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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1186 - (show annotations)
Sun Oct 28 17:57:32 2012 UTC (6 years, 11 months ago) by ph10
File size: 238337 byte(s)
Error occurred while calculating annotation data.
Use CHAR_NULL in pcre_compile when checking for character zero.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.32
5 ------------
6
7 1. Improved JIT compiler optimizations for first character search and single
8 character iterators.
9
10 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11 Patch by Daniel Richard G.
12
13 3. Single character iterator optimizations in the JIT compiler.
14
15 4. Improved JIT compiler optimizations for character ranges.
16
17 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18 Reported by Giuseppe D'Angelo.
19
20 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21 of a line, was being set incorrectly in cases where .* appeared inside
22 atomic brackets at the start of a pattern, or where there was a subsequent
23 *PRUNE or *SKIP.
24
25 7. Improved instruction cache flush for POWER/PowerPC.
26 Patch by Daniel Richard G.
27
28 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
29 grep:
30
31 (a) There is now no limit to the number of patterns to be matched.
32
33 (b) An error is given if a pattern is too long.
34
35 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
36 are now supported.
37
38 (d) --exclude-from and --include-from (multiple use) have been added.
39
40 (e) Exclusions and inclusions now apply to all files and directories, not
41 just to those obtained from scanning a directory recursively.
42
43 (f) Multiple uses of -f and --file-list are now supported.
44
45 (g) In a Windows environment, the default for -d has been changed from
46 "read" (the GNU grep default) to "skip", because otherwise the presence
47 of a directory in the file list provokes an error.
48
49 (h) The documentation has been revised and clarified in places.
50
51 9. Improve the matching speed of capturing brackets.
52
53 10. Changed the meaning of \X so that it now matches a Unicode extended
54 grapheme cluster.
55
56 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
57 out POSIX threads when JIT support is configured.
58
59 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
60
61 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
62 structure could go wrong in environments where size_t is not the same size
63 as int.
64
65 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
66 set.
67
68 15. The EBCDIC support had decayed; later updates to the code had included
69 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
70 general tidy up of EBCDIC-related issues, and the documentation was also
71 not quite right. There is now a test that can be run on ASCII systems to
72 check some of the EBCDIC-related things (but is it not a full test).
73
74 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
75 in a small tidy to the code.
76
77 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
78
79 18. If the --only-matching (-o) option in pcregrep is specified multiple
80 times, each one causes appropriate output. For example, -o1 -o2 outputs the
81 substrings matched by the 1st and 2nd capturing parentheses. A separating
82 string can be specified by --om-separator (default empty).
83
84 19. Improving the first n character searches.
85
86 20. Turn case lists for horizontal and vertical white space into macros so that
87 they are defined only once.
88
89 21. This set of changes together give more compatible Unicode case-folding
90 behaviour for characters that have more than one other case when UCP
91 support is available.
92
93 (a) The Unicode property table now has offsets into a new table of sets of
94 three or more characters that are case-equivalent. The MultiStage2.py
95 script that generates these tables (the pcre_ucd.c file) now scans
96 CaseFolding.txt instead of UnicodeData.txt for character case
97 information.
98
99 (b) The code for adding characters or ranges of characters to a character
100 class has been abstracted into a generalized function that also handles
101 case-independence. In UTF-mode with UCP support, this uses the new data
102 to handle characters with more than one other case.
103
104 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
105 whose other case is greater than 256 are now correctly matched
106 caselessly. Previously, the high codepoint matched the low one, but not
107 vice versa.
108
109 (d) The processing of \h, \H, \v, and \ in character classes now makes use
110 of the new class addition function, using character lists defined as
111 macros alongside the case definitions of 20 above.
112
113 (e) Caseless back references now work with characters that have more than
114 one other case.
115
116 (f) General caseless matching of characters with more than one other case
117 is supported.
118
119 22. Unicode character properties were updated from Unicode 6.2.0
120
121 23. Improved CMake support under Windows. Patch by Daniel Richard G.
122
123 24. Add support for 32-bit character strings, and UTF-32
124
125 25. Major JIT compiler update (code refactoring and bugfixing).
126 Experimental Sparc 32 support is added.
127
128 26. Applied a modified version of Daniel Richard G's patch to create
129 pcre.h.generic and config.h.generic by "make" instead of in the
130 PrepareRelease script.
131
132 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
133 pcre_compile.c when checking for a zero character.
134
135
136 Version 8.31 06-July-2012
137 -------------------------
138
139 1. Fixing a wrong JIT test case and some compiler warnings.
140
141 2. Removed a bashism from the RunTest script.
142
143 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
144 to unsigned type, result still unsigned" that was given by an MS compiler
145 on encountering the code "-sizeof(xxx)".
146
147 4. Partial matching support is added to the JIT compiler.
148
149 5. Fixed several bugs concerned with partial matching of items that consist
150 of more than one character:
151
152 (a) /^(..)\1/ did not partially match "aba" because checking references was
153 done on an "all or nothing" basis. This also applied to repeated
154 references.
155
156 (b) \R did not give a hard partial match if \r was found at the end of the
157 subject.
158
159 (c) \X did not give a hard partial match after matching one or more
160 characters at the end of the subject.
161
162 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
163 a partial match for the string "\r".
164
165 (e) When newline was set to CRLF, the metacharacter "." did not recognize
166 a partial match for a CR character at the end of the subject string.
167
168 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
169 running pcretest, the text "(JIT)" added to the output whenever JIT is
170 actually used to run the match.
171
172 7. Individual JIT compile options can be set in pcretest by following -s+[+]
173 or /S+[+] with a digit between 1 and 7.
174
175 8. OP_NOT now supports any UTF character not just single-byte ones.
176
177 9. (*MARK) control verb is now supported by the JIT compiler.
178
179 10. The command "./RunTest list" lists the available tests without actually
180 running any of them. (Because I keep forgetting what they all are.)
181
182 11. Add PCRE_INFO_MAXLOOKBEHIND.
183
184 12. Applied a (slightly modified) user-supplied patch that improves performance
185 when the heap is used for recursion (compiled with --disable-stack-for-
186 recursion). Instead of malloc and free for each heap frame each time a
187 logical recursion happens, frames are retained on a chain and re-used where
188 possible. This sometimes gives as much as 30% improvement.
189
190 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
191 call.
192
193 14. As documented, (*COMMIT) is now confined to within a positive assertion.
194
195 15. It is now possible to link pcretest with libedit as an alternative to
196 libreadline.
197
198 16. (*COMMIT) control verb is now supported by the JIT compiler.
199
200 17. The Unicode data tables have been updated to Unicode 6.1.0.
201
202 18. Added --file-list option to pcregrep.
203
204 19. Added binary file support to pcregrep, including the -a, --binary-files,
205 -I, and --text options.
206
207 20. The madvise function is renamed for posix_madvise for QNX compatibility
208 reasons. Fixed by Giuseppe D'Angelo.
209
210 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
211
212 22. Changed the option for creating symbolic links for 16-bit man pages from
213 -s to -sf so that re-installing does not cause issues.
214
215 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
216
217 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
218 matches in certain environments (the workspace was not being correctly
219 retained). Also added to pcre_dfa_exec() a simple plausibility check on
220 some of the workspace data at the beginning of a restart.
221
222 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
223 was not doing so when it should - probably a typo introduced by SVN 528
224 (change 8.10/14).
225
226 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
227 \w+ when the character tables indicated that \x{c4} was a word character.
228 There were several related cases, all because the tests for doing a table
229 lookup were testing for characters less than 127 instead of 255.
230
231 27. If a pattern contains capturing parentheses that are not used in a match,
232 their slots in the ovector are set to -1. For those that are higher than
233 any matched groups, this happens at the end of processing. In the case when
234 there were back references that the ovector was too small to contain
235 (causing temporary malloc'd memory to be used during matching), and the
236 highest capturing number was not used, memory off the end of the ovector
237 was incorrectly being set to -1. (It was using the size of the temporary
238 memory instead of the true size.)
239
240 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
241 ovector size, it uses memory at the end of the block that it has got.
242
243 29. Check for an overlong MARK name and give an error at compile time. The
244 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
245
246 30. JIT compiler update.
247
248 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
249 Rill for the patch.
250
251 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
252
253 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
254
255 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
256 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
257 was enabled.
258
259 35. Improve JIT code generation for greedy plus quantifier.
260
261 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
262 1 to "aa" instead of to an empty string. The bug affected repeated groups
263 that could potentially match an empty string.
264
265 37. Optimizing single character iterators in JIT.
266
267 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
268 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
269 codepoints that are too big for the mode are faulted, and in a UTF mode,
270 disallowed codepoints are also faulted.
271
272 39. If PCRE was compiled with UTF support, in three places in the DFA
273 matcher there was code that should only have been obeyed in UTF mode, but
274 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
275 processing when bytes with values greater than 127 were present. In 16-bit
276 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
277 both cases the values are those that cannot be the first data item in a UTF
278 character. The three items that might have provoked this were recursions,
279 possessively repeated groups, and atomic groups.
280
281 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
282 and pcregrep, because some OS require shared objects to be explicitly
283 passed to ld, causing the link step to fail if they are not.
284
285 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
286 mode, patterns that started with \h* or \R* might be incorrectly matched.
287
288
289 Version 8.30 04-February-2012
290 -----------------------------
291
292 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
293 name is defined in ctype.h.
294
295 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
296 only in quite long subpatterns.
297
298 3. Removed the function pcre_info(), which has been obsolete and deprecated
299 since it was replaced by pcre_fullinfo() in February 2000.
300
301 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
302 match a (*MARK), and the match failed at the start of the subject, a
303 reference to memory before the start of the subject could occur. This bug
304 was introduced by fix 17 of release 8.21.
305
306 5. A reference to an unset group with zero minimum repetition was giving
307 totally wrong answers (in non-JavaScript-compatibility mode). For example,
308 /(another)?(\1?)test/ matched against "hello world test". This bug was
309 introduced in release 8.13.
310
311 6. Add support for 16-bit character strings (a large amount of work involving
312 many changes and refactorings).
313
314 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
315 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
316 from a file.
317
318 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
319 rounding is not applied in this particular case).
320
321 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
322 if they appear, or are escaped, in patterns.
323
324 10. Get rid of a number of -Wunused-but-set-variable warnings.
325
326 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
327 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
328 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
329 also returns the mark "x". This bug applied to capturing parentheses,
330 non-capturing parentheses, and atomic parentheses. It also applied to some
331 assertions.
332
333 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
334 information out of configure.ac instead of relying on pcre.h.generic, which
335 is not stored in the repository.
336
337 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
338 -lreadline.
339
340 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
341
342 15. Applied Graycode's patch to put the top-level frame on the stack rather
343 than the heap when not using the stack for recursion. This gives a
344 performance improvement in many cases when recursion is not deep.
345
346 16. Experimental code added to "pcretest -C" to output the stack frame size.
347
348
349 Version 8.21 12-Dec-2011
350 ------------------------
351
352 1. Updating the JIT compiler.
