, it wraps all the #defines with #ifndefs. This script should be run
2532 before "make dist".
2533
2534 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2535 with Unicode property support.
2536
2537 (a) For a maximizing quantifier, if the two different cases of the
2538 character were of different lengths in their UTF-8 codings (there are
2539 some cases like this - I found 11), and the matching function had to
2540 back up over a mixture of the two cases, it incorrectly assumed they
2541 were both the same length.
2542
2543 (b) When PCRE was configured to use the heap rather than the stack for
2544 recursion during matching, it was not correctly preserving the data for
2545 the other case of a UTF-8 character when checking ahead for a match
2546 while processing a minimizing repeat. If the check also involved
2547 matching a wide character, but failed, corruption could cause an
2548 erroneous result when trying to check for a repeat of the original
2549 character.
2550
2551 13. Some tidying changes to the testing mechanism:
2552
2553 (a) The RunTest script now detects the internal link size and whether there
2554 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2555 values substituted by "configure". (The RunGrepTest script already did
2556 this for UTF-8.) The configure.ac script no longer substitutes the
2557 relevant variables.
2558
2559 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2560 with length and offset values. This means that the output is different
2561 for different internal link sizes. Test 2 is skipped for link sizes
2562 other than 2 because of this, bypassing the problem. Unfortunately,
2563 there was also a test in test 3 (the locale tests) that used /B and
2564 failed for link sizes other than 2. Rather than cut the whole test out,
2565 I have added a new /Z option to pcretest that replaces the length and
2566 offset values with spaces. This is now used to make test 3 independent
2567 of link size. (Test 2 will be tidied up later.)
2568
2569 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2570 segmentation fault instead of returning the appropriate error message.
2571
2572 15. In multiline mode when the newline sequence was set to "any", the pattern
2573 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2574 This doesn't seem right; it now treats the CRLF combination as the line
2575 ending, and so does not match in that case. It's only a pattern such as ^$
2576 that would hit this one: something like ^ABC$ would have failed after \r
2577 and then tried again after \r\n.
2578
2579 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2580 in an attempt to make files that differ only in their line terminators
2581 compare equal. This works on Linux.
2582
2583 17. Under certain error circumstances pcregrep might try to free random memory
2584 as it exited. This is now fixed, thanks to valgrind.
2585
2586 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2587 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2588 was because its rules for how to advance for /g after matching an empty
2589 string at the end of a line did not allow for this case. They now check for
2590 it specially.
2591
2592 20. pcretest is supposed to handle patterns and data of any length, by
2593 extending its buffers when necessary. It was getting this wrong when the
2594 buffer for a data line had to be extended.
2595
2596 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2597 CRLF as a newline sequence.
2598
2599 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2600 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2601 I have nevertheless tidied it up.
2602
2603 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2604
2605 24. Added a man page for pcre-config.
2606
2607
2608 Version 7.0 19-Dec-06
2609 ---------------------
2610
2611 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2612 moving to gcc 4.1.1.
2613
2614 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2615 sys/time.h, which is documented as needed for this function. It doesn't
2616 seem to matter on Linux, but it showed up on some releases of OS X.
2617
2618 3. It seems that there are systems where bytes whose values are greater than
2619 127 match isprint() in the "C" locale. The "C" locale should be the
2620 default when a C program starts up. In most systems, only ASCII printing
2621 characters match isprint(). This difference caused the output from pcretest
2622 to vary, making some of the tests fail. I have changed pcretest so that:
2623
2624 (a) When it is outputting text in the compiled version of a pattern, bytes
2625 other than 32-126 are always shown as hex escapes.
2626
2627 (b) When it is outputting text that is a matched part of a subject string,
2628 it does the same, unless a different locale has been set for the match
2629 (using the /L modifier). In this case, it uses isprint() to decide.
2630
2631 4. Fixed a major bug that caused incorrect computation of the amount of memory
2632 required for a compiled pattern when options that changed within the
2633 pattern affected the logic of the preliminary scan that determines the
2634 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2635 that the computed length was too small. The symptoms of this bug were
2636 either the PCRE error "internal error: code overflow" from pcre_compile(),
2637 or a glibc crash with a message such as "pcretest: free(): invalid next
2638 size (fast)". Examples of patterns that provoked this bug (shown in
2639 pcretest format) are:
2640
2641 /(?-x: )/x
2642 /(?x)(?-x: \s*#\s*)/
2643 /((?i)[\x{c0}])/8
2644 /(?i:[\x{c0}])/8
2645
2646 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2647 is now done differently.
2648
2649 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2650 wrapper classes; (b) implement a new function in the C++ scanner that is
2651 more efficient than the old way of doing things because it avoids levels of
2652 recursion in the regex matching; (c) add a paragraph to the documentation
2653 for the FullMatch() function.
2654
2655 6. The escape sequence \n was being treated as whatever was defined as
2656 "newline". Not only was this contrary to the documentation, which states
2657 that \n is character 10 (hex 0A), but it also went horribly wrong when
2658 "newline" was defined as CRLF. This has been fixed.
2659
2660 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2661 was being set to -1 for the "end of line" case (supposedly a value that no
2662 character can have). Though this value is never used (the check for end of
2663 line is "zero bytes in current character"), it caused compiler complaints.
2664 I've changed it to 0xffffffff.
2665
2666 8. In pcre_version.c, the version string was being built by a sequence of
2667 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2668 string (as it is for production releases) called a macro with an empty
2669 argument. The C standard says the result of this is undefined. The gcc
2670 compiler treats it as an empty string (which was what was wanted) but it is
2671 reported that Visual C gives an error. The source has been hacked around to
2672 avoid this problem.
2673
2674 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2675 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2676 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2677 of them did).
2678
2679 10. Originally, pcretest opened its input and output without "b"; then I was
2680 told that "b" was needed in some environments, so it was added for release
2681 5.0 to both the input and output. (It makes no difference on Unix-like
2682 systems.) Later I was told that it is wrong for the input on Windows. I've
2683 now abstracted the modes into two macros, to make it easier to fiddle with
2684 them, and removed "b" from the input mode under Windows.
2685
2686 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2687
2688 12. Added -help and --help to pcretest as an official way of being reminded
2689 of the options.
2690
2691 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2692 and pcrecpp.cc because they annoy compilers at high warning levels.
2693
2694 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2695
2696 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2697 scripts are not C programs :-) and which was not noticed because it works
2698 on Linux.
2699
2700 16. pcretest is supposed to handle any length of pattern and data line (as one
2701 line or as a continued sequence of lines) by extending its input buffer if
2702 necessary. This feature was broken for very long pattern lines, leading to
2703 a string of junk being passed to pcre_compile() if the pattern was longer
2704 than about 50K.
2705
2706 17. I have done a major re-factoring of the way pcre_compile() computes the
2707 amount of memory needed for a compiled pattern. Previously, there was code
2708 that made a preliminary scan of the pattern in order to do this. That was
2709 OK when PCRE was new, but as the facilities have expanded, it has become
2710 harder and harder to keep it in step with the real compile phase, and there
2711 have been a number of bugs (see for example, 4 above). I have now found a
2712 cunning way of running the real compile function in a "fake" mode that
2713 enables it to compute how much memory it would need, while actually only
2714 ever using a few hundred bytes of working memory and without too many
2715 tests of the mode. This should make future maintenance and development
2716 easier. A side effect of this work is that the limit of 200 on the nesting
2717 depth of parentheses has been removed (though this was never a serious
2718 limitation, I suspect). However, there is a downside: pcre_compile() now
2719 runs more slowly than before (30% or more, depending on the pattern). I
2720 hope this isn't a big issue. There is no effect on runtime performance.
2721
2722 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2723 newline (only possible for the last line of a file) and it was a
2724 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2725
2726 19. Added additional timing features to pcretest. (1) The -tm option now times
2727 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2728 separate command line item, by a number that specifies the number of
2729 repeats to use when timing. The default is 50000; this gives better
2730 precision, but takes uncomfortably long for very large patterns.
2731
2732 20. Extended pcre_study() to be more clever in cases where a branch of a
2733 subpattern has no definite first character. For example, (a*|b*)[cd] would
2734 previously give no result from pcre_study(). Now it recognizes that the
2735 first character must be a, b, c, or d.
2736
2737 21. There was an incorrect error "recursive call could loop indefinitely" if
2738 a subpattern (or the entire pattern) that was being tested for matching an
2739 empty string contained only one non-empty item after a nested subpattern.
2740 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2741 incorrectly, because the \d was being skipped in the check.
2742
2743 22. The pcretest program now has a new pattern option /B and a command line
2744 option -b, which is equivalent to adding /B to every pattern. This causes
2745 it to show the compiled bytecode, without the additional information that
2746 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2747 is the same as /B/I).
2748
2749 23. A new optimization is now able automatically to treat some sequences such
2750 as a*b as a*+b. More specifically, if something simple (such as a character
2751 or a simple class like \d) has an unlimited quantifier, and is followed by
2752 something that cannot possibly match the quantified thing, the quantifier
2753 is automatically "possessified".
2754
2755 24. A recursive reference to a subpattern whose number was greater than 39
2756 went wrong under certain circumstances in UTF-8 mode. This bug could also
2757 have affected the operation of pcre_study().
2758
2759 25. Realized that a little bit of performance could be had by replacing
2760 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2761
2762 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2763
2764 27. Possessive quantifiers such as a++ were previously implemented by turning
2765 them into atomic groups such as ($>a+). Now they have their own opcodes,
2766 which improves performance. This includes the automatically created ones
2767 from 23 above.
2768
2769 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2770 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2771 the first matched character to be a colon. This applied both to named and
2772 numbered groups.
2773
2774 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2775
2776 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2777 building PCRE on Windows easier, so I have included it in the distribution.
2778
2779 31. There is now a check in pcretest against a ridiculously large number being
2780 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2781 loop, the loop is abandoned.
2782
2783 32. Forward references to subpatterns in conditions such as (?(2)...) where
2784 subpattern 2 is defined later cause pcre_compile() to search forwards in
2785 the pattern for the relevant set of parentheses. This search went wrong
2786 when there were unescaped parentheses in a character class, parentheses
2787 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2788
2789 33. "Subroutine" calls and backreferences were previously restricted to
2790 referencing subpatterns earlier in the regex. This restriction has now
2791 been removed.
2792
2793 34. Added a number of extra features that are going to be in Perl 5.10. On the
2794 whole, these are just syntactic alternatives for features that PCRE had
2795 previously implemented using the Python syntax or my own invention. The
2796 other formats are all retained for compatibility.
2797
2798 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2799 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2800 also .NET compatible.
2801
2802 (b) A recursion or subroutine call to a named group can now be defined as
2803 (?&name) as well as (?P>name).
2804
2805 (c) A backreference to a named group can now be defined as \k<name> or
2806 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2807 5.10, are also .NET compatible.
2808
2809 (d) A conditional reference to a named group can now use the syntax
2810 (?(<name>) or (?('name') as well as (?(name).
2811
2812 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2813 groups (named and numbered) that are never evaluated inline, but can be
2814 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2815 is always false. There may be only one alternative in such a group.
2816
2817 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2818 as the simple (?(R). The condition is true only if the most recent
2819 recursion is that of the given number or name. It does not search out
2820 through the entire recursion stack.
2821
2822 (g) The escape \gN or \g{N} has been added, where N is a positive or
2823 negative number, specifying an absolute or relative reference.