353
354 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
355 are added as well.
356
357 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
358 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
359 calling _pcre_jit_exec. Some extra comments are added.
360
361 4. (*MARK) settings inside atomic groups that do not contain any capturing
362 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
363 was introduced by change 18 for 8.20.
364
365 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
366 ECMA-262 standard.
367
368 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
369 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
370 This bug was probably introduced by change 9 of 8.13.
371
372 7. While fixing 6 above, I noticed that a number of other items were being
373 incorrectly rejected as "not fixed length". This arose partly because newer
374 opcodes had not been added to the fixed-length checking code. I have (a)
375 corrected the bug and added tests for these items, and (b) arranged for an
376 error to occur if an unknown opcode is encountered while checking for fixed
377 length instead of just assuming "not fixed length". The items that were
378 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
379 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
380 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
381
382 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
383 being incorrectly compiled and would have given unpredicatble results.
384
385 9. A possessively repeated subpattern with minimum repeat count greater than
386 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
387 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
388 the first (A) could occur when it should not.
389
390 10. Add a cast and remove a redundant test from the code.
391
392 11. JIT should use pcre_malloc/pcre_free for allocation.
393
394 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
395 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
396 is anything other than /usr, -L is still shown).
397
398 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
399
400 14. Perl does not support \N without a following name in a [] class; PCRE now
401 also gives an error.
402
403 15. If a forward reference was repeated with an upper limit of around 2000,
404 it caused the error "internal error: overran compiling workspace". The
405 maximum number of forward references (including repeats) was limited by the
406 internal workspace, and dependent on the LINK_SIZE. The code has been
407 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
408 the default depends on LINK_SIZE. There is a new upper limit (for safety)
409 of around 200,000 forward references. While doing this, I also speeded up
410 the filling in of repeated forward references.
411
412 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
413 incorrectly expecting the subject to contain another "a" after the start.
414
415 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
416 in the match, the SKIP should be ignored. This was not happening; instead
417 the SKIP was being treated as NOMATCH. For patterns such as
418 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
419 tested.
420
421 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
422 now much more compatible with Perl, in particular in cases where the result
423 is a non-match for a non-anchored pattern. For example, if
424 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
425 "m", where previously it did not return a name. A side effect of this
426 change is that for partial matches, the last encountered mark name is
427 returned, as for non matches. A number of tests that were previously not
428 Perl-compatible have been moved into the Perl-compatible test files. The
429 refactoring has had the pleasing side effect of removing one argument from
430 the match() function, thus reducing its stack requirements.
431
432 19. If the /S+ option was used in pcretest to study a pattern using JIT,
433 subsequent uses of /S (without +) incorrectly behaved like /S+.
434
435 21. Retrieve executable code size support for the JIT compiler and fixing
436 some warnings.
437
438 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
439 not work when the shorter character appeared right at the end of the
440 subject string.
441
442 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
443 systems.
444
445 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
446 output it when the /M option is used in pcretest.
447
448 25. The CheckMan script was not being included in the distribution. Also, added
449 an explicit "perl" to run Perl scripts from the PrepareRelease script
450 because this is reportedly needed in Windows.
451
452 26. If study data was being save in a file and studying had not found a set of
453 "starts with" bytes for the pattern, the data written to the file (though
454 never used) was taken from uninitialized memory and so caused valgrind to
455 complain.
456
457 27. Updated RunTest.bat as provided by Sheri Pierce.
458
459 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
460
461 29. Computation of memory usage for the table of capturing group names was
462 giving an unnecessarily large value.
463
464
465 Version 8.20 21-Oct-2011
466 ------------------------
467
468 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
469 a POSIX class. After further experiments with Perl, which convinced me that
470 Perl has bugs and confusions, a closing square bracket is no longer allowed
471 in a POSIX name. This bug also affected patterns with classes that started
472 with full stops.
473
474 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
475 captured substring, but while checking the failing first alternative,
476 substring 1 is temporarily captured. If the output vector supplied to
477 pcre_exec() was not big enough for this capture, the yield of the function
478 was still zero ("insufficient space for captured substrings"). This cannot
479 be totally fixed without adding another stack variable, which seems a lot
480 of expense for a edge case. However, I have improved the situation in cases
481 such as /(a)(b)x|abc/ matched against "abc", where the return code
482 indicates that fewer than the maximum number of slots in the ovector have
483 been set.
484
485 3. Related to (2) above: when there are more back references in a pattern than
486 slots in the output vector, pcre_exec() uses temporary memory during
487 matching, and copies in the captures as far as possible afterwards. It was
488 using the entire output vector, but this conflicts with the specification
489 that only 2/3 is used for passing back captured substrings. Now it uses
490 only the first 2/3, for compatibility. This is, of course, another edge
491 case.
492
493 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
494 main code base, and can be used by building with --enable-jit. When this is
495 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
496 runtime --no-jit option is given.
497
498 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
499 ovector, the return from the function was zero, implying that there were
500 other matches that did not fit. The correct "exactly full" value is now
501 returned.
502
503 6. If a subpattern that was called recursively or as a subroutine contained
504 (*PRUNE) or any other control that caused it to give a non-standard return,
505 invalid errors such as "Error -26 (nested recursion at the same subject
506 position)" or even infinite loops could occur.
507
508 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
509 computing the minimum length on reaching *ACCEPT, and so ended up with the
510 wrong value of 1 rather than 0. Further investigation indicates that
511 computing a minimum subject length in the presence of *ACCEPT is difficult
512 (think back references, subroutine calls), and so I have changed the code
513 so that no minimum is registered for a pattern that contains *ACCEPT.
514
515 8. If (*THEN) was present in the first (true) branch of a conditional group,
516 it was not handled as intended. [But see 16 below.]
517
518 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
519 Sheri Pierce.
520
521 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
522 the first byte in a match must be "a".
523
524 11. Change 17 for 8.13 increased the recursion depth for patterns like
525 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
526 pattern contains any instances of (*THEN). If it does not, the old
527 optimizations are restored. It would be nice to do this on a per-group
528 basis, but at the moment that is not feasible.
529
530 12. In some environments, the output of pcretest -C is CRLF terminated. This
531 broke RunTest's code that checks for the link size. A single white space
532 character after the value is now allowed for.
533
534 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
535 For "fr", it uses the Windows-specific input and output files.
536
537 14. If (*THEN) appeared in a group that was called recursively or as a
538 subroutine, it did not work as intended. [But see next item.]
539
540 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
541 pattern fragments (but not containing any | characters). If A and B are
542 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
543 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
544 D. In other words, Perl considers parentheses that do not contain any |
545 characters to be part of a surrounding alternative, whereas PCRE was
546 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
547 differently. PCRE now behaves in the same way as Perl, except in the case
548 of subroutine/recursion calls such as (?1) which have in any case always
549 been different (but PCRE had them first :-).
550
551 16. Related to 15 above: Perl does not treat the | in a conditional group as
552 creating alternatives. Such a group is treated in the same way as an
553 ordinary group without any | characters when processing (*THEN). PCRE has
554 been changed to match Perl's behaviour.
555
556 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
557 RunGrepTest script failed.
558
559 18. Change 22 for version 13 caused atomic groups to use more stack. This is
560 inevitable for groups that contain captures, but it can lead to a lot of
561 stack use in large patterns. The old behaviour has been restored for atomic
562 groups that do not contain any capturing parentheses.
563
564 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
565 suppress the check for a minimum subject length at run time. (If it was
566 given to pcre_exec() or pcre_dfa_exec() it did work.)
567
568 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
569 fail to work when decoding hex characters in data strings in EBCDIC
570 environments.
571
572 21. It appears that in at least one Mac OS environment, the isxdigit() function
573 is implemented as a macro that evaluates to its argument more than once,
574 contravening the C 90 Standard (I haven't checked a later standard). There
575 was an instance in pcretest which caused it to go wrong when processing
576 \x{...} escapes in subject strings. The has been rewritten to avoid using
577 things like p++ in the argument of isxdigit().
578
579
580 Version 8.13 16-Aug-2011
581 ------------------------
582
583 1. The Unicode data tables have been updated to Unicode 6.0.0.
584
585 2. Two minor typos in pcre_internal.h have been fixed.
586
587 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
588 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
589 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
590
591 4. There were a number of related bugs in the code for matching backrefences
592 caselessly in UTF-8 mode when codes for the characters concerned were
593 different numbers of bytes. For example, U+023A and U+2C65 are an upper
594 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
595 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
596 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
597 2-byte code at the end of the subject (it thought there wasn't enough data
598 left).
599
600 5. Comprehensive information about what went wrong is now returned by
601 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
602 as the output vector has at least 2 elements. The offset of the start of
603 the failing character and a reason code are placed in the vector.
604
605 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
606 now returned is for the first byte of the failing character, instead of the
607 last byte inspected. This is an incompatible change, but I hope it is small
608 enough not to be a problem. It makes the returned offset consistent with
609 pcre_exec() and pcre_dfa_exec().
610
611 7. pcretest now gives a text phrase as well as the error number when
612 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
613 failure, the offset and reason code are output.
614
615 8. When \R was used with a maximizing quantifier it failed to skip backwards
616 over a \r\n pair if the subsequent match failed. Instead, it just skipped
617 back over a single character (\n). This seems wrong (because it treated the
618 two characters as a single entity when going forwards), conflicts with the
619 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
620 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
621 has been changed.
622
623 9. Some internal refactoring has changed the processing so that the handling
624 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
625 time (the PCRE_DOTALL option was changed this way some time ago: version
626 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
627 which was always a bit of a fudge. It also means that there is one less
628 argument for the match() function, which reduces its stack requirements
629 slightly. This change also fixes an incompatibility with Perl: the pattern
630 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
631
632 10. More internal refactoring has drastically reduced the number of recursive
633 calls to match() for possessively repeated groups such as (abc)++ when
634 using pcre_exec().
635
636 11. While implementing 10, a number of bugs in the handling of groups were
637 discovered and fixed:
638
639 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
640 (a|)*(?1) gave a compile-time internal error.
641 ((a|)+)+ did not notice that the outer group could match an empty string.
642 (^a|^)+ was not marked as anchored.
643 (.*a|.*)+ was not marked as matching at start or after a newline.
644
645 12. Yet more internal refactoring has removed another argument from the match()
646 function. Special calls to this function are now indicated by setting a
647 value in a variable in the "match data" data block.
648
649 13. Be more explicit in pcre_study() instead of relying on "default" for
650 opcodes that mean there is no starting character; this means that when new
651 ones are added and accidentally left out of pcre_study(), testing should
652 pick them up.
653
654 14. The -s option of pcretest has been documented for ages as being an old
655 synonym of -m (show memory usage). I have changed it to mean "force study
656 for every regex", that is, assume /S for every regex. This is similar to -i
657 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
658 using it. It makes it easier to run collections of tests with and without
659 study enabled, and thereby test pcre_study() more easily. All the standard
660 tests are now run with and without -s (but some patterns can be marked as
661 "never study" - see 20 below).
662
663 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
664 restoration of the capturing data to the outer values was not happening
665 correctly.
666
667 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
668 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
669 pattern had matched an empty string, and so incorrectly returned a no
670 match.
671
672 17. There was optimizing code for the last branch of non-capturing parentheses,
673 and also for the obeyed branch of a conditional subexpression, which used
674 tail recursion to cut down on stack usage. Unfortunately, now that there is
675 the possibility of (*THEN) occurring in these branches, tail recursion is
676 no longer possible because the return has to be checked for (*THEN). These
677 two optimizations have therefore been removed. [But see 8.20/11 above.]
678
679 18. If a pattern containing \R was studied, it was assumed that \R always
680 matched two bytes, thus causing the minimum subject length to be
681 incorrectly computed because \R can also match just one byte.
682
683 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
684 was incorrectly computed.
685
686 20. If /S is present twice on a test pattern in pcretest input, it now
687 *disables* studying, thereby overriding the use of -s on the command line
688 (see 14 above). This is necessary for one or two tests to keep the output
689 identical in both cases.
690
691 21. When (*ACCEPT) was used in an assertion that matched an empty string and
692 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
693
694 22. When an atomic group that contained a capturing parenthesis was
695 successfully matched, but the branch in which it appeared failed, the
696 capturing was not being forgotten if a higher numbered group was later
697 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
698 group 1 to "a", when in fact it should be unset. This applied to multi-
699 branched capturing and non-capturing groups, repeated or not, and also to
700 positive assertions (capturing in negative assertions does not happen
701 in PCRE) and also to nested atomic groups.
702
703 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
704 subject after a captured substring, to make it easier to tell which of a
705 number of identical substrings has been captured.
706
707 24. The way atomic groups are processed by pcre_exec() has been changed so that
708 if they are repeated, backtracking one repetition now resets captured
709 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
710 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
711 "aaa". Previously, it would have been "a". As part of this code
712 refactoring, the way recursive calls are handled has also been changed.
713
714 25. If an assertion condition captured any substrings, they were not passed
715 back unless some other capturing happened later. For example, if
716 (?(?=(a))a) was matched against "a", no capturing was returned.
717
718 26. When studying a pattern that contained subroutine calls or assertions,
719 the code for finding the minimum length of a possible match was handling
720 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
721 group 1 called group 2 while simultaneously a separate group 2 called group
722 1). A stack overflow occurred in this case. I have fixed this by limiting
723 the recursion depth to 10.
724
725 27. Updated RunTest.bat in the distribution to the version supplied by Tom
726 Fortmann. This supports explicit test numbers on the command line, and has
727 argument validation and error reporting.
728
729 28. An instance of \X with an unlimited repeat could fail if at any point the
730 first character it looked at was a mark character.
731
732 29. Some minor code refactoring concerning Unicode properties and scripts
733 should reduce the stack requirement of match() slightly.
734
735 30. Added the '=' option to pcretest to check the setting of unused capturing
736 slots at the end of the pattern, which are documented as being -1, but are
737 not included in the return count.
738
739 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
740 compiled something random. Now it gives a compile-time error (as does
741 Perl).
742
743 32. A *MARK encountered during the processing of a positive assertion is now
744 recorded and passed back (compatible with Perl).
745
746 33. If --only-matching or --colour was set on a pcregrep call whose pattern
747 had alternative anchored branches, the search for a second match in a line
748 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
749 matched the line "0102" twice. The same bug affected patterns that started
750 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
751 twice.
752
753 34. Previously, PCRE did not allow quantification of assertions. However, Perl
754 does, and because of capturing effects, quantifying parenthesized
755 assertions may at times be useful. Quantifiers are now allowed for
756 parenthesized assertions.
757
758 35. A minor code tidy in pcre_compile() when checking options for \R usage.
759
760 36. \g was being checked for fancy things in a character class, when it should
761 just be a literal "g".
762
763 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
764 appearance of a nested POSIX class supersedes an apparent external class.
765 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
766 unescaped square brackets may also appear as part of class names. For
767 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
768 more like Perl. (But see 8.20/1 above.)
769
770 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
771 was because it thought it was \N{name}, which is not supported).
772
773 39. Add minix to OS list not supporting the -S option in pcretest.
774
775 40. PCRE tries to detect cases of infinite recursion at compile time, but it
776 cannot analyze patterns in sufficient detail to catch mutual recursions
777 such as ((?1))((?2)). There is now a runtime test that gives an error if a
778 subgroup is called recursively as a subpattern for a second time at the
779 same position in the subject string. In previous releases this might have
780 been caught by the recursion limit, or it might have run out of stack.
781
782 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
783 happen only once. PCRE was, however incorrectly giving a compile time error
784 "recursive call could loop indefinitely" because it cannot analyze the
785 pattern in sufficient detail. The compile time test no longer happens when
786 PCRE is compiling a conditional subpattern, but actual runaway loops are
787 now caught at runtime (see 40 above).
788
789 42. It seems that Perl allows any characters other than a closing parenthesis
790 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
791 has been changed to be the same.
792
793 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
794 as not to get warnings when autogen.sh is called. Also changed
795 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
796
797 44. To help people who use pcregrep to scan files containing exceedingly long
798 lines, the following changes have been made:
799
800 (a) The default value of the buffer size parameter has been increased from
801 8K to 20K. (The actual buffer used is three times this size.)
802
803 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
804 PCRE is built.
805
806 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
807 to be set at run time.
808
809 (d) Numerical values in pcregrep options can be followed by K or M, for
810 example --buffer-size=50K.
811
812 (e) If a line being scanned overflows pcregrep's buffer, an error is now
813 given and the return code is set to 2.
814
815 45. Add a pointer to the latest mark to the callout data block.
816
817 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
818 partial match of an empty string instead of no match. This was specific to
819 the use of ".".
820
821 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
822 complete match instead of a partial match. This bug was dependent on both
823 the PCRE_UTF8 and PCRE_DOTALL options being set.
824
825 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
826 starting byte set, because \b was not being ignored.
827
828
829 Version 8.12 15-Jan-2011
830 ------------------------
831
832 1. Fixed some typos in the markup of the man pages, and wrote a script that
833 checks for such things as part of the documentation building process.
834
835 2. On a big-endian 64-bit system, pcregrep did not correctly process the
836 --match-limit and --recursion-limit options (added for 8.11). In
837 particular, this made one of the standard tests fail. (The integer value
838 went into the wrong half of a long int.)
839
840 3. If the --colour option was given to pcregrep with -v (invert match), it
841 did strange things, either producing crazy output, or crashing. It should,
842 of course, ignore a request for colour when reporting lines that do not
843 match.
844
845 4. Another pcregrep bug caused similar problems if --colour was specified with
846 -M (multiline) and the pattern match finished with a line ending.
847
848 5. In pcregrep, when a pattern that ended with a literal newline sequence was
849 matched in multiline mode, the following line was shown as part of the
850 match. This seems wrong, so I have changed it.
851
852 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
853 the check for further matches in the same line (so they could be coloured)
854 to overrun the end of the current line. If another match was found, it was
855 incorrectly shown (and then shown again when found in the next line).
856
857 7. If pcregrep was compiled under Windows, there was a reference to the
858 function pcregrep_exit() before it was defined. I am assuming this was
859 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
860 reported by a user. I've moved the definition above the reference.
861
862
863 Version 8.11 10-Dec-2010
864 ------------------------
865
866 1. (*THEN) was not working properly if there were untried alternatives prior
867 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
868 backtracked to try for "b" instead of moving to the next alternative branch
869 at the same level (in this case, to look for "c"). The Perl documentation
870 is clear that when (*THEN) is backtracked onto, it goes to the "next
871 alternative in the innermost enclosing group".
872
873 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
874 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
875 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
876 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
877 (*THEN).
878
879 3. If \s appeared in a character class, it removed the VT character from
880 the class, even if it had been included by some previous item, for example
881 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
882 of \s, but is part of the POSIX "space" class.)
883
884 4. A partial match never returns an empty string (because you can always
885 match an empty string at the end of the subject); however the checking for
886 an empty string was starting at the "start of match" point. This has been
887 changed to the "earliest inspected character" point, because the returned
888 data for a partial match starts at this character. This means that, for
889 example, /(?<=abc)def/ gives a partial match for the subject "abc"
890 (previously it gave "no match").
891
892 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
893 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
894 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
895 has an implication that the given string is incomplete (because a partial
896 match is preferred over a full match). For this reason, these items now
897 give a partial match in this situation. [Aside: previously, the one case
898 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
899 match rather than a full match, which was wrong by the old rules, but is
900 now correct.]
901
902 6. There was a bug in the handling of #-introduced comments, recognized when
903 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
904 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
905 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
906 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
907 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
908 places in pcre_compile().
909
910 7. Related to (6) above, when pcre_compile() was skipping #-introduced
911 comments when looking ahead for named forward references to subpatterns,
912 the only newline sequence it recognized was NL. It now handles newlines
913 according to the set newline convention.
914
915 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
916 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
917 cater for a lack of strerror(). These oversights have been fixed.
918
919 9. Added --match-limit and --recursion-limit to pcregrep.
920
921 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
922
923 11. When the -o option was used, pcregrep was setting a return code of 1, even
924 when matches were found, and --line-buffered was not being honoured.
925
926 12. Added an optional parentheses number to the -o and --only-matching options
927 of pcregrep.
928
929 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
930 can match an empty string. The code to do it in pcretest and pcredemo
931 needed fixing:
932
933 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
934 only one byte after an empty string match just before CRLF (this case
935 just got forgotten; "any" and "anycrlf" were OK).
936
937 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
938 mode when an empty string match preceded an ASCII character followed by
939 a non-ASCII character. (The code for advancing by one character rather
940 than one byte was nonsense.)
941
942 (c) The pcredemo.c sample program did not have any code at all to handle
943 the cases when CRLF is a valid newline sequence.
944
945 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
946 as a starting offset was within the subject string. There is now a new
947 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
948 negative or greater than the length of the string. In order to test this,
949 pcretest is extended to allow the setting of negative starting offsets.
950
951 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
952 starting offset points to the beginning of a UTF-8 character was
953 unnecessarily clumsy. I tidied it up.
954
955 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
956 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
957
958 17. Nobody had reported that the --include_dir option, which was added in
959 release 7.7 should have been called --include-dir (hyphen, not underscore)
960 for compatibility with GNU grep. I have changed it to --include-dir, but
961 left --include_dir as an undocumented synonym, and the same for
962 --exclude-dir, though that is not available in GNU grep, at least as of
963 release 2.5.4.
964
965 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
966 characters from a string of bytes) have been redefined so as not to use
967 loops, in order to improve performance in some environments. At the same
968 time, I abstracted some of the common code into auxiliary macros to save
969 repetition (this should not affect the compiled code).
970
971 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
972 compile-time error is now given if \c is not followed by an ASCII
973 character, that is, a byte less than 128. (In EBCDIC mode, the code is
974 different, and any byte value is allowed.)