2824
2825 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2826 some "unreachable code" warnings.
2827
2828 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2829 things, this adds five new scripts.
2830
2831 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2832 There were also incompatibilities regarding the handling of \Q..\E inside
2833 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2834 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2835
2836 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2837 matches an empty string, and forcibly breaks the loop. There were bugs in
2838 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2839 against aaaa the result was just "a" rather than "aaaa", for example. Two
2840 separate and independent bugs (that affected different cases) have been
2841 fixed.
2842
2843 39. Refactored the code to abolish the use of different opcodes for small
2844 capturing bracket numbers. This is a tidy that I avoided doing when I
2845 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2846 The new approach is not only tidier, it makes it possible to reduce the
2847 memory needed to fix the previous bug (38).
2848
2849 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2850 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2851 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2852 mode.
2853
2854 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2855 report.
2856
2857 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2858 copy construction and assignment in the C++ wrapper.
2859
2860 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2861 couple of bugs that could have given wrong results in the "--newline=crlf"
2862 case.
2863
2864 44. Added a number of casts and did some reorganization of signed/unsigned int
2865 variables following suggestions from Dair Grant. Also renamed the variable
2866 "this" as "item" because it is a C++ keyword.
2867
2868 45. Arranged for dftables to add
2869
2870 #include "pcre_internal.h"
2871
2872 to pcre_chartables.c because without it, gcc 4.x may remove the array
2873 definition from the final binary if PCRE is built into a static library and
2874 dead code stripping is activated.
2875
2876 46. For an unanchored pattern, if a match attempt fails at the start of a
2877 newline sequence, and the newline setting is CRLF or ANY, and the next two
2878 characters are CRLF, advance by two characters instead of one.
2879
2880
2881 Version 6.7 04-Jul-06
2882 ---------------------
2883
2884 1. In order to handle tests when input lines are enormously long, pcretest has
2885 been re-factored so that it automatically extends its buffers when
2886 necessary. The code is crude, but this _is_ just a test program. The
2887 default size has been increased from 32K to 50K.
2888
2889 2. The code in pcre_study() was using the value of the re argument before
2890 testing it for NULL. (Of course, in any sensible call of the function, it
2891 won't be NULL.)
2892
2893 3. The memmove() emulation function in pcre_internal.h, which is used on
2894 systems that lack both memmove() and bcopy() - that is, hardly ever -
2895 was missing a "static" storage class specifier.
2896
2897 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2898 containing an extended class (one that cannot be represented by a bitmap
2899 because it contains high-valued characters or Unicode property items, e.g.
2900 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2901 pattern, but PCRE should not loop if you do not (it no longer does).
2902 [Detail: two cases were found: (a) a repeated subpattern containing an
2903 extended class; (b) a recursive reference to a subpattern that followed a
2904 previous extended class. It wasn't skipping over the extended class
2905 correctly when UTF-8 mode was not set.]
2906
2907 5. A negated single-character class was not being recognized as fixed-length
2908 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2909 compile error "lookbehind assertion is not fixed length".
2910
2911 6. The RunPerlTest auxiliary script was showing an unexpected difference
2912 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2913 write a Perl script that can interpret lines of an input file either as
2914 byte characters or as UTF-8, which is what "perltest" was being required to
2915 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2916 can't do is switch easily at run time between having the "use utf8;" pragma
2917 or not. In the end, I fudged it by using the RunPerlTest script to insert
2918 "use utf8;" explicitly for the UTF-8 tests.
2919
2920 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2921 the end of the subject string, contrary to the documentation and to what
2922 Perl does. This was true of both matching functions. Now it matches only at
2923 the start of the subject and immediately after *internal* newlines.
2924
2925 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2926 a pointer to an int instead of a pointer to an unsigned long int. This
2927 caused problems on 64-bit systems.
2928
2929 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2930 instance of the 'standard' template library not being so standard".
2931
2932 10. There was no check on the number of named subpatterns nor the maximum
2933 length of a subpattern name. The product of these values is used to compute
2934 the size of the memory block for a compiled pattern. By supplying a very
2935 long subpattern name and a large number of named subpatterns, the size
2936 computation could be caused to overflow. This is now prevented by limiting
2937 the length of names to 32 characters, and the number of named subpatterns
2938 to 10,000.
2939
2940 11. Subpatterns that are repeated with specific counts have to be replicated in
2941 the compiled pattern. The size of memory for this was computed from the
2942 length of the subpattern and the repeat count. The latter is limited to
2943 65535, but there was no limit on the former, meaning that integer overflow
2944 could in principle occur. The compiled length of a repeated subpattern is
2945 now limited to 30,000 bytes in order to prevent this.
2946
2947 12. Added the optional facility to have named substrings with the same name.
2948
2949 13. Added the ability to use a named substring as a condition, using the
2950 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2951 are numbers (not recommended). Forward references are permitted.
2952
2953 14. Added forward references in named backreferences (if you see what I mean).
2954
2955 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2956 pattern could run off the end of the subject. For example, the pattern
2957 "(?s)(.{1,5})"8 did this with the subject "ab".
2958
2959 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2960 PCRE_CASELESS was set when matching characters that were quantified with ?
2961 or *.
2962
2963 17. A character class other than a single negated character that had a minimum
2964 but no maximum quantifier - for example [ab]{6,} - was not handled
2965 correctly by pce_dfa_exec(). It would match only one character.
2966
2967 18. A valid (though odd) pattern that looked like a POSIX character
2968 class but used an invalid character after [ (for example [[,abc,]]) caused
2969 pcre_compile() to give the error "Failed: internal error: code overflow" or
2970 in some cases to crash with a glibc free() error. This could even happen if
2971 the pattern terminated after [[ but there just happened to be a sequence of
2972 letters, a binary zero, and a closing ] in the memory that followed.
2973
2974 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2975 over the years. Originally (before any Unicode support), just the bottom 8
2976 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2977 output from "man perlunicode" includes this:
2978
2979 The regular expression compiler produces polymorphic opcodes. That
2980 is, the pattern adapts to the data and automatically switches to
2981 the Unicode character scheme when presented with Unicode data--or
2982 instead uses a traditional byte scheme when presented with byte
2983 data.
2984
2985 Sadly, a wide octal escape does not cause a switch, and in a string with
2986 no other multibyte characters, these octal escapes are treated as before.
2987 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2988 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2989 Unicode string.
2990
2991 I have not perpetrated such confusion in PCRE. Up till now, it took just
2992 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2993 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2994 translate to the appropriate multibyte character.
2995
2996 29. Applied some refactoring to reduce the number of warnings from Microsoft
2997 and Borland compilers. This has included removing the fudge introduced
2998 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2999 a warning about an unused variable.
3000
3001 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3002 characters since release 4.0, because Perl (from release 5.004) does not.
3003 [Or at least, is documented not to: some releases seem to be in conflict
3004 with the documentation.] However, when a pattern was studied with
3005 pcre_study() and all its branches started with \s, PCRE still included VT
3006 as a possible starting character. Of course, this did no harm; it just
3007 caused an unnecessary match attempt.
3008
3009 22. Removed a now-redundant internal flag bit that recorded the fact that case
3010 dependency changed within the pattern. This was once needed for "required
3011 byte" processing, but is no longer used. This recovers a now-scarce options
3012 bit. Also moved the least significant internal flag bit to the most-
3013 significant bit of the word, which was not previously used (hangover from
3014 the days when it was an int rather than a uint) to free up another bit for
3015 the future.
3016
3017 23. Added support for CRLF line endings as well as CR and LF. As well as the
3018 default being selectable at build time, it can now be changed at runtime
3019 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3020 specify that it is scanning data with non-default line endings.
3021
3022 24. Changed the definition of CXXLINK to make it agree with the definition of
3023 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3024
3025 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3026 recursions. This makes a big different to stack usage for some patterns.
3027
3028 26. If a subpattern containing a named recursion or subroutine reference such
3029 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3030 the space required for the compiled pattern went wrong and gave too small a
3031 value. Depending on the environment, this could lead to "Failed: internal
3032 error: code overflow at offset 49" or "glibc detected double free or
3033 corruption" errors.
3034
3035 27. Applied patches from Google (a) to support the new newline modes and (b) to
3036 advance over multibyte UTF-8 characters in GlobalReplace.
3037
3038 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3039 difference for some implementation of PCRE in some Windows version.
3040
3041 29. Added some extra testing facilities to pcretest:
3042
3043 \q<number> in a data line sets the "match limit" value
3044 \Q<number> in a data line sets the "match recursion limt" value
3045 -S <number> sets the stack size, where <number> is in megabytes
3046
3047 The -S option isn't available for Windows.
3048
3049
3050 Version 6.6 06-Feb-06
3051 ---------------------
3052
3053 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3054 in pcreposix.h. I have copied the definition from pcre.h.
3055
3056 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3057 because pcre.h is no longer a built file.
3058
3059 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3060 not normally included in the compiled code.
3061
3062
3063 Version 6.5 01-Feb-06
3064 ---------------------
3065
3066 1. When using the partial match feature with pcre_dfa_exec(), it was not
3067 anchoring the second and subsequent partial matches at the new starting
3068 point. This could lead to incorrect results. For example, with the pattern
3069 /1234/, partially matching against "123" and then "a4" gave a match.
3070
3071 2. Changes to pcregrep:
3072
3073 (a) All non-match returns from pcre_exec() were being treated as failures
3074 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3075 error message is output. Some extra information is given for the
3076 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3077 probably the only errors that are likely to be caused by users (by
3078 specifying a regex that has nested indefinite repeats, for instance).
3079 If there are more than 20 of these errors, pcregrep is abandoned.
3080
3081 (b) A binary zero was treated as data while matching, but terminated the
3082 output line if it was written out. This has been fixed: binary zeroes
3083 are now no different to any other data bytes.
3084
3085 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3086 used to set a locale for matching. The --locale=xxxx long option has
3087 been added (no short equivalent) to specify a locale explicitly on the
3088 pcregrep command, overriding the environment variables.
3089
3090 (d) When -B was used with -n, some line numbers in the output were one less
3091 than they should have been.
3092
3093 (e) Added the -o (--only-matching) option.
3094
3095 (f) If -A or -C was used with -c (count only), some lines of context were
3096 accidentally printed for the final match.
3097
3098 (g) Added the -H (--with-filename) option.
3099
3100 (h) The combination of options -rh failed to suppress file names for files
3101 that were found from directory arguments.
3102
3103 (i) Added the -D (--devices) and -d (--directories) options.
3104
3105 (j) Added the -F (--fixed-strings) option.
3106
3107 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3108
3109 (l) Added the --colo(u)r option.
3110
3111 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3112 is not present by default.
3113
3114 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3115 items such as (?R) or (?1), when the recursion could match a number of
3116 alternatives. If it matched one of the alternatives, but subsequently,
3117 outside the recursion, there was a failure, the code tried to back up into
3118 the recursion. However, because of the way PCRE is implemented, this is not
3119 possible, and the result was an incorrect result from the match.
3120
3121 In order to prevent this happening, the specification of recursion has
3122 been changed so that all such subpatterns are automatically treated as
3123 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3124
3125 4. I had overlooked the fact that, in some locales, there are characters for
3126 which isalpha() is true but neither isupper() nor islower() are true. In
3127 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3128 and ordfeminine) are like this. This affected the treatment of \w and \W
3129 when they appeared in character classes, but not when they appeared outside
3130 a character class. The bit map for "word" characters is now created
3131 separately from the results of isalnum() instead of just taking it from the
3132 upper, lower, and digit maps. (Plus the underscore character, of course.)
3133
3134 5. The above bug also affected the handling of POSIX character classes such as
3135 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3136 permanent tables. Instead, the bit maps for such a class were previously
3137 created as the appropriate unions of the upper, lower, and digit bitmaps.
3138 Now they are created by subtraction from the [[:word:]] class, which has
3139 its own bitmap.
3140
3141 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3142 It is created by subtracting the vertical space characters (\x09, \x0a,
3143 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3144 subtraction was done in the overall bitmap for a character class, meaning
3145 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3146 be recognized. This bug has been fixed.
3147
3148 7. Patches from the folks at Google:
3149
3150 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3151 real life, but is still worth protecting against".
3152
3153 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3154 regular expressions".
3155
3156 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3157 have it.
3158
3159 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3160 "configure" and the latter not, in order to fix a problem somebody had
3161 with compiling the Arg class on HP-UX.
3162
3163 (e) Improve the error-handling of the C++ wrapper a little bit.
3164
3165 (f) New tests for checking recursion limiting.
3166
3167 8. The pcre_memmove() function, which is used only if the environment does not
3168 have a standard memmove() function (and is therefore rarely compiled),
3169 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3170 returning a result (though PCRE never actually uses the result).
3171
3172 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3173 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3174 returned instead of calling malloc() with an overflowing number that would
3175 most likely cause subsequent chaos.
3176
3177 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3178
3179 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3180 with this option is matched, the nmatch and pmatch options of regexec() are
3181 ignored.
3182
3183 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3184 provided in case anyone wants to the the POSIX interface with UTF-8
3185 strings.
3186
3187 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3188 C++ linking (needed for some HP-UX environments).
3189
3190 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3191 (unused parameter) and in the pcre_printint() function (omitted "default"
3192 switch label when the default is to do nothing).
3193
3194 15. Added some code to make it possible, when PCRE is compiled as a C++
3195 library, to replace subject pointers for pcre_exec() with a smart pointer
3196 class, thus making it possible to process discontinuous strings.
3197
3198 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3199 much the same function. They were added by different people who were trying
3200 to make PCRE easy to compile on non-Unix systems. It has been suggested
3201 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3202 for compiling on Windows systems. I have therefore replaced it with
3203 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3204 defaults to "extern" for C or "extern C" for C++, which works fine on
3205 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3206 SCOPE with something explicit in config.h. In addition:
3207
3208 (a) pcreposix.h still had just "extern" instead of either of these macros;
3209 I have replaced it with PCRE_DATA_SCOPE.
3210
3211 (b) Functions such as _pcre_xclass(), which are internal to the library,
3212 but external in the C sense, all had PCRE_EXPORT in their definitions.
3213 This is apparently wrong for the Windows case, so I have removed it.
3214 (It makes no difference on Unix-like systems.)
3215
3216 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3217 of recursive calls to match(). This is different to MATCH_LIMIT because
3218 that limits the total number of calls to match(), not all of which increase
3219 the depth of recursion. Limiting the recursion depth limits the amount of
3220 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3221 when PCRE is compiled, and changed at run time. A patch from Google adds
3222 this functionality to the C++ interface.
3223
3224 18. Changes to the handling of Unicode character properties:
3225
3226 (a) Updated the table to Unicode 4.1.0.
3227
3228 (b) Recognize characters that are not in the table as "Cn" (undefined).
3229
3230 (c) I revised the way the table is implemented to a much improved format
3231 which includes recognition of ranges. It now supports the ranges that
3232 are defined in UnicodeData.txt, and it also amalgamates other
3233 characters into ranges. This has reduced the number of entries in the
3234 table from around 16,000 to around 3,000, thus reducing its size
3235 considerably. I realized I did not need to use a tree structure after
3236 all - a binary chop search is just as efficient. Having reduced the
3237 number of entries, I extended their size from 6 bytes to 8 bytes to
3238 allow for more data.
3239
3240 (d) Added support for Unicode script names via properties such as \p{Han}.
3241
3242 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3243 matching that character.
3244
3245 20. When matching a repeated Unicode property with a minimum greater than zero,
3246 (for example \pL{2,}), PCRE could look past the end of the subject if it
3247 reached it while seeking the minimum number of characters. This could
3248 happen only if some of the characters were more than one byte long, because
3249 there is a check for at least the minimum number of bytes.
3250
3251 21. Refactored the implementation of \p and \P so as to be more general, to
3252 allow for more different types of property in future. This has changed the
3253 compiled form incompatibly. Anybody with saved compiled patterns that use
3254 \p or \P will have to recompile them.
3255
3256 22. Added "Any" and "L&" to the supported property types.
3257
3258 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3259 but give a compile time error if the value is greater than 0xff.
3260
3261 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3262 accidentally not being installed or uninstalled.
3263
3264 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3265 made were to insert the current release number. This seemed silly, because
3266 it made things harder for people building PCRE on systems that don't run
3267 "configure". I have turned pcre.h into a distributed file, no longer built
3268 by "configure", with the version identification directly included. There is
3269 no longer a pcre.h.in file.
3270
3271 However, this change necessitated a change to the pcre-config script as
3272 well. It is built from pcre-config.in, and one of the substitutions was the
3273 release number. I have updated configure.ac so that ./configure now finds
3274 the release number by grepping pcre.h.
3275
3276 26. Added the ability to run the tests under valgrind.
3277
3278
3279 Version 6.4 05-Sep-05
3280 ---------------------
3281
3282 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3283 "--" to be printed when multiple files were scanned, even when none of the
3284 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3285 consider it to be a bug, and have restored the previous behaviour.
3286
3287 2. A couple of code tidies to get rid of compiler warnings.
3288
3289 3. The pcretest program used to cheat by referring to symbols in the library
3290 whose names begin with _pcre_. These are internal symbols that are not
3291 really supposed to be visible externally, and in some environments it is
3292 possible to suppress them. The cheating is now confined to including
3293 certain files from the library's source, which is a bit cleaner.
3294
3295 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3296 file's purpose clearer.
3297
3298 5. Reorganized pcre_ucp_findchar().
3299
3300
3301 Version 6.3 15-Aug-05
3302 ---------------------
3303
3304 1. The file libpcre.pc.in did not have general read permission in the tarball.
3305
3306 2. There were some problems when building without C++ support:
3307
3308 (a) If C++ support was not built, "make install" and "make test" still
3309 tried to test it.
3310
3311 (b) There were problems when the value of CXX was explicitly set. Some
3312 changes have been made to try to fix these, and ...
3313
3314 (c) --disable-cpp can now be used to explicitly disable C++ support.
3315
3316 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3317 backslash in a target when C++ was disabled. This confuses some
3318 versions of "make", apparently. Using an intermediate variable solves
3319 this. (Same for CPP_LOBJ.)
3320
3321 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3322 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3323 necessary on certain architectures.
3324
3325 4. Added a setting of -export-symbols-regex to the link command to remove
3326 those symbols that are exported in the C sense, but actually are local
3327 within the library, and not documented. Their names all begin with
3328 "_pcre_". This is not a perfect job, because (a) we have to except some
3329 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3330 available (and never for static libraries). I have made a note to try to
3331 find a way round (a) in the future.
3332
3333
3334 Version 6.2 01-Aug-05
3335 ---------------------
3336
3337 1. There was no test for integer overflow of quantifier values. A construction
3338 such as {1111111111111111} would give undefined results. What is worse, if
3339 a minimum quantifier for a parenthesized subpattern overflowed and became
3340 negative, the calculation of the memory size went wrong. This could have
3341 led to memory overwriting.
3342
3343 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3344
3345 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3346 operating environments where this matters.
3347
3348 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3349 PCRE options from within the C++ wrapper.
3350
3351 5. Named capturing subpatterns were not being correctly counted when a pattern
3352 was compiled. This caused two problems: (a) If there were more than 100
3353 such subpatterns, the calculation of the memory needed for the whole
3354 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3355 back references of the form \12, where the number was greater than 9, were
3356 not recognized as back references, even though there were sufficient
3357 previous subpatterns.
3358
3359 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3360 versions of gcc, e.g. 2.95.4.
3361
3362
3363 Version 6.1 21-Jun-05
3364 ---------------------
3365
3366 1. There was one reference to the variable "posix" in pcretest.c that was not
3367 surrounded by "#if !defined NOPOSIX".
3368
3369 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3370 the cross-check on the old pcre_info() function, for the benefit of the
3371 cut-down version of PCRE that is currently imported into Exim.
3372
3373 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3374 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3375 patterns that start (?i) but I don't think that matters. The use of (?i) is
3376 just an example; this all applies to the other options as well.
3377
3378 4. Since libtool seems to echo the compile commands it is issuing, the output
3379 from "make" can be reduced a bit by putting "@" in front of each libtool
3380 compile command.
3381
3382 5. Patch from the folks at Google for configure.in to be a bit more thorough
3383 in checking for a suitable C++ installation before trying to compile the
3384 C++ stuff. This should fix a reported problem when a compiler was present,
3385 but no suitable headers.
3386
3387 6. The man pages all had just "PCRE" as their title. I have changed them to
3388 be the relevant file name. I have also arranged that these names are
3389 retained in the file doc/pcre.txt, which is a concatenation in text format
3390 of all the man pages except the little individual ones for each function.
3391
3392 7. The NON-UNIX-USE file had not been updated for the different set of source
3393 files that come with release 6. I also added a few comments about the C++
3394 wrapper.
3395
3396
3397 Version 6.0 07-Jun-05
3398 ---------------------
3399
3400 1. Some minor internal re-organization to help with my DFA experiments.
3401
3402 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3403 didn't matter for the library itself when fully configured, but did matter
3404 when compiling without UCP support, or within Exim, where the ucp files are
3405 not imported.
3406
3407 3. Refactoring of the library code to split up the various functions into
3408 different source modules. The addition of the new DFA matching code (see
3409 below) to a single monolithic source would have made it really too
3410 unwieldy, quite apart from causing all the code to be include in a
3411 statically linked application, when only some functions are used. This is
3412 relevant even without the DFA addition now that patterns can be compiled in
3413 one application and matched in another.
3414
3415 The downside of splitting up is that there have to be some external
3416 functions and data tables that are used internally in different modules of
3417 the library but which are not part of the API. These have all had their
3418 names changed to start with "_pcre_" so that they are unlikely to clash
3419 with other external names.
3420
3421 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3422 a different (DFA) algorithm. Although it is slower than the original
3423 function, it does have some advantages for certain types of matching
3424 problem.
3425
3426 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3427 including restarting after a partial match.
3428
3429 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3430 defined when compiling for Windows was sent to me. I have put it into the
3431 code, though I have no means of testing or verifying it.
3432
3433 7. Added the pcre_refcount() auxiliary function.
3434
3435 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3436 match before or at the first newline in the subject string. In pcretest,
3437 the /f option on a pattern can be used to set this.
3438
3439 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3440 would behave wrongly. This has been present in PCRE since release 4.0.
3441
3442 10. A number of changes to the pcregrep command:
3443
3444 (a) Refactored how -x works; insert ^(...)$ instead of setting
3445 PCRE_ANCHORED and checking the length, in preparation for adding
3446 something similar for -w.