975
976 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
977 START_OPTIMIZE option, which is now allowed at compile time - but just
978 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
979 to pcregrep and other applications that have no direct access to PCRE
980 options. The new /Y option in pcretest sets this option when calling
981 pcre_compile().
982
983 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
984 back references. Groups containing recursive back references were forced to
985 be atomic by that change, but in the case of named groups, the amount of
986 memory required was incorrectly computed, leading to "Failed: internal
987 error: code overflow". This has been fixed.
988
989 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
990 pcretest.c, to avoid build problems in some Borland environments.
991
992
993 Version 8.10 25-Jun-2010
994 ------------------------
995
996 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
997 THEN.
998
999 2. (*ACCEPT) was not working when inside an atomic group.
1000
1001 3. Inside a character class, \B is treated as a literal by default, but
1002 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1003 causes the error). The code is unchanged, but I tidied the documentation.
1004
1005 4. Inside a character class, PCRE always treated \R and \X as literals,
1006 whereas Perl faults them if its -w option is set. I have changed PCRE so
1007 that it faults them when PCRE_EXTRA is set.
1008
1009 5. Added support for \N, which always matches any character other than
1010 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1011
1012 6. When compiling pcregrep with newer versions of gcc which may have
1013 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1014 declared with attribute warn_unused_result" were given. Just casting the
1015 result to (void) does not stop the warnings; a more elaborate fudge is
1016 needed. I've used a macro to implement this.
1017
1018 7. Minor change to pcretest.c to avoid a compiler warning.
1019
1020 8. Added four artifical Unicode properties to help with an option to make
1021 \s etc use properties (see next item). The new properties are: Xan
1022 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1023
1024 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1025 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1026 this option. Modified pcretest to add /W to test this facility. Added
1027 REG_UCP to make it available via the POSIX interface.
1028
1029 10. Added --line-buffered to pcregrep.
1030
1031 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1032 studied, and the match started with a letter with a code point greater than
1033 127 whose first byte was different to the first byte of the other case of
1034 the letter, the other case of this starting letter was not recognized
1035 (#976).
1036
1037 12. If a pattern that was studied started with a repeated Unicode property
1038 test, for example, \p{Nd}+, there was the theoretical possibility of
1039 setting up an incorrect bitmap of starting bytes, but fortunately it could
1040 not have actually happened in practice until change 8 above was made (it
1041 added property types that matched character-matching opcodes).
1042
1043 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1044 possible starting bytes for non-anchored patterns.
1045
1046 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1047 \R, and also a number of cases that involve Unicode properties, both
1048 explicit and implicit when PCRE_UCP is set.
1049
1050 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1051 input, it could crash or give wrong results if characters with values
1052 greater than 0xc0 were present in the subject string. (Detail: it assumed
1053 UTF-8 input when processing these items.)
1054
1055 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1056 size_t is 64-bit (#991).
1057
1058 17. Added a check for running out of memory when PCRE is compiled with
1059 --disable-stack-for-recursion (#990).
1060
1061 18. If the last data line in a file for pcretest does not have a newline on
1062 the end, a newline was missing in the output.
1063
1064 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1065 less than 128) in its various bitmaps. However, there is a facility for
1066 generating tables according to the current locale when PCRE is compiled. It
1067 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1068 characters, are recognized by isspace() and therefore were getting set in
1069 these tables, and indeed these tables seem to approximate to ISO 8859. This
1070 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1071 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1072 which of course cannot start UTF-8 characters. I have changed the code so
1073 that only real ASCII characters (less than 128) and the correct starting
1074 bytes for UTF-8 encodings are set for characters greater than 127 when in
1075 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1076 altogether.)
1077
1078 20. Added the /T option to pcretest so as to be able to run tests with non-
1079 standard character tables, thus making it possible to include the tests
1080 used for 19 above in the standard set of tests.
1081
1082 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1083 reference to a subpattern the other side of a comment that contains an
1084 opening parenthesis caused either an internal compiling error, or a
1085 reference to the wrong subpattern.
1086
1087
1088 Version 8.02 19-Mar-2010
1089 ------------------------
1090
1091 1. The Unicode data tables have been updated to Unicode 5.2.0.
1092
1093 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1094 configured.
1095
1096 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1097 original author of that file, following a query about its status.
1098
1099 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1100 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1101
1102 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1103 quantifier applied to a forward-referencing subroutine call, could compile
1104 incorrect code or give the error "internal error: previously-checked
1105 referenced subpattern not found".
1106
1107 6. Both MS Visual Studio and Symbian OS have problems with initializing
1108 variables to point to external functions. For these systems, therefore,
1109 pcre_malloc etc. are now initialized to local functions that call the
1110 relevant global functions.
1111
1112 7. There were two entries missing in the vectors called coptable and poptable
1113 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1114 I've fixed the data, and added a kludgy way of testing at compile time that
1115 the lengths are correct (equal to the number of opcodes).
1116
1117 8. Following on from 7, I added a similar kludge to check the length of the
1118 eint vector in pcreposix.c.
1119
1120 9. Error texts for pcre_compile() are held as one long string to avoid too
1121 much relocation at load time. To find a text, the string is searched,
1122 counting zeros. There was no check for running off the end of the string,
1123 which could happen if a new error number was added without updating the
1124 string.
1125
1126 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1127
1128 11. \K was not working if it appeared in an atomic group or in a group that
1129 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1130 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1131 the assertion is positive, but not if it is negative.
1132
1133 12. Change 11 fortuitously reduced the size of the stack frame used in the
1134 "match()" function of pcre_exec.c by one pointer. Forthcoming
1135 implementation of support for (*MARK) will need an extra pointer on the
1136 stack; I have reserved it now, so that the stack frame size does not
1137 decrease.
1138
1139 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1140 item in branch that calls a recursion is a subroutine call - as in the
1141 second branch in the above example - was incorrectly given the compile-
1142 time error "recursive call could loop indefinitely" because pcre_compile()
1143 was not correctly checking the subroutine for matching a non-empty string.
1144
1145 14. The checks for overrunning compiling workspace could trigger after an
1146 overrun had occurred. This is a "should never occur" error, but it can be
1147 triggered by pathological patterns such as hundreds of nested parentheses.
1148 The checks now trigger 100 bytes before the end of the workspace.
1149
1150 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1151
1152
1153 Version 8.01 19-Jan-2010
1154 ------------------------
1155
1156 1. If a pattern contained a conditional subpattern with only one branch (in
1157 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1158 computed the wrong minimum data length (which is of course zero for such
1159 subpatterns). This could cause incorrect "no match" results.
1160
1161 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1162 the pattern is reset in the first branch, pcre_compile() failed with
1163 "internal error: code overflow at offset...". This happened only when
1164 the reset was to the original external option setting. (An optimization
1165 abstracts leading options settings into an external setting, which was the
1166 cause of this.)
1167
1168 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1169 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1170 assertion pattern did not match (meaning that the assertion was true), it
1171 was incorrectly treated as false if the SKIP had been reached during the
1172 matching. This also applied to assertions used as conditions.
1173
1174 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1175 assertion subpattern, including such a pattern used as a condition,
1176 unpredictable results occurred, instead of the error return
1177 PCRE_ERROR_DFA_UITEM.
1178
1179 5. The C++ GlobalReplace function was not working like Perl for the special
1180 situation when an empty string is matched. It now does the fancy magic
1181 stuff that is necessary.
1182
1183 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1184 removed. (These were left over from very, very early versions of PCRE.)
1185
1186 7. Some cosmetic changes to the code to make life easier when compiling it
1187 as part of something else:
1188
1189 (a) Change DEBUG to PCRE_DEBUG.
1190
1191 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1192 called "current" as "current_branch", to prevent a collision with the
1193 Linux macro when compiled as a kernel module.
1194
1195 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1196 prevent a collision with the Linux macro when compiled as a kernel
1197 module.
1198
1199 8. In pcre_compile() there are some checks for integer overflows that used to
1200 cast potentially large values to (double). This has been changed to that
1201 when building, a check for int64_t is made, and if it is found, it is used
1202 instead, thus avoiding the use of floating point arithmetic. (There is no
1203 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1204 double.
1205
1206 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1207 2005 (difference between two addresses compared to an unsigned value).
1208
1209 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1210 custom one, because of the following reported problem in Windows:
1211
1212 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1213 under Win32.
1214 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1215 therefore missing the function definition.
1216 - The compiler thus generates a "C" signature for the test function.
1217 - The linker fails to find the "C" function.
1218 - PCRE fails to configure if asked to do so against libbz2.
1219
1220 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1221 messages were output:
1222
1223 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1224 rerunning libtoolize, to keep the correct libtool macros in-tree.
1225 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1226
1227 I have done both of these things.
1228
1229 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1230 most of the time, it *can* run out if it is given a pattern that contains a
1231 runaway infinite recursion. I updated the discussion in the pcrestack man
1232 page.
1233
1234 13. Now that we have gone to the x.xx style of version numbers, the minor
1235 version may start with zero. Using 08 or 09 is a bad idea because users
1236 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1237 interpreted as invalid octal numbers. I've updated the previous comment in
1238 configure.ac, and also added a check that gives an error if 08 or 09 are
1239 used.
1240
1241 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1242 causing partial matching to fail when the end of the subject matched \W
1243 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1244
1245 15. There were some discrepancies between the declarations in pcre_internal.h
1246 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1247 their definitions. The declarations used "const uschar *" and the
1248 definitions used USPTR. Even though USPTR is normally defined as "const
1249 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1250 reported that: "This difference in casting confuses some C++ compilers, for
1251 example, SunCC recognizes above declarations as different functions and
1252 generates broken code for hbpcre." I have changed the declarations to use
1253 USPTR.
1254
1255 16. GNU libtool is named differently on some systems. The autogen.sh script now
1256 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1257 (FreeBSD).
1258
1259 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1260 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1261 comment: "Figure out how to create a longlong from a string: strtoll and
1262 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1263 instance, but it only takes 2 args instead of 3!"
1264
1265 18. A subtle bug concerned with back references has been fixed by a change of
1266 specification, with a corresponding code fix. A pattern such as
1267 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1268 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1269 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1270 same bug. Such groups have to be quantified to be useful, or contained
1271 inside another quantified group. (If there's no repetition, the reference
1272 can never match.) The problem arises because, having left the group and
1273 moved on to the rest of the pattern, a later failure that backtracks into
1274 the group uses the captured value from the final iteration of the group
1275 rather than the correct earlier one. I have fixed this in PCRE by forcing
1276 any group that contains a reference to itself to be an atomic group; that
1277 is, there cannot be any backtracking into it once it has completed. This is
1278 similar to recursive and subroutine calls.
1279
1280
1281 Version 8.00 19-Oct-09
1282 ----------------------
1283
1284 1. The table for translating pcre_compile() error codes into POSIX error codes
1285 was out-of-date, and there was no check on the pcre_compile() error code
1286 being within the table. This could lead to an OK return being given in
1287 error.
1288
1289 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1290 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1291 in a Windows environment.
1292
1293 3. The pcregrep --count option prints the count for each file even when it is
1294 zero, as does GNU grep. However, pcregrep was also printing all files when
1295 --files-with-matches was added. Now, when both options are given, it prints
1296 counts only for those files that have at least one match. (GNU grep just
1297 prints the file name in this circumstance, but including the count seems
1298 more useful - otherwise, why use --count?) Also ensured that the
1299 combination -clh just lists non-zero counts, with no names.
1300
1301 4. The long form of the pcregrep -F option was incorrectly implemented as
1302 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1303 but it seems right to fix it, and I didn't think it was worth preserving
1304 the old behaviour.
1305
1306 5. The command line items --regex=pattern and --regexp=pattern were not
1307 recognized by pcregrep, which required --regex pattern or --regexp pattern
1308 (with a space rather than an '='). The man page documented the '=' forms,
1309 which are compatible with GNU grep; these now work.
1310
1311 6. No libpcreposix.pc file was created for pkg-config; there was just
1312 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1313
1314 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1315 when UCP support is not needed, by modifying the Python script that
1316 generates it from Unicode data files. This should not matter if the module
1317 is correctly used as a library, but I received one complaint about 50K of
1318 unwanted data. My guess is that the person linked everything into his
1319 program rather than using a library. Anyway, it does no harm.
1320
1321 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1322 was a minimum greater than 1 for a wide character in a possessive
1323 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1324 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1325 character. Chaos in the form of incorrect output or a compiling loop could
1326 result.
1327
1328 9. The restrictions on what a pattern can contain when partial matching is
1329 requested for pcre_exec() have been removed. All patterns can now be
1330 partially matched by this function. In addition, if there are at least two
1331 slots in the offset vector, the offset of the earliest inspected character
1332 for the match and the offset of the end of the subject are set in them when
1333 PCRE_ERROR_PARTIAL is returned.
1334
1335 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1336 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1337 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1338 and may be more useful for multi-segment matching.
1339
1340 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1341 used to be given if ever the end of the subject was reached; now it is
1342 given only if matching could not proceed because another character was
1343 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1344 string "Z", which now yields "no match" instead of "partial match". In the
1345 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1346 final character ended with (*FAIL).
1347
1348 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1349 if the pattern had a "must contain" character that was already found in the
1350 earlier partial match, unless partial matching was again requested. For
1351 example, with the pattern /dog.(body)?/, the "must contain" character is
1352 "g". If the first part-match was for the string "dog", restarting with
1353 "sbody" failed. This bug has been fixed.
1354
1355 13. The string returned by pcre_dfa_exec() after a partial match has been
1356 changed so that it starts at the first inspected character rather than the
1357 first character of the match. This makes a difference only if the pattern
1358 starts with a lookbehind assertion or \b or \B (\K is not supported by
1359 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1360 matching functions compatible, and I think it's the right thing to do.
1361
1362 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1363 so that the demonstration program is easily available in environments where
1364 PCRE has not been installed from source.
1365
1366 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1367 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1368 library.
1369
1370 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1371 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1372 is not the first non-POSIX option to be added. Clearly some people find
1373 these options useful.
1374
1375 17. If a caller to the POSIX matching function regexec() passes a non-zero
1376 value for nmatch with a NULL value for pmatch, the value of
1377 nmatch is forced to zero.
1378
1379 18. RunGrepTest did not have a test for the availability of the -u option of
1380 the diff command, as RunTest does. It now checks in the same way as
1381 RunTest, and also checks for the -b option.
1382
1383 19. If an odd number of negated classes containing just a single character
1384 interposed, within parentheses, between a forward reference to a named
1385 subpattern and the definition of the subpattern, compilation crashed with
1386 an internal error, complaining that it could not find the referenced
1387 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1388 [The bug was that it was starting one character too far in when skipping
1389 over the character class, thus treating the ] as data rather than
1390 terminating the class. This meant it could skip too much.]
1391
1392 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1393 /g option in pcretest when the pattern contains \K, which makes it possible
1394 to have an empty string match not at the start, even when the pattern is
1395 anchored. Updated pcretest and pcredemo to use this option.
1396
1397 21. If the maximum number of capturing subpatterns in a recursion was greater
1398 than the maximum at the outer level, the higher number was returned, but
1399 with unset values at the outer level. The correct (outer level) value is
1400 now given.
1401
1402 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1403 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1404 make it do so. The string so far is captured, making this feature
1405 compatible with Perl.
1406
1407 23. The tests have been re-organized, adding tests 11 and 12, to make it
1408 possible to check the Perl 5.10 features against Perl 5.10.
1409
1410 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1411 pattern matches a fixed length string. PCRE did not allow this; now it
1412 does. Neither allows recursion.
1413
1414 25. I finally figured out how to implement a request to provide the minimum
1415 length of subject string that was needed in order to match a given pattern.
1416 (It was back references and recursion that I had previously got hung up
1417 on.) This code has now been added to pcre_study(); it finds a lower bound
1418 to the length of subject needed. It is not necessarily the greatest lower
1419 bound, but using it to avoid searching strings that are too short does give
1420 some useful speed-ups. The value is available to calling programs via
1421 pcre_fullinfo().
1422
1423 26. While implementing 25, I discovered to my embarrassment that pcretest had
1424 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1425 study optimizations had never been tested with that matching function.
1426 Oops. What is worse, even when it was passed study data, there was a bug in
1427 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1428 were also very few tests of studied patterns with pcre_dfa_exec().
1429
1430 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1431 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1432 on the other side of the coin, they are no longer allowed to have different
1433 names, because these cannot be distinguished in PCRE, and this has caused
1434 confusion. (This is a difference from Perl.)
1435
1436 28. When duplicate subpattern names are present (necessarily with different
1437 numbers, as required by 27 above), and a test is made by name in a
1438 conditional pattern, either for a subpattern having been matched, or for
1439 recursion in such a pattern, all the associated numbered subpatterns are
1440 tested, and the overall condition is true if the condition is true for any
1441 one of them. This is the way Perl works, and is also more like the way
1442 testing by number works.
1443
1444
1445 Version 7.9 11-Apr-09
1446 ---------------------
1447
1448 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1449 (pcretest), all targets were linked against these libraries. This included
1450 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1451 libraries. This caused unwanted dependencies to be created. This problem
1452 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1453 pcretest is linked with readline.
1454
1455 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1456 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1457 moved outside it again, because FALSE and TRUE are already defined in AIX,
1458 but BOOL is not.
1459
1460 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1461 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1462
1463 4. The pcregrep documentation said spaces were inserted as well as colons (or
1464 hyphens) following file names and line numbers when outputting matching
1465 lines. This is not true; no spaces are inserted. I have also clarified the
1466 wording for the --colour (or --color) option.
1467
1468 5. In pcregrep, when --colour was used with -o, the list of matching strings
1469 was not coloured; this is different to GNU grep, so I have changed it to be
1470 the same.
1471
1472 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1473 each matching line was coloured. Now it goes on to look for further matches
1474 of any of the test patterns, which is the same behaviour as GNU grep.
1475
1476 7. A pattern that could match an empty string could cause pcregrep to loop; it
1477 doesn't make sense to accept an empty string match in pcregrep, so I have
1478 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1479 seems to be how GNU grep behaves.
1480
1481 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1482 start or after a newline", because the conditional assertion was not being
1483 correctly handled. The rule now is that both the assertion and what follows
1484 in the first alternative must satisfy the test.
1485
1486 9. If auto-callout was enabled in a pattern with a conditional group whose
1487 condition was an assertion, PCRE could crash during matching, both with
1488 pcre_exec() and pcre_dfa_exec().
1489
1490 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1491 used for matching.
1492
1493 11. Unicode property support in character classes was not working for
1494 characters (bytes) greater than 127 when not in UTF-8 mode.
1495
1496 12. Added the -M command line option to pcretest.
1497
1498 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1499
1500 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1501
1502 16. Added comments and documentation about mis-use of no_arg in the C++
1503 wrapper.
1504
1505 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1506 from Martin Jerabek that uses macro names for all relevant character and
1507 string constants.
1508
1509 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1510 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1511 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1512 these, but not everybody uses configure.
1513
1514 19. A conditional group that had only one branch was not being correctly
1515 recognized as an item that could match an empty string. This meant that an
1516 enclosing group might also not be so recognized, causing infinite looping
1517 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1518 with the subject "ab", where knowledge that the repeated group can match
1519 nothing is needed in order to break the loop.
1520
1521 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1522 exec(), but without supplying a callout function, matching went wrong.
1523
1524 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1525 leak if the size of the offset vector was greater than 30. When the vector
1526 is smaller, the saved offsets during recursion go onto a local stack
1527 vector, but for larger vectors malloc() is used. It was failing to free
1528 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1529 error, in fact).
1530
1531 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1532 heapframe that is used only when UTF-8 support is enabled. This caused no
1533 problem, but was untidy.
1534
1535 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1536 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1537 included within another project.
1538
1539 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1540 slightly modified by me:
1541
1542 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1543 not building pcregrep.
1544
1545 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1546 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1547
1548 25. Forward references, both numeric and by name, in patterns that made use of
1549 duplicate group numbers, could behave incorrectly or give incorrect errors,
1550 because when scanning forward to find the reference group, PCRE was not
1551 taking into account the duplicate group numbers. A pattern such as
1552 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1553
1554 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1555 the feature of a custom pointer more persuasive (as requested by a user).
1556
1557 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1558 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1559 pre-defined.
1560
1561 28. Added support for (*UTF8) at the start of a pattern.
1562
1563 29. Arrange for flags added by the "release type" setting in CMake to be shown
1564 in the configuration summary.
1565
1566
1567 Version 7.8 05-Sep-08
1568 ---------------------
1569
1570 1. Replaced UCP searching code with optimized version as implemented for Ad
1571 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1572 stage table and inline lookup instead of a function, giving speed ups of 2
1573 to 5 times on some simple patterns that I tested. Permission was given to
1574 distribute the MultiStage2.py script that generates the tables (it's not in
1575 the tarball, but is in the Subversion repository).
1576
1577 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1578 scripts.
1579
1580 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1581 a group with a zero qualifier. The result of the study could be incorrect,
1582 or the function might crash, depending on the pattern.
1583
1584 4. Caseless matching was not working for non-ASCII characters in back
1585 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1586 It now works when Unicode Property Support is available.
1587
1588 5. In pcretest, an escape such as \x{de} in the data was always generating
1589 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1590 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1591 truncation.
1592
1593 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1594
1595 7. Added two (int) casts to pcregrep when printing the difference of two
1596 pointers, in case they are 64-bit values.
1597
1598 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1599 test 2 if it fails.
1600
1601 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1602 and a #define of that name to empty if it is not externally set. This is to
1603 allow users of MSVC to set it if necessary.
1604
1605 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1606 the convenience functions in the pcre_get.c source file.