3447
3448 (b) Added the -w (match as a word) option.
3449
3450 (c) Refactored the way lines are read and buffered so as to have more
3451 than one at a time available.
3452
3453 (d) Implemented a pcregrep test script.
3454
3455 (e) Added the -M (multiline match) option. This allows patterns to match
3456 over several lines of the subject. The buffering ensures that at least
3457 8K, or the rest of the document (whichever is the shorter) is available
3458 for matching (and similarly the previous 8K for lookbehind assertions).
3459
3460 (f) Changed the --help output so that it now says
3461
3462 -w, --word-regex(p)
3463
3464 instead of two lines, one with "regex" and the other with "regexp"
3465 because that confused at least one person since the short forms are the
3466 same. (This required a bit of code, as the output is generated
3467 automatically from a table. It wasn't just a text change.)
3468
3469 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3470 option but starts with a hyphen. Could be a pattern or a path name
3471 starting with a hyphen, for instance.
3472
3473 (h) "-" can be given as a file name to represent stdin.
3474
3475 (i) When file names are being printed, "(standard input)" is used for
3476 the standard input, for compatibility with GNU grep. Previously
3477 "<stdin>" was used.
3478
3479 (j) The option --label=xxx can be used to supply a name to be used for
3480 stdin when file names are being printed. There is no short form.
3481
3482 (k) Re-factored the options decoding logic because we are going to add
3483 two more options that take data. Such options can now be given in four
3484 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3485
3486 (l) Added the -A, -B, and -C options for requesting that lines of context
3487 around matches be printed.
3488
3489 (m) Added the -L option to print the names of files that do not contain
3490 any matching lines, that is, the complement of -l.
3491
3492 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3493 continue to scan other files.
3494
3495 (o) The -s option was incorrectly implemented. For compatibility with other
3496 greps, it now suppresses the error message for a non-existent or non-
3497 accessible file (but not the return code). There is a new option called
3498 -q that suppresses the output of matching lines, which was what -s was
3499 previously doing.
3500
3501 (p) Added --include and --exclude options to specify files for inclusion
3502 and exclusion when recursing.
3503
3504 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3505 Hopefully, it now does.
3506
3507 12. Missing cast in pcre_study().
3508
3509 13. Added an "uninstall" target to the makefile.
3510
3511 14. Replaced "extern" in the function prototypes in Makefile.in with
3512 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3513 world, but is set differently for Windows.
3514
3515 15. Added a second compiling function called pcre_compile2(). The only
3516 difference is that it has an extra argument, which is a pointer to an
3517 integer error code. When there is a compile-time failure, this is set
3518 non-zero, in addition to the error test pointer being set to point to an
3519 error message. The new argument may be NULL if no error number is required
3520 (but then you may as well call pcre_compile(), which is now just a
3521 wrapper). This facility is provided because some applications need a
3522 numeric error indication, but it has also enabled me to tidy up the way
3523 compile-time errors are handled in the POSIX wrapper.
3524
3525 16. Added VPATH=.libs to the makefile; this should help when building with one
3526 prefix path and installing with another. (Or so I'm told by someone who
3527 knows more about this stuff than I do.)
3528
3529 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3530 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3531 match everything, including newlines. This is not POSIX-compatible, but
3532 somebody wanted the feature. From pcretest it can be activated by using
3533 both the P and the s flags.
3534
3535 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3536
3537 19. libpcre.pc was being incorrectly installed as executable.
3538
3539 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3540 it now also looks for '\r' so that it will work unmodified on Windows.
3541
3542 21. Added Google's contributed C++ wrapper to the distribution.
3543
3544 22. Added some untidy missing memory free() calls in pcretest, to keep
3545 Electric Fence happy when testing.
3546
3547
3548
3549 Version 5.0 13-Sep-04
3550 ---------------------
3551
3552 1. Internal change: literal characters are no longer packed up into items
3553 containing multiple characters in a single byte-string. Each character
3554 is now matched using a separate opcode. However, there may be more than one
3555 byte in the character in UTF-8 mode.
3556
3557 2. The pcre_callout_block structure has two new fields: pattern_position and
3558 next_item_length. These contain the offset in the pattern to the next match
3559 item, and its length, respectively.
3560
3561 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3562 insertion of callouts before each pattern item. Added the /C option to
3563 pcretest to make use of this.
3564
3565 4. On the advice of a Windows user, the lines
3566
3567 #if defined(_WIN32) || defined(WIN32)
3568 _setmode( _fileno( stdout ), 0x8000 );
3569 #endif /* defined(_WIN32) || defined(WIN32) */
3570
3571 have been added to the source of pcretest. This apparently does useful
3572 magic in relation to line terminators.
3573
3574 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3575 for the benefit of those environments where the "b" makes a difference.
3576
3577 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3578 to know about it. I have put a hack into configure.in that adds in code
3579 to set GCC=yes if CC=icc. This seems to end up at a point in the
3580 generated configure script that is early enough to affect the setting of
3581 compiler options, which is what is needed, but I have no means of testing
3582 whether it really works. (The user who reported this had patched the
3583 generated configure script, which of course I cannot do.)
3584
3585 LATER: After change 22 below (new libtool files), the configure script
3586 seems to know about icc (and also ecc). Therefore, I have commented out
3587 this hack in configure.in.
3588
3589 7. Added support for pkg-config (2 patches were sent in).
3590
3591 8. Negated POSIX character classes that used a combination of internal tables
3592 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3593 [[:^ascii]]. Typically, they would match almost any characters. The other
3594 POSIX classes were not broken in this way.
3595
3596 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3597 to find the match, as PCRE was deluded into thinking that the match had to
3598 start at the start point or following a newline. The same bug applied to
3599 patterns with negative forward assertions or any backward assertions
3600 preceding ".*" at the start, unless the pattern required a fixed first
3601 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3602
3603 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3604 starting at the last subject character, bytes beyond the end of the subject
3605 string were read.
3606
3607 11. Renamed the variable "class" as "classbits" to make life easier for C++
3608 users. (Previously there was a macro definition, but it apparently wasn't
3609 enough.)
3610
3611 12. Added the new field "tables" to the extra data so that tables can be passed
3612 in at exec time, or the internal tables can be re-selected. This allows
3613 a compiled regex to be saved and re-used at a later time by a different
3614 program that might have everything at different addresses.
3615
3616 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3617 -R library as well as a -L library.
3618
3619 14. The debugging options of pcretest (-d on the command line or D on a
3620 pattern) showed incorrect output for anything following an extended class
3621 that contained multibyte characters and which was followed by a quantifier.
3622
3623 15. Added optional support for general category Unicode character properties
3624 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3625 support. It adds about 90K to the size of the library. The meanings of the
3626 inbuilt class escapes such as \d and \s have NOT been changed.
3627
3628 16. Updated pcredemo.c to include calls to free() to release the memory for the
3629 compiled pattern.
3630
3631 17. The generated file chartables.c was being created in the source directory
3632 instead of in the building directory. This caused the build to fail if the
3633 source directory was different from the building directory, and was
3634 read-only.
3635
3636 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3637 file. No doubt somebody will tell me if they don't make sense... Also added
3638 Dan Mooney's comments about building on OpenVMS.
3639
3640 19. Added support for partial matching via the PCRE_PARTIAL option for
3641 pcre_exec() and the \P data escape in pcretest.
3642
3643 20. Extended pcretest with 3 new pattern features:
3644
3645 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3646 write the compiled pattern to the file whose name is "rest-of-line".
3647 This is a straight binary dump of the data, with the saved pointer to
3648 the character tables forced to be NULL. The study data, if any, is
3649 written too. After writing, pcretest reads a new pattern.
3650
3651 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3652 compiled pattern from the given file. There must not be any
3653 occurrences of "<" in the file name (pretty unlikely); if there are,
3654 pcretest will instead treat the initial "<" as a pattern delimiter.
3655 After reading in the pattern, pcretest goes on to read data lines as
3656 usual.
3657
3658 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3659 and 16-bit fields in a compiled pattern, to simulate a pattern that
3660 was compiled on a host of opposite endianness.
3661
3662 21. The pcre-exec() function can now cope with patterns that were compiled on
3663 hosts of opposite endianness, with this restriction:
3664
3665 As for any compiled expression that is saved and used later, the tables
3666 pointer field cannot be preserved; the extra_data field in the arguments
3667 to pcre_exec() should be used to pass in a tables address if a value
3668 other than the default internal tables were used at compile time.
3669
3670 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3671 now diagnosed as an error. Previously, most of the time, a negative number
3672 would have been treated as zero, but if in addition "ovector" was passed as
3673 NULL, a crash could occur.
3674
3675 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3676 new versions from the libtool 1.5 distribution (the last one is a copy of
3677 a file called libtool.m4). This seems to have fixed the need to patch
3678 "configure" to support Darwin 1.3 (which I used to do). However, I still
3679 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3680 workstation).
3681
3682 24. Changed the PCRE licence to be the more standard "BSD" licence.
3683
3684
3685 Version 4.5 01-Dec-03
3686 ---------------------
3687
3688 1. There has been some re-arrangement of the code for the match() function so
3689 that it can be compiled in a version that does not call itself recursively.
3690 Instead, it keeps those local variables that need separate instances for
3691 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3692 needs to "recurse". Keeping track of where control must go is done by means
3693 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3694 hide most of the details from the main code, and operates only if
3695 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3696 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3697 operating.
3698
3699 To make it easier for callers to provide specially tailored get/free
3700 functions for this usage, two new functions, pcre_stack_malloc, and
3701 pcre_stack_free, are used. They are always called in strict stacking order,
3702 and the size of block requested is always the same.
3703
3704 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3705 PCRE has been compiled to use the stack or the heap for recursion. The
3706 -C option of pcretest uses this to show which version is compiled.
3707
3708 A new data escape \S, is added to pcretest; it causes the amounts of store
3709 obtained and freed by both kinds of malloc/free at match time to be added
3710 to the output.
3711
3712 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3713 what's available on my current Linux desktop machine.
3714
3715 3. When matching a UTF-8 string, the test for a valid string at the start has
3716 been extended. If start_offset is not zero, PCRE now checks that it points
3717 to a byte that is the start of a UTF-8 character. If not, it returns
3718 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3719 this is necessary because there may be backward assertions in the pattern.
3720 When matching the same subject several times, it may save resources to use
3721 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3722
3723 4. The code for checking the validity of UTF-8 strings has been tightened so
3724 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3725 containing "overlong sequences".
3726
3727 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3728 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3729 should have been "&", but it just so happened that all the cases this let
3730 through by mistake were picked up later in the function.
3731
3732 6. I had used a variable called "isblank" - this is a C99 function, causing
3733 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3734
3735 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3736 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3737 so the version is shown; (c) stop "make test" echoing "./RunTest".
3738
3739 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3740
3741 9. The prototype for memmove() for systems that don't have it was using
3742 size_t, but the inclusion of the header that defines size_t was later. I've
3743 moved the #includes for the C headers earlier to avoid this.
3744
3745 10. Added some adjustments to the code to make it easier to compiler on certain
3746 special systems:
3747
3748 (a) Some "const" qualifiers were missing.
3749 (b) Added the macro EXPORT before all exported functions; by default this
3750 is defined to be empty.
3751 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3752 that it reads its output file name as an argument instead of writing
3753 to the standard output and assuming this can be redirected.