1607
1608 11. An option change at the start of a pattern that had top-level alternatives
1609 could cause overwriting and/or a crash. This command provoked a crash in
1610 some environments:
1611
1612 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1613
1614 This potential security problem was recorded as CVE-2008-2371.
1615
1616 12. For a pattern where the match had to start at the beginning or immediately
1617 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1618 pcre_dfa_exec() could read past the end of the passed subject if there was
1619 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1620 pcretest so that it places the subject at the end of its malloc-ed buffer.
1621
1622 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1623 exec() might read past the end of the data buffer in UTF-8 mode.
1624
1625 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1626 the data contained the byte 0x85 as part of a UTF-8 character within its
1627 first line. This applied both to normal and DFA matching.
1628
1629 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1630 /^[^d]*?$/8 failed to match "abc".
1631
1632 16. Added a missing copyright notice to pcrecpp_internal.h.
1633
1634 17. Make it more clear in the documentation that values returned from
1635 pcre_exec() in ovector are byte offsets, not character counts.
1636
1637 18. Tidied a few places to stop certain compilers from issuing warnings.
1638
1639 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1640 supplied by Stefan Weber. I made a further small update for 7.8 because
1641 there is a change of source arrangements: the pcre_searchfuncs.c module is
1642 replaced by pcre_ucd.c.
1643
1644
1645 Version 7.7 07-May-08
1646 ---------------------
1647
1648 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1649 a string to a long long, pretend we don't even have a long long." This is
1650 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1651
1652 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1653 pre-7.6 versions, which defined a global no_arg variable instead of putting
1654 it in the RE class. (See also #8 below.)
1655
1656 3. Remove a line of dead code, identified by coverity and reported by Nuno
1657 Lopes.
1658
1659 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1660
1661 (1) The include/exclude patterns were being applied to the whole pathnames
1662 of files, instead of just to the final components.
1663
1664 (2) If there was more than one level of directory, the subdirectories were
1665 skipped unless they satisfied the include/exclude conditions. This is
1666 inconsistent with GNU grep (and could even be seen as contrary to the
1667 pcregrep specification - which I improved to make it absolutely clear).
1668 The action now is always to scan all levels of directory, and just
1669 apply the include/exclude patterns to regular files.
1670
1671 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1672 --exclude_dir in the tests to avoid scanning .svn directories.
1673
1674 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1675 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1676 doesn't support NULs in patterns.
1677
1678 7. Added some missing "const"s to declarations of static tables in
1679 pcre_compile.c and pcre_dfa_exec.c.
1680
1681 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1682 caused by fix #2 above. (Subsequently also a second patch to fix the
1683 first patch. And a third patch - this was a messy problem.)
1684
1685 9. Applied Craig's patch to remove the use of push_back().
1686
1687 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1688 matching function regexec().
1689
1690 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1691 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1692 references. PCRE supports relative numbers with this syntax (I don't think
1693 Oniguruma does).
1694
1695 12. Previously, a group with a zero repeat such as (...){0} was completely
1696 omitted from the compiled regex. However, this means that if the group
1697 was called as a subroutine from elsewhere in the pattern, things went wrong
1698 (an internal error was given). Such groups are now left in the compiled
1699 pattern, with a new opcode that causes them to be skipped at execution
1700 time.
1701
1702 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1703 to the way PCRE behaves:
1704
1705 (a) A lone ] character is dis-allowed (Perl treats it as data).
1706
1707 (b) A back reference to an unmatched subpattern matches an empty string
1708 (Perl fails the current match path).
1709
1710 (c) A data ] in a character class must be notated as \] because if the
1711 first data character in a class is ], it defines an empty class. (In
1712 Perl it is not possible to have an empty class.) The empty class []
1713 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1714 The negative empty class [^] matches any one character, independently
1715 of the DOTALL setting.
1716
1717 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1718 non-existent subpattern following a character class starting with ']' and
1719 containing () gave an internal compiling error instead of "reference to
1720 non-existent subpattern". Fortunately, when the pattern did exist, the
1721 compiled code was correct. (When scanning forwards to check for the
1722 existencd of the subpattern, it was treating the data ']' as terminating
1723 the class, so got the count wrong. When actually compiling, the reference
1724 was subsequently set up correctly.)
1725
1726 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1727 it was being rejected as not supported by pcre_dfa_exec(), even though
1728 other assertions are supported. I have made pcre_dfa_exec() support
1729 (*FAIL).
1730
1731 16. The implementation of 13c above involved the invention of a new opcode,
1732 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1733 cannot be changed at match time, I realized I could make a small
1734 improvement to matching performance by compiling OP_ALLANY instead of
1735 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1736 on the OP_ANY path.
1737
1738 17. Compiling pcretest on Windows with readline support failed without the
1739 following two fixes: (1) Make the unistd.h include conditional on
1740 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1741
1742 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1743 ncurses library to be included for pcretest when ReadLine support is
1744 requested, but also to allow for it to be overridden. This patch came from
1745 Daniel Bergström.
1746
1747 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1748 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1749 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1750 spotting this.
1751
1752
1753 Version 7.6 28-Jan-08
1754 ---------------------
1755
1756 1. A character class containing a very large number of characters with
1757 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1758 overflow.
1759
1760 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1761 HAVE_LONG_LONG is not defined.
1762
1763 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1764 bring them up to date and include new features. This patch includes:
1765
1766 - Fixed PH's badly added libz and libbz2 support.
1767 - Fixed a problem with static linking.
1768 - Added pcredemo. [But later removed - see 7 below.]
1769 - Fixed dftables problem and added an option.
1770 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1771 HAVE_LONG_LONG.
1772 - Added readline support for pcretest.
1773 - Added an listing of the option settings after cmake has run.
1774
1775 4. A user submitted a patch to Makefile that makes it easy to create
1776 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1777 Makefile.am that cause it to include this special target, without
1778 affecting anything else. Note that the same mingw target plus all
1779 the other distribution libraries and programs are now supported
1780 when configuring with CMake (see 6 below) instead of with
1781 Configure/Make.
1782
1783 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1784 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1785 exported in the Windows port". It has not yet been confirmed that the patch
1786 solves the problem, but it does no harm.
1787
1788 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1789 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1790 with CMake, and also correct the comment about stack recursion.
1791
1792 7. Remove the automatic building of pcredemo from the ./configure system and
1793 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1794 of a program that users should build themselves after PCRE is installed, so
1795 building it automatically is not really right. What is more, it gave
1796 trouble in some build environments.
1797
1798 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1799
1800
1801 Version 7.5 10-Jan-08
1802 ---------------------
1803
1804 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1805 values in parens when parsing an RE using the C++ wrapper."
1806
1807 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1808 Characters greater than 255 were excluded from the class instead of being
1809 included.
1810
1811 3. The same bug as (2) above applied to negated POSIX classes such as
1812 [:^space:].
1813
1814 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1815 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1816 I have changed it.
1817
1818 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1819 first named subpattern) and a construct such as (?&a) would reference the
1820 first named subpattern whose name started with "a" (in other words, the
1821 length check was missing). Both these problems are fixed. "Subpattern name
1822 expected" is now given for (?&) (a zero-length name), and this patch also
1823 makes it give the same error for \k'' (previously it complained that that
1824 was a reference to a non-existent subpattern).
1825
1826 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1827 this is right because (?- can be followed by option settings as well as by
1828 digits. I have, however, made the messages clearer.
1829
1830 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1831 than the number used in the conditional) now cause a compile-time error.
1832 This is actually not compatible with Perl, which accepts such patterns, but
1833 treats the conditional as always being FALSE (as PCRE used to), but it
1834 seems to me that giving a diagnostic is better.
1835
1836 8. Change "alphameric" to the more common word "alphanumeric" in comments
1837 and messages.
1838
1839 9. Fix two occurrences of "backslash" in comments that should have been
1840 "backspace".
1841
1842 10. Remove two redundant lines of code that can never be obeyed (their function
1843 was moved elsewhere).
1844
1845 11. The program that makes PCRE's Unicode character property table had a bug
1846 which caused it to generate incorrect table entries for sequences of
1847 characters that have the same character type, but are in different scripts.
1848 It amalgamated them into a single range, with the script of the first of
1849 them. In other words, some characters were in the wrong script. There were
1850 thirteen such cases, affecting characters in the following ranges:
1851
1852 U+002b0 - U+002c1
1853 U+0060c - U+0060d
1854 U+0061e - U+00612
1855 U+0064b - U+0065e
1856 U+0074d - U+0076d
1857 U+01800 - U+01805
1858 U+01d00 - U+01d77
1859 U+01d9b - U+01dbf
1860 U+0200b - U+0200f
1861 U+030fc - U+030fe
1862 U+03260 - U+0327f
1863 U+0fb46 - U+0fbb1
1864 U+10450 - U+1049d
1865
1866 12. The -o option (show only the matching part of a line) for pcregrep was not
1867 compatible with GNU grep in that, if there was more than one match in a
1868 line, it showed only the first of them. It now behaves in the same way as
1869 GNU grep.
1870
1871 13. If the -o and -v options were combined for pcregrep, it printed a blank
1872 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1873 does the same. The return code can be used to tell if there were any
1874 non-matching lines.
1875
1876 14. Added --file-offsets and --line-offsets to pcregrep.
1877
1878 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1879 infinitely looping recursion. The bug was that positive lookaheads were not
1880 being skipped when checking for a possible empty match (negative lookaheads
1881 and both kinds of lookbehind were skipped).
1882
1883 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1884 inclusion of <windows.h> to before rather than after the definition of
1885 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1886
1887 17. Specifying a possessive quantifier with a specific limit for a Unicode
1888 character property caused pcre_compile() to compile bad code, which led at
1889 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1890 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1891 caused the error; without that there was no problem.
1892
1893 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1894
1895 19. Added --enable-pcretest-libreadline.
1896
1897 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1898 RE::GlobalReplace(). As a result, the number of replacements returned was
1899 double what it should be. I removed one of the increments, but Craig sent a
1900 later patch that removed the other one (the right fix) and added unit tests
1901 that check the return values (which was not done before).
1902
1903 21. Several CMake things:
1904
1905 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1906 the names libpcre and libpcreposix, not just pcre and pcreposix.
1907
1908 (2) The above change means that pcretest and pcregrep are now correctly
1909 linked with the newly-built libraries, not previously installed ones.
1910
1911 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1912
1913 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1914 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1915 UTF-8 newline character). The key issue is that the pattern starts .*;
1916 this means that the match must be either at the beginning, or after a
1917 newline. The bug was in the code for advancing after a failed match and
1918 checking that the new position followed a newline. It was not taking
1919 account of UTF-8 characters correctly.
1920
1921 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1922 character classes. PCRE was not treating the sequence [:...:] as a
1923 character class unless the ... were all letters. Perl, however, seems to
1924 allow any characters between [: and :], though of course it rejects as
1925 unknown any "names" that contain non-letters, because all the known class
1926 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1927 for example, whereas PCRE did not - it did not recognize a POSIX character
1928 class. This seemed a bit dangerous, so the code has been changed to be
1929 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1930 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1931 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1932 Perl does, and where it didn't before.