3754
3755 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3756 class containing characters with values greater than 255, PCRE compilation
3757 went into a loop.
3758
3759 12. A recursive reference to a subpattern that was within another subpattern
3760 that had a minimum quantifier of zero caused PCRE to crash. For example,
3761 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3762 recursion. If the recursively-called subpattern itself had a zero repeat,
3763 that was OK.
3764
3765 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3766 buffer into which it was copied (for escape processing) was still set at
3767 1024, so long lines caused crashes.
3768
3769 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3770 "internal error: code overflow...". This applied to any character class
3771 that was followed by a possessive quantifier.
3772
3773 15. Modified the Makefile to add libpcre.la as a prerequisite for
3774 libpcreposix.la because I was told this is needed for a parallel build to
3775 work.
3776
3777 16. If a pattern that contained .* following optional items at the start was
3778 studied, the wrong optimizing data was generated, leading to matching
3779 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3780 matching string must start with a or b or c. The correct conclusion for
3781 this pattern is that a match can start with any character.
3782
3783
3784 Version 4.4 13-Aug-03
3785 ---------------------
3786
3787 1. In UTF-8 mode, a character class containing characters with values between
3788 127 and 255 was not handled correctly if the compiled pattern was studied.
3789 In fixing this, I have also improved the studying algorithm for such
3790 classes (slightly).
3791
3792 2. Three internal functions had redundant arguments passed to them. Removal
3793 might give a very teeny performance improvement.
3794
3795 3. Documentation bug: the value of the capture_top field in a callout is *one
3796 more than* the number of the hightest numbered captured substring.
3797
3798 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3799 in incorrectly linking with a previously installed version. They now link
3800 explicitly with libpcre.la.
3801
3802 5. configure.in no longer needs to recognize Cygwin specially.
3803
3804 6. A problem in pcre.in for Windows platforms is fixed.
3805
3806 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3807 pcretest, it used to include the size of the study block as part of its
3808 output. Unfortunately, the structure contains a field that has a different
3809 size on different hardware architectures. This meant that the tests that
3810 showed this size failed. As the block is currently always of a fixed size,
3811 this information isn't actually particularly useful in pcretest output, so
3812 I have just removed it.
3813
3814 8. Three pre-processor statements accidentally did not start in column 1.
3815 Sadly, there are *still* compilers around that complain, even though
3816 standard C has not required this for well over a decade. Sigh.
3817
3818 9. In pcretest, the code for checking callouts passed small integers in the
3819 callout_data field, which is a void * field. However, some picky compilers
3820 complained about the casts involved for this on 64-bit systems. Now
3821 pcretest passes the address of the small integer instead, which should get
3822 rid of the warnings.
3823
3824 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3825 both compile and run time, and gives an error if an invalid UTF-8 sequence
3826 is found. There is a option for disabling this check in cases where the
3827 string is known to be correct and/or the maximum performance is wanted.
3828
3829 11. In response to a bug report, I changed one line in Makefile.in from
3830
3831 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3832 to
3833 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3834
3835 to look similar to other lines, but I have no way of telling whether this
3836 is the right thing to do, as I do not use Windows. No doubt I'll get told
3837 if it's wrong...
3838
3839
3840 Version 4.3 21-May-03
3841 ---------------------
3842
3843 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3844 Makefile.
3845
3846 2. Some refactoring to improve the quality of the code:
3847
3848 (i) The utf8_table... variables are now declared "const".
3849
3850 (ii) The code for \cx, which used the "case flipping" table to upper case
3851 lower case letters, now just substracts 32. This is ASCII-specific,
3852 but the whole concept of \cx is ASCII-specific, so it seems
3853 reasonable.
3854
3855 (iii) PCRE was using its character types table to recognize decimal and
3856 hexadecimal digits in the pattern. This is silly, because it handles
3857 only 0-9, a-f, and A-F, but the character types table is locale-
3858 specific, which means strange things might happen. A private
3859 table is now used for this - though it costs 256 bytes, a table is
3860 much faster than multiple explicit tests. Of course, the standard
3861 character types table is still used for matching digits in subject
3862 strings against \d.
3863
3864 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3865 ending in _t are). So I've renamed it as ESC_tee.
3866
3867 3. The first argument for regexec() in the POSIX wrapper should have been
3868 defined as "const".
3869
3870 4. Changed pcretest to use malloc() for its buffers so that they can be
3871 Electric Fenced for debugging.
3872
3873 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3874 to read one or more bytes before the start of the subject string. Often this
3875 had no effect on PCRE's behaviour, but in some circumstances it could
3876 provoke a segmentation fault.
3877
3878 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3879 to try to read one or more bytes before the start of the subject string.
3880
3881 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3882 UTF-8 support could misbehave in various ways if the subject string
3883 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3884 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3885 back over UTF-8 characters.)
3886
3887
3888 Version 4.2 14-Apr-03
3889 ---------------------
3890
3891 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3892
3893 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3894 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3895 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3896 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3897 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3898 and BUILD_EXEEXT
3899 Note: automatic setting of the BUILD variables is not yet working
3900 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3901 compile-time but not at link-time
3902 [LINK]: use for linking executables only
3903 make different versions for Windows and non-Windows
3904 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3905 libraries
3906 [LINK_FOR_BUILD]: new variable
3907 [OBJEXT]: use throughout
3908 [EXEEXT]: use throughout
3909 <winshared>: new target
3910 <wininstall>: new target
3911 <dftables.o>: use native compiler
3912 <dftables>: use native linker
3913 <install>: handle Windows platform correctly
3914 <clean>: ditto
3915 <check>: ditto
3916 copy DLL to top builddir before testing
3917
3918 As part of these changes, -no-undefined was removed again. This was reported
3919 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3920 in any case.
3921
3922 3. Some tidies to get rid of compiler warnings:
3923
3924 . In the match_data structure, match_limit was an unsigned long int, whereas
3925 match_call_count was an int. I've made them both unsigned long ints.
3926
3927 . In pcretest the fact that a const uschar * doesn't automatically cast to
3928 a void * provoked a warning.
3929
3930 . Turning on some more compiler warnings threw up some "shadow" variables
3931 and a few more missing casts.
3932
3933 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3934 option, a class that contained a single character with a value between 128
3935 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3936
3937 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3938 option, a class that contained several characters, but with at least one
3939 whose value was between 128 and 255 caused PCRE to crash.
3940
3941
3942 Version 4.1 12-Mar-03
3943 ---------------------
3944
3945 1. Compiling with gcc -pedantic found a couple of places where casts were
3946 needed, and a string in dftables.c that was longer than standard compilers are
3947 required to support.
3948
3949 2. Compiling with Sun's compiler found a few more places where the code could
3950 be tidied up in order to avoid warnings.
3951
3952 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3953 first of these names is deprecated in the latest Autoconf in favour of the name
3954 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3955 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3956 analogy I have changed it to CFLAGS_FOR_BUILD.
3957
3958 4. Added -no-undefined to the linking command in the Makefile, because this is
3959 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3960 linking step for the pcreposix library.
3961
3962 5. PCRE was failing to diagnose the case of two named groups with the same
3963 name.
3964
3965 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3966 literal character that is needed in the subject for a match, and scans along to
3967 ensure that it is present before embarking on the full matching process. This
3968 saves time in cases of nested unlimited repeats that are never going to match.
3969 Problem: the scan can take a lot of time if the subject is very long (e.g.
3970 megabytes), thus penalizing straightforward matches. It is now done only if the
3971 amount of subject to be scanned is less than 1000 bytes.
3972
3973 7. A lesser problem with the same optimization is that it was recording the
3974 first character of an anchored pattern as "needed", thus provoking a search
3975 right along the subject, even when the first match of the pattern was going to
3976 fail. The "needed" character is now not set for anchored patterns, unless it
3977 follows something in the pattern that is of non-fixed length. Thus, it still
3978 fulfils its original purpose of finding quick non-matches in cases of nested
3979 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3980
3981
3982 Version 4.0 17-Feb-03
3983 ---------------------
3984
3985 1. If a comment in an extended regex that started immediately after a meta-item
3986 extended to the end of string, PCRE compiled incorrect data. This could lead to
3987 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3988
3989 2. Moved to autoconf 2.53 and libtool 1.4.2.
3990
3991 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3992 the special perltest8 script is no longer needed - all the tests can be run
3993 from a single perltest script.
3994
3995 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3996 by \s. It has now been removed in PCRE. This means it isn't recognized as
3997 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3998 class [:space:] *does* include VT, thereby creating a mess.
3999
4000 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4001 space and tab.
4002
4003 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4004 its new features into the main test script, reducing the number of scripts.
4005
4006 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4007 were backward compatible, and made the (?i) apply to the whole pattern, as if
4008 /i were given. Now it behaves more logically, and applies the option setting
4009 only to what follows. PCRE has been changed to follow suit. However, if it
4010 finds options settings right at the start of the pattern, it extracts them into
4011 the global options, as before. Thus, they show up in the info data.
4012
4013 8. Added support for the \Q...\E escape sequence. Characters in between are
4014 treated as literals. This is slightly different from Perl in that $ and @ are
4015 also handled as literals inside the quotes. In Perl, they will cause variable
4016 interpolation. Note the following examples:
4017
4018 Pattern PCRE matches Perl matches
4019
4020 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4021 \Qabc\$xyz\E abc\$xyz abc\$xyz
4022 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4023
4024 For compatibility with Perl, \Q...\E sequences are recognized inside character
4025 classes as well as outside them.
4026
4027 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4028 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4029 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4030 signed/unsigned warnings.
4031
4032 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4033 option for pcretest, so I've replaced it by a simple function that does just
4034 that job.
4035
4036 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4037 "pcregrep -".
4038
4039 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4040 Java package. This provides some syntactic sugar for simple cases of what my
4041 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4042 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4043 item, you can use this simplified notation. Note that only makes sense with
4044 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4045 greediness, whatever the setting of the PCRE_UNGREEDY option.
4046
4047 13. A change of greediness default within a pattern was not taking effect at
4048 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4049 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4050 was abstracted outside.
4051
4052 14. PCRE now supports the \G assertion. It is true when the current matching
4053 position is at the start point of the match. This differs from \A when the
4054 starting offset is non-zero. Used with the /g option of pcretest (or similar
4055 code), it works in the same way as it does for Perl's /g option. If all
4056 alternatives of a regex begin with \G, the expression is anchored to the start
4057 match position, and the "anchored" flag is set in the compiled expression.
4058
4059 15. Some bugs concerning the handling of certain option changes within patterns
4060 have been fixed. These applied to options other than (?ims). For example,
4061 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4062 been the other way round. Some of this was related to change 7 above.
4063
4064 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4065 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4066 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4067 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4068
4069 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4070 mode. Unlike ".", it always matches newline, whatever the setting of
4071 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4072 assertions. Perl allows it, but it doesn't (in general) work because it can't
4073 calculate the length of the lookbehind. At least, that's the case for Perl
4074 5.8.0 - I've been told they are going to document that it doesn't work in
4075 future.
4076
4077 18. Added an error diagnosis for escapes that PCRE does not support: these are
4078 \L, \l, \N, \P, \p, \U, \u, and \X.
4079
4080 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4081 reading past the end of the pattern in cases such as /[abcd/.
4082
4083 20. PCRE was getting more memory than necessary for patterns with classes that
4084 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4085
4086 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4087 compiling PCRE for use with Virtual Pascal.
4088
4089 22. Small fix to the Makefile to make it work properly if the build is done
4090 outside the source tree.
4091
4092 23. Added a new extension: a condition to go with recursion. If a conditional
4093 subpattern starts with (?(R) the "true" branch is used if recursion has
4094 happened, whereas the "false" branch is used only at the top level.
4095
4096 24. When there was a very long string of literal characters (over 255 bytes
4097 without UTF support, over 250 bytes with UTF support), the computation of how
4098 much memory was required could be incorrect, leading to segfaults or other
4099 strange effects.
4100
4101 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4102 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4103 there was a subsequent back reference to those brackets. This meant that, for
4104 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4105 possible to check for precisely this case. All we can do is abandon the
4106 optimization if .* occurs inside capturing brackets when there are any back
4107 references whatsoever. (See below for a better fix that came later.)
4108
4109 26. The handling of the optimization for finding the first character of a
4110 non-anchored pattern, and for finding a character that is required later in the
4111 match were failing in some cases. This didn't break the matching; it just
4112 failed to optimize when it could. The way this is done has been re-implemented.
4113
4114 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4115
4116 28. Added a new feature that provides some of the functionality that Perl
4117 provides with (?{...}). The facility is termed a "callout". The way it is done
4118 in PCRE is for the caller to provide an optional function, by setting
4119 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4120 global variable. By default it is unset, which disables all calling out. To get
4121 the function called, the regex must include (?C) at appropriate points. This
4122 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4123 This provides a means of identifying different callout points. When PCRE
4124 reaches such a point in the regex, if pcre_callout has been set, the external
4125 function is called. It is provided with data in a structure called
4126 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4127 matching continues; if it returns a non-zero value, the match at the current
4128 point fails. However, backtracking will occur if possible. [This was changed
4129 later and other features added - see item 49 below.]
4130
4131 29. pcretest is upgraded to test the callout functionality. It provides a
4132 callout function that displays information. By default, it shows the start of
4133 the match and the current position in the text. There are some new data escapes
4134 to vary what happens:
4135
4136 \C+ in addition, show current contents of captured substrings
4137 \C- do not supply a callout function
4138 \C!n return 1 when callout number n is reached
4139 \C!n!m return 1 when callout number n is reached for the mth time
4140
4141 30. If pcregrep was called with the -l option and just a single file name, it
4142 output "<stdin>" if a match was found, instead of the file name.
4143
4144 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4145 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4146 pcre_exec(). This saves a malloc/free per call. The default value of
4147 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4148 when configuring.
4149
4150 32. The default maximum size of a compiled pattern is 64K. There have been a
4151 few cases of people hitting this limit. The code now uses macros to handle the
4152 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4153 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4154 configuring. Tests 2 and 5 work only with 2-byte links because they output
4155 debugging information about compiled patterns.
4156
4157 33. Internal code re-arrangements:
4158
4159 (a) Moved the debugging function for printing out a compiled regex into
4160 its own source file (printint.c) and used #include to pull it into
4161 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4162 separate copies.
4163
4164 (b) Defined the list of op-code names for debugging as a macro in
4165 internal.h so that it is next to the definition of the opcodes.
4166
4167 (c) Defined a table of op-code lengths for simpler skipping along compiled
4168 code. This is again a macro in internal.h so that it is next to the
4169 definition of the opcodes.
4170
4171 34. Added support for recursive calls to individual subpatterns, along the
4172 lines of Robin Houston's patch (but implemented somewhat differently).
4173
4174 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4175 allow it to read and process whole directories in Win32. This code was
4176 contributed by Lionel Fourquaux; it has not been tested by me.
4177
4178 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4179 used to name a group. Names consist of alphanumerics and underscores, and must
4180 be unique. Back references use the syntax (?P=name) and recursive calls use
4181 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4182 numbers. The function pcre_fullinfo() can be used after compilation to extract
4183 a name/number map. There are three relevant calls:
4184
4185 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4186 PCRE_INFO_NAMECOUNT yields the number of entries
4187 PCRE_INFO_NAMETABLE yields a pointer to the map.
4188
4189 The map is a vector of fixed-size entries. The size of each entry depends on
4190 the length of the longest name used. The first two bytes of each entry are the
4191 group number, most significant byte first. There follows the corresponding
4192 name, zero terminated. The names are in alphabetical order.
4193
4194 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4195 case instead of 255. Making it the same both with and without UTF-8 support
4196 means that the same test output works with both.
4197
4198 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4199 calling malloc() with a zero argument.
4200
4201 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4202 optimization. I've improved things by keeping a bitmap of backreferences with
4203 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4204 fact referenced, the optimization can be applied. It is unlikely that a
4205 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4206 the match to follow \n) will appear inside brackets with a number greater than
4207 31, but if it does, any back reference > 31 suppresses the optimization.
4208
4209 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4210 of disabling numbered capturing parentheses. Any opening parenthesis that is
4211 not followed by ? behaves as if it were followed by ?: but named parentheses
4212 can still be used for capturing (and they will acquire numbers in the usual
4213 way).
4214
4215 41. Redesigned the return codes from the match() function into yes/no/error so
4216 that errors can be passed back from deep inside the nested calls. A malloc
4217 failure while inside a recursive subpattern call now causes the
4218 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4219
4220 42. It is now possible to set a limit on the number of times the match()
4221 function is called in a call to pcre_exec(). This facility makes it possible to
4222 limit the amount of recursion and backtracking, though not in a directly
4223 obvious way, because the match() function is used in a number of different
4224 circumstances. The count starts from zero for each position in the subject
4225 string (for non-anchored patterns). The default limit is, for compatibility, a
4226 large number, namely 10 000 000. You can change this in two ways:
4227
4228 (a) When configuring PCRE before making, you can use --with-match-limit=n
4229 to set a default value for the compiled library.
4230
4231 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4232 a different value is set. See 45 below.
4233
4234 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4235
4236 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4237 of things that can be changed at compile time. The first argument specifies
4238 what is wanted and the second points to where the information is to be placed.
4239 The current list of available information is:
4240
4241 PCRE_CONFIG_UTF8
4242
4243 The output is an integer that is set to one if UTF-8 support is available;
4244 otherwise it is set to zero.
4245
4246 PCRE_CONFIG_NEWLINE
4247
4248 The output is an integer that it set to the value of the code that is used for
4249 newline. It is either LF (10) or CR (13).
4250
4251 PCRE_CONFIG_LINK_SIZE
4252
4253 The output is an integer that contains the number of bytes used for internal
4254 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4255
4256 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4257
4258 The output is an integer that contains the threshold above which the POSIX
4259 interface uses malloc() for output vectors. See item 31 above.
4260
4261 PCRE_CONFIG_MATCH_LIMIT
4262
4263 The output is an unsigned integer that contains the default limit of the number
4264 of match() calls in a pcre_exec() execution. See 42 above.
4265
4266 44. pcretest has been upgraded by the addition of the -C option. This causes it
4267 to extract all the available output from the new pcre_config() function, and to
4268 output it. The program then exits immediately.
4269
4270 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4271 order to support additional features. One way would have been to define
4272 pcre_exec2() (for example) with extra arguments, but this would not have been
4273 extensible, and would also have required all calls to the original function to
4274 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4275 is used for passing in "extra" data from pcre_study().
4276
4277 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4278 contains the following fields:
4279
4280 flags a bitmap indicating which of the following fields are set
4281 study_data opaque data from pcre_study()
4282 match_limit a way of specifying a limit on match() calls for a specific
4283 call to pcre_exec()
4284 callout_data data for callouts (see 49 below)
4285
4286 The flag bits are also defined in pcre.h, and are
4287
4288 PCRE_EXTRA_STUDY_DATA
4289 PCRE_EXTRA_MATCH_LIMIT
4290 PCRE_EXTRA_CALLOUT_DATA
4291
4292 The pcre_study() function now returns one of these new pcre_extra blocks, with
4293 the actual study data pointed to by the study_data field, and the
4294 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4295 before. That is, this change is entirely upwards-compatible and requires no
4296 change to existing code.
4297
4298 If you want to pass in additional data to pcre_exec(), you can either place it
4299 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4300 block.
4301
4302 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4303 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4304 times with different match limits, until it finds the minimum value needed for
4305 pcre_exec() to complete. The value is then output. This can be instructive; for
4306 most simple matches the number is quite small, but for pathological cases it
4307 gets very large very quickly.
4308
4309 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4310 returns the size of the data block pointed to by the study_data field in a
4311 pcre_extra block, that is, the value that was passed as the argument to
4312 pcre_malloc() when PCRE was getting memory in which to place the information
4313 created by pcre_study(). The fourth argument should point to a size_t variable.
4314 pcretest has been extended so that this information is shown after a successful
4315 pcre_study() call when information about the compiled regex is being displayed.
4316
4317 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4318 because what follows is always an absolute path. (Later: it turns out that this
4319 is more than cosmetic for MinGW, because it doesn't like empty path
4320 components.)
4321
4322 49. Some changes have been made to the callout feature (see 28 above):
4323
4324 (i) A callout function now has three choices for what it returns:
4325
4326 0 => success, carry on matching
4327 > 0 => failure at this point, but backtrack if possible
4328 < 0 => serious error, return this value from pcre_exec()
4329
4330 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4331 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4332 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4333 use by callout functions. It will never be used by PCRE itself.
4334
4335 (ii) The pcre_extra structure (see 45 above) has a void * field called
4336 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4337 pcre_callout_block structure has a field of the same name. The contents of
4338 the field passed in the pcre_extra structure are passed to the callout
4339 function in the corresponding field in the callout block. This makes it
4340 easier to use the same callout-containing regex from multiple threads. For
4341 testing, the pcretest program has a new data escape
4342
4343 \C*n pass the number n (may be negative) as callout_data
4344
4345 If the callout function in pcretest receives a non-zero value as
4346 callout_data, it returns that value.
4347
4348 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4349 there were some redundant $(CFLAGS) in commands that are now specified as
4350 $(LINK), which already includes $(CFLAGS).
4351
4352 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4353 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4354 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4355 one-byte characters throughout. Note that case-insensitive matching applies
4356 only to characters whose values are less than 256. PCRE doesn't support the
4357 notion of cases for higher-valued characters.
4358
4359 (i) A character class whose characters are all within 0-255 is handled as
4360 a bit map, and the map is inverted for negative classes. Previously, a
4361 character > 255 always failed to match such a class; however it should
4362 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4363
4364 (ii) A negated character class with a single character < 255 is coded as
4365 "not this character" (OP_NOT). This wasn't working properly when the test
4366 character was multibyte, either singly or repeated.
4367
4368 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4369 mode, for example: \x{100}{2,3}.
4370
4371 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4372 singly or repeated) now correctly test multibyte characters. However,
4373 PCRE doesn't recognize any characters with values greater than 255 as
4374 digits, spaces, or word characters. Such characters always match \D, \S,
4375 and \W, and never match \d, \s, or \w.
4376
4377 (v) Classes may now contain characters and character ranges with values
4378 greater than 255. For example: [ab\x{100}-\x{400}].
4379
4380 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4381 PCRE in UTF-8 mode.
4382
4383 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4384 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4385 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4386 value.)