1933
1934 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1935 Windows environments %n is disabled by default.
1936
1937
1938 Version 7.4 21-Sep-07
1939 ---------------------
1940
1941 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1942 means that a class such as [\s] counted as "explicit reference to CR or
1943 LF". That isn't really right - the whole point of the change was to try to
1944 help when there was an actual mention of one of the two characters. So now
1945 the change happens only if \r or \n (or a literal CR or LF) character is
1946 encountered.
1947
1948 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1949 of both had grown to the point where there were only 3 bits left.
1950 Fortunately, there was spare space in the data structure, and so I have
1951 moved the internal flags into a new 16-bit field to free up more option
1952 bits.
1953
1954 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1955 but did not set the internal JCHANGED flag - either of these is enough to
1956 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1957 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1958 start sets both bits.
1959
1960 4. Added options (at build time, compile time, exec time) to change \R from
1961 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1962
1963 5. doc/pcresyntax.html was missing from the distribution.
1964
1965 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1966 compatibility, even though it is no longer used.
1967
1968 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1969 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1970 windows.h file is present (where different names are used). [This was
1971 reversed later after testing - see 16 below.]
1972
1973 8. Changed all #include <config.h> to #include "config.h". There were also
1974 some further <pcre.h> cases that I changed to "pcre.h".
1975
1976 9. When pcregrep was used with the --colour option, it missed the line ending
1977 sequence off the lines that it output.
1978
1979 10. It was pointed out to me that arrays of string pointers cause lots of
1980 relocations when a shared library is dynamically loaded. A technique of
1981 using a single long string with a table of offsets can drastically reduce
1982 these. I have refactored PCRE in four places to do this. The result is
1983 dramatic:
1984
1985 Originally: 290
1986 After changing UCP table: 187
1987 After changing error message table: 43
1988 After changing table of "verbs" 36
1989 After changing table of Posix names 22
1990
1991 Thanks to the folks working on Gregex for glib for this insight.
1992
1993 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1994 unicode-properties was also set.
1995
1996 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1997
1998 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1999 checked only for CRLF.
2000
2001 14. Added casts to pcretest.c to avoid compiler warnings.
2002
2003 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2004
2005 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2006 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2007 entirely. This removes changes made in 7 above.
2008
2009 17. The CMake files have been updated, and there is now more information about
2010 building with CMake in the NON-UNIX-USE document.
2011
2012
2013 Version 7.3 28-Aug-07
2014 ---------------------
2015
2016 1. In the rejigging of the build system that eventually resulted in 7.1, the
2017 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2018 brackets there is not right, since it causes compilers to look for an
2019 installed pcre.h, not the version that is in the source that is being
2020 compiled (which of course may be different). I have changed it back to:
2021
2022 #include "pcre.h"
2023
2024 I have a vague recollection that the change was concerned with compiling in
2025 different directories, but in the new build system, that is taken care of
2026 by the VPATH setting the Makefile.
2027
2028 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2029 when the subject happened to end in the byte 0x85 (e.g. if the last
2030 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2031 characters but of course it shouldn't be taken as a newline when it is part
2032 of another character. The bug was that, for an unlimited repeat of . in
2033 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2034 characters when looking for a newline.
2035
2036 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2037
2038 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2039 in debug output.
2040
2041 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2042 long printing in the pcrecpp unittest when running under MinGW.
2043
2044 6. ESC_K was left out of the EBCDIC table.
2045
2046 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2047 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2048 limit also applies to "virtual nesting" when a pattern is recursive, and in
2049 this case 1000 isn't so big. I have been able to remove this limit at the
2050 expense of backing off one optimization in certain circumstances. Normally,
2051 when pcre_exec() would call its internal match() function recursively and
2052 immediately return the result unconditionally, it uses a "tail recursion"
2053 feature to save stack. However, when a subpattern that can match an empty
2054 string has an unlimited repetition quantifier, it no longer makes this
2055 optimization. That gives it a stack frame in which to save the data for
2056 checking that an empty string has been matched. Previously this was taken
2057 from the 1000-entry workspace that had been reserved. So now there is no
2058 explicit limit, but more stack is used.
2059
2060 8. Applied Daniel's patches to solve problems with the import/export magic
2061 syntax that is required for Windows, and which was going wrong for the
2062 pcreposix and pcrecpp parts of the library. These were overlooked when this
2063 problem was solved for the main library.
2064
2065 9. There were some crude static tests to avoid integer overflow when computing
2066 the size of patterns that contain repeated groups with explicit upper
2067 limits. As the maximum quantifier is 65535, the maximum group length was
2068 set at 30,000 so that the product of these two numbers did not overflow a
2069 32-bit integer. However, it turns out that people want to use groups that
2070 are longer than 30,000 bytes (though not repeat them that many times).
2071 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2072 made it possible to implement the integer overflow checks in a much more
2073 dynamic way, which I have now done. The artificial limitation on group
2074 length has been removed - we now have only the limit on the total length of
2075 the compiled pattern, which depends on the LINK_SIZE setting.
2076
2077 10. Fixed a bug in the documentation for get/copy named substring when
2078 duplicate names are permitted. If none of the named substrings are set, the
2079 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2080 empty string.
2081
2082 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2083 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2084 because the ] is interpreted as the first data character and the
2085 terminating ] is not found. PCRE has been made compatible with Perl in this
2086 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2087 cause memory overwriting.
2088
2089 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2090 string has been matched (to stop an infinite loop). It was not recognizing
2091 a conditional subpattern that could match an empty string if that
2092 subpattern was within another subpattern. For example, it looped when
2093 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2094 condition was not nested. This bug has been fixed.
2095
2096 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2097 past the start of the subject in the presence of bytes with the top bit
2098 set, for example "\x8aBCD".
2099
2100 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2101 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2102
2103 14. Optimized (?!) to (*FAIL).
2104
2105 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2106 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2107 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2108 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2109 does: it's just the validity check that is more restrictive.
2110
2111 16. Inserted checks for integer overflows during escape sequence (backslash)
2112 processing, and also fixed erroneous offset values for syntax errors during
2113 backslash processing.
2114
2115 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2116 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2117
2118 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2119 caused an overrun.
2120
2121 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2122 something other than just ASCII characters) inside a group that had an
2123 unlimited repeat caused a loop at compile time (while checking to see
2124 whether the group could match an empty string).
2125
2126 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2127 [\P{Any}] did so. (Error in the code for printing property names.)
2128
2129 21. An orphan \E inside a character class could cause a crash.
2130
2131 22. A repeated capturing bracket such as (A)? could cause a wild memory
2132 reference during compilation.
2133
2134 23. There are several functions in pcre_compile() that scan along a compiled
2135 expression for various reasons (e.g. to see if it's fixed length for look
2136 behind). There were bugs in these functions when a repeated \p or \P was
2137 present in the pattern. These operators have additional parameters compared
2138 with \d, etc, and these were not being taken into account when moving along
2139 the compiled data. Specifically:
2140
2141 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2142 length.
2143
2144 (b) An item such as \pL+ within a repeated group could cause crashes or
2145 loops.
2146
2147 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2148 "reference to non-existent subpattern" error.
2149
2150 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2151
2152 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2153 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2154
2155 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2156
2157 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2158 character were causing crashes (broken optimization).
2159
2160 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2161 \p or \P) caused a compile-time loop.
2162
2163 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2164 break. For example, the unstudied pattern [\r\n]A does not match the string
2165 "\r\nA" because change 7.0/46 below moves the current point on by two
2166 characters after failing to match at the start. However, the pattern \nA
2167 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2168 the same is true. There doesn't seem any very clean way out of this, but
2169 what I have chosen to do makes the common cases work: PCRE now takes note
2170 of whether there can be an explicit match for \r or \n anywhere in the
2171 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2172 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2173 pattern has explicit CR or LF references.
2174
2175 29. Added (*CR) etc for changing newline setting at start of pattern.
2176
2177
2178 Version 7.2 19-Jun-07
2179 ---------------------
2180
2181 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2182 which is apparently normally available under Windows.
2183
2184 2. Re-jig the pcregrep tests with different newline settings in an attempt
2185 to make them independent of the local environment's newline setting.
2186
2187 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2188
2189 4. Some of the "internals" tests were previously cut out when the link size
2190 was not 2, because the output contained actual offsets. The recent new
2191 "Z" feature of pcretest means that these can be cut out, making the tests
2192 usable with all link sizes.
2193
2194 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2195 stack recursion. This gives a massive performance boost under BSD, but just
2196 a small improvement under Linux. However, it saves one field in the frame
2197 in all cases.
2198
2199 6. Added more features from the forthcoming Perl 5.10:
2200
2201 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2202 recursion call. It refers to the nth most recently opened parentheses.
2203
2204 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2205 to be opened parentheses.
2206
2207 (c) Conditions that refer to capturing parentheses can be specified
2208 relatively, for example, (?(-2)... or (?(+3)...
2209
2210 (d) \K resets the start of the current match so that everything before
2211 is not part of it.
2212
2213 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2214
2215 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2216 reference syntax.
2217
2218 (g) (?| introduces a group in which the numbering of parentheses in each
2219 alternative starts with the same number.
2220
2221 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2222
2223 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2224 PCRE_INFO_JCHANGED.
2225
2226 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2227 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2228 for detecting groups that can match an empty string.
2229
2230 9. A pattern with a very large number of alternatives (more than several
2231 hundred) was running out of internal workspace during the pre-compile
2232 phase, where pcre_compile() figures out how much memory will be needed. A
2233 bit of new cunning has reduced the workspace needed for groups with
2234 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2235 workspace instead of running out of the 4096 that are available.
2236
2237 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2238
2239 11. Applied patch from Google to remove an optimization that didn't quite work.
2240 The report of the bug said:
2241
2242 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2243 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2244 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2245
2246 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2247 it matched the wrong number of bytes.
2248
2249
2250 Version 7.1 24-Apr-07
2251 ---------------------
2252
2253 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2254 that is more "standard", making use of automake and other Autotools. There
2255 is some re-arrangement of the files and adjustment of comments consequent
2256 on this.
2257
2258 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2259 for recursive directory scanning broke on some systems because the files
2260 are not scanned in any specific order and on different systems the order
2261 was different. A call to "sort" has been inserted into RunGrepTest for the
2262 approprate test as a short-term fix. In the longer term there may be an
2263 alternative.
2264
2265 3. I had an email from Eric Raymond about problems translating some of PCRE's
2266 man pages to HTML (despite the fact that I distribute HTML pages, some
2267 people do their own conversions for various reasons). The problems
2268 concerned the use of low-level troff macros .br and .in. I have therefore
2269 removed all such uses from the man pages (some were redundant, some could
2270 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2271 HTML has been updated to handle .nf/.fi and to complain if it encounters
2272 .br or .in.