4387
4388 53. The single man page has become too large. I have therefore split it up into
4389 a number of separate man pages. These also give rise to individual HTML pages;
4390 these are now put in a separate directory, and there is an index.html page that
4391 lists them all. Some hyperlinking between the pages has been installed.
4392
4393 54. Added convenience functions for handling named capturing parentheses.
4394
4395 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4396 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4397 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4398 are faulted.
4399
4400 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4401 calling configure. These values are used when compiling the dftables.c program
4402 which is run to generate the source of the default character tables. They
4403 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4404 you will need to set these values.
4405
4406 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4407
4408
4409 Version 3.9 02-Jan-02
4410 ---------------------
4411
4412 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4413
4414 2. If --disable-static was given, the building process failed when trying to
4415 build pcretest and pcregrep. (For some reason it was using libtool to compile
4416 them, which is not right, as they aren't part of the library.)
4417
4418
4419 Version 3.8 18-Dec-01
4420 ---------------------
4421
4422 1. The experimental UTF-8 code was completely screwed up. It was packing the
4423 bytes in the wrong order. How dumb can you get?
4424
4425
4426 Version 3.7 29-Oct-01
4427 ---------------------
4428
4429 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4430 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4431 this didn't happen under Solaris 8, where I normally test things.
4432
4433 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4434 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4435 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4436 this fix an hour or so after the initial 3.7 release.)
4437
4438
4439 Version 3.6 23-Oct-01
4440 ---------------------
4441
4442 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4443 offsets passed as NULL with zero offset count.
4444
4445 2. The config.guess and config.sub files had not been updated when I moved to
4446 the latest autoconf.
4447
4448
4449 Version 3.5 15-Aug-01
4450 ---------------------
4451
4452 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4453 had been forgotten.
4454
4455 2. By using declared but undefined structures, we can avoid using "void"
4456 definitions in pcre.h while keeping the internal definitions of the structures
4457 private.
4458
4459 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4460 user point of view, this means that both static and shared libraries are built
4461 by default, but this can be individually controlled. More of the work of
4462 handling this static/shared cases is now inside libtool instead of PCRE's make
4463 file.
4464
4465 4. The pcretest utility is now installed along with pcregrep because it is
4466 useful for users (to test regexs) and by doing this, it automatically gets
4467 relinked by libtool. The documentation has been turned into a man page, so
4468 there are now .1, .txt, and .html versions in /doc.
4469
4470 5. Upgrades to pcregrep:
4471 (i) Added long-form option names like gnu grep.
4472 (ii) Added --help to list all options with an explanatory phrase.
4473 (iii) Added -r, --recursive to recurse into sub-directories.
4474 (iv) Added -f, --file to read patterns from a file.
4475
4476 6. pcre_exec() was referring to its "code" argument before testing that
4477 argument for NULL (and giving an error if it was NULL).
4478
4479 7. Upgraded Makefile.in to allow for compiling in a different directory from
4480 the source directory.
4481
4482 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4483 options bits, the pointer it was passed was to an int instead of to an unsigned
4484 long int. This mattered only on 64-bit systems.
4485
4486 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4487 generated) instead of pcre.in, which it its source. Also made the same change
4488 in several of the .c files.
4489
4490 10. A new release of gcc defines printf() as a macro, which broke pcretest
4491 because it had an ifdef in the middle of a string argument for printf(). Fixed
4492 by using separate calls to printf().
4493
4494 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4495 script, to force use of CR or LF instead of \n in the source. On non-Unix
4496 systems, the value can be set in config.h.
4497
4498 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4499 absolute limit. Changed the text of the error message to make this clear, and
4500 likewise updated the man page.
4501
4502 13. The limit of 99 on the number of capturing subpatterns has been removed.
4503 The new limit is 65535, which I hope will not be a "real" limit.
4504
4505
4506 Version 3.4 22-Aug-00
4507 ---------------------
4508
4509 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4510
4511 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4512
4513
4514 Version 3.3 01-Aug-00
4515 ---------------------
4516
4517 1. If an octal character was given, but the value was greater than \377, it
4518 was not getting masked to the least significant bits, as documented. This could
4519 lead to crashes in some systems.
4520
4521 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4522 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4523
4524 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4525 These just pass their arguments on to (pcre_free)(), but they are provided
4526 because some uses of PCRE bind it to non-C systems that can call its functions,
4527 but cannot call free() or pcre_free() directly.
4528
4529 4. Add "make test" as a synonym for "make check". Corrected some comments in
4530 the Makefile.
4531
4532 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4533 Makefile.
4534
4535 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4536 command called pgrep for grepping around the active processes.
4537
4538 7. Added the beginnings of support for UTF-8 character strings.
4539
4540 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4541 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4542 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4543 out for the ar command.)
4544
4545
4546 Version 3.2 12-May-00
4547 ---------------------
4548
4549 This is purely a bug fixing release.
4550
4551 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4552 of ZA. This was just one example of several cases that could provoke this bug,
4553 which was introduced by change 9 of version 2.00. The code for breaking
4554 infinite loops after an iteration that matches an empty string was't working
4555 correctly.
4556
4557 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4558 when matched against abbab (for example). After matching an empty string, it
4559 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4560 caused it to match further down the string than it should.
4561
4562 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4563 was there because it doesn't seem to be needed, and it causes trouble on some
4564 systems, as it is not a Standard C header. It has been removed.
4565
4566 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4567 were reported on the Macintosh. The changes were from
4568
4569 while ((c = *(++ptr)) != 0 && c != '\n');
4570 to
4571 while ((c = *(++ptr)) != 0 && c != '\n') ;
4572
4573 Totally extraordinary, but if that's what it takes...
4574
4575 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4576 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4577 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4578 assumes the way PCRE uses memmove() (always moving upwards).
4579
4580 6. PCRE is being used in one environment where strchr() is not available. There
4581 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4582 faster code anyway.
4583
4584
4585 Version 3.1 09-Feb-00
4586 ---------------------
4587
4588 The only change in this release is the fixing of some bugs in Makefile.in for
4589 the "install" target:
4590
4591 (1) It was failing to install pcreposix.h.
4592
4593 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4594
4595
4596 Version 3.0 01-Feb-00
4597 ---------------------
4598
4599 1. Add support for the /+ modifier to perltest (to output $` like it does in
4600 pcretest).
4601
4602 2. Add support for the /g modifier to perltest.
4603
4604 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4605 matches null strings.
4606
4607 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4608 pattern. Perl treats empty patterns specially - it reuses the most recent
4609 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4610 effect.
4611
4612 5. The POSIX interface was broken in that it was just handing over the POSIX
4613 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4614 required a bigger vector, with some working space on the end. This means that
4615 the POSIX wrapper now has to get and free some memory, and copy the results.
4616
4617 6. Added some simple autoconf support, placing the test data and the
4618 documentation in separate directories, re-organizing some of the
4619 information files, and making it build pcre-config (a GNU standard). Also added
4620 libtool support for building PCRE as a shared library, which is now the
4621 default.
4622
4623 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4624 09 are not valid octal constants. Single digits will be used for minor values
4625 less than 10.
4626
4627 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4628 existing programs that set these in the POSIX interface can use PCRE without
4629 modification.
4630
4631 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4632 return all that pcre_info() returns, plus additional data. The pcre_info()
4633 function is retained for compatibility, but is considered to be obsolete.
4634
4635 10. Added experimental recursion feature (?R) to handle one common case that
4636 Perl 5.6 will be able to do with (?p{...}).
4637
4638 11. Added support for POSIX character classes like [:alpha:], which Perl is
4639 adopting.
4640
4641
4642 Version 2.08 31-Aug-99
4643 ----------------------
4644
4645 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4646 trying to match at the startoffset position, but instead was moving forward to
4647 the next newline as if a previous match had failed.
4648
4649 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4650 and could get into a loop if a null string was matched other than at the start
4651 of the subject.
4652
4653 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4654 be distinguished at compile time, and for completeness also added PCRE_DATE.
4655
4656 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4657 in GnuWin32 environments.
4658
4659
4660 Version 2.07 29-Jul-99
4661 ----------------------
4662
4663 1. The documentation is now supplied in plain text form and HTML as well as in
4664 the form of man page sources.
4665
4666 2. C++ compilers don't like assigning (void *) values to other pointer types.
4667 In particular this affects malloc(). Although there is no problem in Standard
4668 C, I've put in casts to keep C++ compilers happy.
4669
4670 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4671 should be (const char *).
4672
4673 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4674 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4675 However, I haven't made this a standard facility. The documentation doesn't
4676 mention it, and the Makefile doesn't support it.
4677
4678 5. The Makefile now contains an "install" target, with editable destinations at
4679 the top of the file. The pcretest program is not installed.
4680
4681 6. pgrep -V now gives the PCRE version number and date.
4682
4683 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4684 causing the entire string to be ignored, instead of just the last character.
4685
4686 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4687 non-matching string, it can take a very, very long time, even for strings of
4688 quite modest length, because of the nested recursion. PCRE now does better in
4689 some of these cases. It does this by remembering the last required literal
4690 character in the pattern, and pre-searching the subject to ensure it is present
4691 before running the real match. In other words, it applies a heuristic to detect
4692 some types of certain failure quickly, and in the above example, if presented
4693 with a string that has no trailing " it gives "no match" very quickly.
4694
4695 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4696 other alternatives are tried instead.
4697
4698
4699 Version 2.06 09-Jun-99
4700 ----------------------
4701
4702 1. Change pcretest's output for amount of store used to show just the code
4703 space, because the remainder (the data block) varies in size between 32-bit and
4704 64-bit systems.
4705
4706 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4707 start matching at. This allows lookbehinds to work when searching for multiple
4708 occurrences in a string.
4709
4710 3. Added additional options to pcretest for testing multiple occurrences:
4711
4712 /+ outputs the rest of the string that follows a match
4713 /g loops for multiple occurrences, using the new startoffset argument
4714 /G loops for multiple occurrences by passing an incremented pointer
4715
4716 4. PCRE wasn't doing the "first character" optimization for patterns starting
4717 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4718 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4719 the letter 'x'. On long subject strings, this gives a significant speed-up.
4720
4721
4722 Version 2.05 21-Apr-99
4723 ----------------------
4724
4725 1. Changed the type of magic_number from int to long int so that it works
4726 properly on 16-bit systems.
4727
4728 2. Fixed a bug which caused patterns starting with .* not to work correctly
4729 when the subject string contained newline characters. PCRE was assuming
4730 anchoring for such patterns in all cases, which is not correct because .* will
4731 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4732 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4733 must be retried after every newline in the subject.
4734
4735
4736 Version 2.04 18-Feb-99
4737 ----------------------
4738
4739 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4740 computation of the store needed to hold the pattern was incorrect (too large).
4741 If such patterns were nested a few deep, this could multiply and become a real
4742 problem.
4743
4744 2. Added /M option to pcretest to show the memory requirement of a specific
4745 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4746
4747 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4748 compiled in such a way that the backtracking after subsequent failure was
4749 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4750 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4751
4752
4753 Version 2.03 02-Feb-99
4754 ----------------------
4755
4756 1. Fixed typo and small mistake in man page.
4757
4758 2. Added 4th condition (GPL supersedes if conflict) and created separate
4759 LICENCE file containing the conditions.
4760
4761 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4762 Perl, that is the internal \ allows the delimiter to be included in the
4763 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4764 the final delimiter, add \ to the end of the pattern (to test the error).
4765
4766 4. Added the convenience functions for extracting substrings after a successful
4767 match. Updated pcretest to make it able to test these functions.
4768
4769
4770 Version 2.02 14-Jan-99
4771 ----------------------
4772
4773 1. Initialized the working variables associated with each extraction so that
4774 their saving and restoring doesn't refer to uninitialized store.
4775
4776 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4777 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4778 fix the problem.
4779
4780 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4781 calls, and wasn't printing the correct value for compiling calls. Increased the
4782 default value of LOOPREPEAT, and the number of significant figures in the
4783 times.
4784
4785 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4786
4787 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4788 a building problem on Windows NT with a FAT file system.
4789
4790
4791 Version 2.01 21-Oct-98
4792 ----------------------
4793
4794 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4795 to character tables built by pcre_maketables() in the current locale. If NULL
4796 is passed, the default tables are used.
4797
4798
4799 Version 2.00 24-Sep-98
4800 ----------------------
4801
4802 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4803 it any more.
4804
4805 2. Allow quantification of (?>) groups, and make it work correctly.
4806
4807 3. The first character computation wasn't working for (?>) groups.
4808
4809 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4810 end of the subject) and add 5.005's \z, which really does match only at the
4811 very end of the subject.
4812
4813 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4814
4815 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4816 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4817 localized options. All options to pcre_study() were also removed.
4818
4819 7. Add other new features from 5.005:
4820
4821 $(?<= positive lookbehind
4822 $(?<! negative lookbehind
4823 (?imsx-imsx) added the unsetting capability
4824 such a setting is global if at outer level; local otherwise
4825 (?imsx-imsx:) non-capturing groups with option setting
4826 (?(cond)re|re) conditional pattern matching
4827
4828 A backreference to itself in a repeated group matches the previous
4829 captured string.
4830
4831 8. General tidying up of studying (both automatic and via "study")
4832 consequential on the addition of new assertions.
4833
4834 9. As in 5.005, unlimited repeated groups that could match an empty substring
4835 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4836 runtime if any iteration does actually match an empty substring.
4837
4838 10. Include the RunTest script in the distribution.
4839
4840 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4841 discrepancies, some of which were old and were also with respect to 5.004. They
4842 have now been fixed.
4843
4844
4845 Version 1.09 28-Apr-98
4846 ----------------------
4847
4848 1. A negated single character class followed by a quantifier with a minimum
4849 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4850 program crashes, or just wrong answers. This did not apply to negated classes
4851 containing more than one character, or to minima other than one.
4852
4853
4854 Version 1.08 27-Mar-98
4855 ----------------------
4856
4857 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4858
4859 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4860 latter must appear before anything that relies on it in the pattern.
4861
4862
4863 Version 1.07 16-Feb-98
4864 ----------------------
4865
4866 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4867 repeat of a potentially empty string).
4868
4869
4870 Version 1.06 23-Jan-98
4871 ----------------------
4872
4873 1. Added Markus Oberhumer's little patches for C++.
4874
4875 2. Literal strings longer than 255 characters were broken.
4876
4877
4878 Version 1.05 23-Dec-97
4879 ----------------------
4880
4881 1. Negated character classes containing more than one character were failing if
4882 PCRE_CASELESS was set at run time.
4883
4884
4885 Version 1.04 19-Dec-97
4886 ----------------------
4887
4888 1. Corrected the man page, where some "const" qualifiers had been omitted.
4889
4890 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4891 input syntax.
4892
4893 3. Fixed memory leak which occurred when a regex with back references was
4894 matched with an offsets vector that wasn't big enough. The temporary memory
4895 that is used in this case wasn't being freed if the match failed.
4896
4897 4. Tidied pcretest to ensure it frees memory that it gets.
4898
4899 5. Temporary memory was being obtained in the case where the passed offsets
4900 vector was exactly big enough.
4901
4902 6. Corrected definition of offsetof() from change 5 below.
4903
4904 7. I had screwed up change 6 below and broken the rules for the use of
4905 setjmp(). Now fixed.
4906
4907
4908 Version 1.03 18-Dec-97
4909 ----------------------
4910
4911 1. A erroneous regex with a missing opening parenthesis was correctly
4912 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4913 on some systems.
4914
4915 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4916 it was reported that one broken compiler failed on the former because "code" is
4917 also an independent variable.
4918
4919 3. The erroneous regex a[]b caused an array overrun reference.
4920
4921 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4922 fail on data ending with that character. (It was going on too far, and checking
4923 the next character, typically a binary zero.) This was specific to the
4924 optimized code for single-character negative classes.
4925
4926 5. Added a contributed patch from the TIN world which does the following:
4927
4928 + Add an undef for memmove, in case the the system defines a macro for it.
4929
4930 + Add a definition of offsetof(), in case there isn't one. (I don't know
4931 the reason behind this - offsetof() is part of the ANSI standard - but
4932 it does no harm).
4933
4934 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4935 most of the places where whitespace preceded '#'. I have given up and
4936 allowed the remaining 2 cases to be at the margin.
4937
4938 + Rename some variables in pcre to eliminate shadowing. This seems very
4939 pedantic, but does no harm, of course.
4940
4941 6. Moved the call to setjmp() into its own function, to get rid of warnings
4942 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4943
4944 7. Constructs such as \d{8,} were compiling into the equivalent of
4945 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4946 outcome, but in this particular case used more store than had been allocated,
4947 which caused the bug to be discovered because it threw up an internal error.
4948
4949 8. The debugging code in both pcre and pcretest for outputting the compiled
4950 form of a regex was going wrong in the case of back references followed by
4951 curly-bracketed repeats.
4952
4953
4954 Version 1.02 12-Dec-97
4955 ----------------------
4956
4957 1. Typos in pcre.3 and comments in the source fixed.
4958
4959 2. Applied a contributed patch to get rid of places where it used to remove
4960 'const' from variables, and fixed some signed/unsigned and uninitialized
4961 variable warnings.
4962
4963 3. Added the "runtest" target to Makefile.
4964
4965 4. Set default compiler flag to -O2 rather than just -O.
4966
4967
4968 Version 1.01 19-Nov-97
4969 ----------------------
4970
4971 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4972 like /([ab]*)*/, that is, for classes with more than one character in them.
4973
4974 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4975 as /((?>a*))*/ (a PCRE_EXTRA facility).
4976
4977
4978 Version 1.00 18-Nov-97
4979 ----------------------
4980
4981 1. Added compile-time macros to support systems such as SunOS4 which don't have
4982 memmove() or strerror() but have other things that can be used instead.
4983
4984 2. Arranged that "make clean" removes the executables.
4985
4986
4987 Version 0.99 27-Oct-97
4988 ----------------------
4989
4990 1. Fixed bug in code for optimizing classes with only one character. It was
4991 initializing a 32-byte map regardless, which could cause it to run off the end
4992 of the memory it had got.
4993
4994 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4995
4996
4997 Version 0.98 22-Oct-97
4998 ----------------------
4999
5000 1. Fixed bug in code for handling temporary memory usage when there are more
5001 back references than supplied space in the ovector. This could cause segfaults.
5002
5003
5004 Version 0.97 21-Oct-97
5005 ----------------------
5006
5007 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5008
5009 2. Optimized negated single characters not to use a bit map.
5010
5011 3. Brought error texts together as macro definitions; clarified some of them;
5012 fixed one that was wrong - it said "range out of order" when it meant "invalid
5013 escape sequence".
5014
5015 4. Changed some char * arguments to const char *.
5016
5017 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5018
5019 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5020 pcretest.
5021
5022
5023 Version 0.96 16-Oct-97
5024 ----------------------
5025
5026 1. Added a simple "pgrep" utility to the distribution.
5027
5028 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5029 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5030 where "ddd" means "one or more decimal digits".
5031
5032 3. Fixed serious bug. If a pattern had a back reference, but the call to
5033 pcre_exec() didn't supply a large enough ovector to record the related
5034 identifying subpattern, the match always failed. PCRE now remembers the number
5035 of the largest back reference, and gets some temporary memory in which to save
5036 the offsets during matching if necessary, in order to ensure that
5037 backreferences always work.
5038
5039 4. Increased the compatibility with Perl in a number of ways:
5040
5041 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5042 to request this handling. The option can be set at compile or exec time.
5043
5044 (b) $ matches before a terminating newline by default; an option
5045 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5046 mode). The option can be set at compile or exec time.
5047
5048 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5049 the same as Perl's. If the decimal number it represents is less than 10
5050 or there aren't that many previous left capturing parentheses, an octal
5051 escape is read. Inside a character class, it's always an octal escape,
5052 even if it is a single digit.
5053
5054 (d) An escaped but undefined alphabetic character is taken as a literal,
5055 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5056 escapes.
5057
5058 (e) {0} is now permitted. (The previous item is removed from the compiled
5059 pattern).
5060
5061 5. Changed all the names of code files so that the basic parts are no longer
5062 than 10 characters, and abolished the teeny "globals.c" file.
5063
5064 6. Changed the handling of character classes; they are now done with a 32-byte
5065 bit map always.
5066
5067 7. Added the -d and /D options to pcretest to make it possible to look at the
5068 internals of compilation without having to recompile pcre.
5069
5070
5071 Version 0.95 23-Sep-97
5072 ----------------------
5073
5074 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5075 \x20 at the start of a run of normal characters. These were being treated as
5076 real characters, instead of the source characters being re-checked.
5077
5078
5079 Version 0.94 18-Sep-97
5080 ----------------------
5081
5082 1. The functions are now thread-safe, with the caveat that the global variables
5083 containing pointers to malloc() and free() or alternative functions are the
5084 same for all threads.
5085
5086 2. Get pcre_study() to generate a bitmap of initial characters for non-
5087 anchored patterns when this is possible, and use it if passed to pcre_exec().
5088
5089
5090 Version 0.93 15-Sep-97
5091 ----------------------
5092
5093 1. /(b)|(:+)/ was computing an incorrect first character.
5094
5095 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5096 but not actually doing anything yet.
5097
5098 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5099 as Perl does (e.g. [-az] or [az-]).
5100
5101 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5102 all possible positions.
5103
5104 5. Split up into different modules to avoid including unneeded functions in a
5105 compiled binary. However, compile and exec are still in one module. The "study"
5106 function is split off.
5107
5108 6. The character tables are now in a separate module whose source is generated
5109 by an auxiliary program - but can then be edited by hand if required. There are
5110 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5111 toupper() in the code.
5112
5113 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5114 make them global. Abolish the function for setting them, as the caller can now
5115 set them directly.
5116
5117
5118 Version 0.92 11-Sep-97
5119 ----------------------
5120
5121 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5122 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5123
5124 2. Caseless matching was not working in character classes if the characters in
5125 the pattern were in upper case.
5126
5127 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5128
5129 4. Make PCRE_ANCHORED public and accept as a compile option.
5130
5131 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5132 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5133 pass them.
5134
5135 6. Give an error if bad option bits passed at compile or run time.
5136
5137 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5138 pcretest to cause it to pass that flag.
5139
5140 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5141 options, and the first character, if set.
5142
5143 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5144
5145
5146 Version 0.91 10-Sep-97
5147 ----------------------
5148
5149 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5150 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5151
5152 2. PCRE was looping on encountering an indefinitely repeated back reference to
5153 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5154 Perl does - treats the match as successful.
5155
5156 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5