2273
2274 4. Updated comments in configure.ac that get placed in config.h.in and also
2275 arranged for config.h to be included in the distribution, with the name
2276 config.h.generic, for the benefit of those who have to compile without
2277 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2278
2279 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2280 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2281 makevp.bat for the current PCRE, using the additional files
2282 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2283
2284 6. A Windows user reported a minor discrepancy with test 2, which turned out
2285 to be caused by a trailing space on an input line that had got lost in his
2286 copy. The trailing space was an accident, so I've just removed it.
2287
2288 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2289 that is needed.
2290
2291 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2292 as "const" (a) because they are and (b) because it helps the PHP
2293 maintainers who have recently made a script to detect big data structures
2294 in the php code that should be moved to the .rodata section. I remembered
2295 to update Builducptable as well, so it won't revert if ucptable.h is ever
2296 re-created.
2297
2298 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2299 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2300 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2301 support is not required. This saves 1.5-2K of code, which is important in
2302 some applications.
2303
2304 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2305 so as not to refer to the tables, even though these functions will never be
2306 called when UTF-8 support is disabled. Otherwise there are problems with a
2307 shared library.
2308
2309 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2310
2311 (a) It was defining its arguments as char * instead of void *.
2312
2313 (b) It was assuming that all moves were upwards in memory; this was true
2314 a long time ago when I wrote it, but is no longer the case.
2315
2316 The emulated memove() is provided for those environments that have neither
2317 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2318 is clearly not the case, as these two bugs were recently reported.
2319
2320 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2321 and Detrail to create the HTML documentation, the .txt form of the man
2322 pages, and it removes trailing spaces from listed files. It also creates
2323 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2324 case, it wraps all the #defines with #ifndefs. This script should be run
2325 before "make dist".
2326
2327 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2328 with Unicode property support.
2329
2330 (a) For a maximizing quantifier, if the two different cases of the
2331 character were of different lengths in their UTF-8 codings (there are
2332 some cases like this - I found 11), and the matching function had to
2333 back up over a mixture of the two cases, it incorrectly assumed they
2334 were both the same length.
2335
2336 (b) When PCRE was configured to use the heap rather than the stack for
2337 recursion during matching, it was not correctly preserving the data for
2338 the other case of a UTF-8 character when checking ahead for a match
2339 while processing a minimizing repeat. If the check also involved
2340 matching a wide character, but failed, corruption could cause an
2341 erroneous result when trying to check for a repeat of the original
2342 character.
2343
2344 13. Some tidying changes to the testing mechanism:
2345
2346 (a) The RunTest script now detects the internal link size and whether there
2347 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2348 values substituted by "configure". (The RunGrepTest script already did
2349 this for UTF-8.) The configure.ac script no longer substitutes the
2350 relevant variables.
2351
2352 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2353 with length and offset values. This means that the output is different
2354 for different internal link sizes. Test 2 is skipped for link sizes
2355 other than 2 because of this, bypassing the problem. Unfortunately,
2356 there was also a test in test 3 (the locale tests) that used /B and
2357 failed for link sizes other than 2. Rather than cut the whole test out,
2358 I have added a new /Z option to pcretest that replaces the length and
2359 offset values with spaces. This is now used to make test 3 independent
2360 of link size. (Test 2 will be tidied up later.)
2361
2362 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2363 segmentation fault instead of returning the appropriate error message.
2364
2365 15. In multiline mode when the newline sequence was set to "any", the pattern
2366 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2367 This doesn't seem right; it now treats the CRLF combination as the line
2368 ending, and so does not match in that case. It's only a pattern such as ^$
2369 that would hit this one: something like ^ABC$ would have failed after \r
2370 and then tried again after \r\n.
2371
2372 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2373 in an attempt to make files that differ only in their line terminators
2374 compare equal. This works on Linux.
2375
2376 17. Under certain error circumstances pcregrep might try to free random memory
2377 as it exited. This is now fixed, thanks to valgrind.
2378
2379 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2380 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2381 was because its rules for how to advance for /g after matching an empty
2382 string at the end of a line did not allow for this case. They now check for
2383 it specially.
2384
2385 20. pcretest is supposed to handle patterns and data of any length, by
2386 extending its buffers when necessary. It was getting this wrong when the
2387 buffer for a data line had to be extended.
2388
2389 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2390 CRLF as a newline sequence.
2391
2392 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2393 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2394 I have nevertheless tidied it up.
2395
2396 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2397
2398 24. Added a man page for pcre-config.
2399
2400
2401 Version 7.0 19-Dec-06
2402 ---------------------
2403
2404 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2405 moving to gcc 4.1.1.
2406
2407 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2408 sys/time.h, which is documented as needed for this function. It doesn't
2409 seem to matter on Linux, but it showed up on some releases of OS X.
2410
2411 3. It seems that there are systems where bytes whose values are greater than
2412 127 match isprint() in the "C" locale. The "C" locale should be the
2413 default when a C program starts up. In most systems, only ASCII printing
2414 characters match isprint(). This difference caused the output from pcretest
2415 to vary, making some of the tests fail. I have changed pcretest so that:
2416
2417 (a) When it is outputting text in the compiled version of a pattern, bytes
2418 other than 32-126 are always shown as hex escapes.
2419
2420 (b) When it is outputting text that is a matched part of a subject string,
2421 it does the same, unless a different locale has been set for the match
2422 (using the /L modifier). In this case, it uses isprint() to decide.
2423
2424 4. Fixed a major bug that caused incorrect computation of the amount of memory
2425 required for a compiled pattern when options that changed within the
2426 pattern affected the logic of the preliminary scan that determines the
2427 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2428 that the computed length was too small. The symptoms of this bug were
2429 either the PCRE error "internal error: code overflow" from pcre_compile(),
2430 or a glibc crash with a message such as "pcretest: free(): invalid next
2431 size (fast)". Examples of patterns that provoked this bug (shown in
2432 pcretest format) are:
2433
2434 /(?-x: )/x
2435 /(?x)(?-x: \s*#\s*)/
2436 /((?i)[\x{c0}])/8
2437 /(?i:[\x{c0}])/8
2438
2439 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2440 is now done differently.
2441
2442 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2443 wrapper classes; (b) implement a new function in the C++ scanner that is
2444 more efficient than the old way of doing things because it avoids levels of
2445 recursion in the regex matching; (c) add a paragraph to the documentation
2446 for the FullMatch() function.
2447
2448 6. The escape sequence \n was being treated as whatever was defined as
2449 "newline". Not only was this contrary to the documentation, which states
2450 that \n is character 10 (hex 0A), but it also went horribly wrong when
2451 "newline" was defined as CRLF. This has been fixed.
2452
2453 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2454 was being set to -1 for the "end of line" case (supposedly a value that no
2455 character can have). Though this value is never used (the check for end of
2456 line is "zero bytes in current character"), it caused compiler complaints.
2457 I've changed it to 0xffffffff.
2458
2459 8. In pcre_version.c, the version string was being built by a sequence of
2460 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2461 string (as it is for production releases) called a macro with an empty
2462 argument. The C standard says the result of this is undefined. The gcc
2463 compiler treats it as an empty string (which was what was wanted) but it is
2464 reported that Visual C gives an error. The source has been hacked around to
2465 avoid this problem.
2466
2467 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2468 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2469 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2470 of them did).
2471
2472 10. Originally, pcretest opened its input and output without "b"; then I was
2473 told that "b" was needed in some environments, so it was added for release
2474 5.0 to both the input and output. (It makes no difference on Unix-like
2475 systems.) Later I was told that it is wrong for the input on Windows. I've
2476 now abstracted the modes into two macros, to make it easier to fiddle with
2477 them, and removed "b" from the input mode under Windows.
2478
2479 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2480
2481 12. Added -help and --help to pcretest as an official way of being reminded
2482 of the options.
2483
2484 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2485 and pcrecpp.cc because they annoy compilers at high warning levels.
2486
2487 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2488
2489 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2490 scripts are not C programs :-) and which was not noticed because it works
2491 on Linux.
2492
2493 16. pcretest is supposed to handle any length of pattern and data line (as one
2494 line or as a continued sequence of lines) by extending its input buffer if
2495 necessary. This feature was broken for very long pattern lines, leading to
2496 a string of junk being passed to pcre_compile() if the pattern was longer
2497 than about 50K.
2498
2499 17. I have done a major re-factoring of the way pcre_compile() computes the
2500 amount of memory needed for a compiled pattern. Previously, there was code
2501 that made a preliminary scan of the pattern in order to do this. That was
2502 OK when PCRE was new, but as the facilities have expanded, it has become
2503 harder and harder to keep it in step with the real compile phase, and there
2504 have been a number of bugs (see for example, 4 above). I have now found a
2505 cunning way of running the real compile function in a "fake" mode that
2506 enables it to compute how much memory it would need, while actually only
2507 ever using a few hundred bytes of working memory and without too many
2508 tests of the mode. This should make future maintenance and development
2509 easier. A side effect of this work is that the limit of 200 on the nesting
2510 depth of parentheses has been removed (though this was never a serious
2511 limitation, I suspect). However, there is a downside: pcre_compile() now
2512 runs more slowly than before (30% or more, depending on the pattern). I
2513 hope this isn't a big issue. There is no effect on runtime performance.
2514
2515 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2516 newline (only possible for the last line of a file) and it was a
2517 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2518
2519 19. Added additional timing features to pcretest. (1) The -tm option now times
2520 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2521 separate command line item, by a number that specifies the number of
2522 repeats to use when timing. The default is 50000; this gives better
2523 precision, but takes uncomfortably long for very large patterns.
2524
2525 20. Extended pcre_study() to be more clever in cases where a branch of a
2526 subpattern has no definite first character. For example, (a*|b*)[cd] would
2527 previously give no result from pcre_study(). Now it recognizes that the
2528 first character must be a, b, c, or d.
2529
2530 21. There was an incorrect error "recursive call could loop indefinitely" if
2531 a subpattern (or the entire pattern) that was being tested for matching an
2532 empty string contained only one non-empty item after a nested subpattern.
2533 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2534 incorrectly, because the \d was being skipped in the check.
2535
2536 22. The pcretest program now has a new pattern option /B and a command line
2537 option -b, which is equivalent to adding /B to every pattern. This causes
2538 it to show the compiled bytecode, without the additional information that
2539 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2540 is the same as /B/I).
2541
2542 23. A new optimization is now able automatically to treat some sequences such
2543 as a*b as a*+b. More specifically, if something simple (such as a character
2544 or a simple class like \d) has an unlimited quantifier, and is followed by
2545 something that cannot possibly match the quantified thing, the quantifier
2546 is automatically "possessified".
2547
2548 24. A recursive reference to a subpattern whose number was greater than 39
2549 went wrong under certain circumstances in UTF-8 mode. This bug could also
2550 have affected the operation of pcre_study().
2551
2552 25. Realized that a little bit of performance could be had by replacing
2553 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2554
2555 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2556
2557 27. Possessive quantifiers such as a++ were previously implemented by turning
2558 them into atomic groups such as ($>a+). Now they have their own opcodes,
2559 which improves performance. This includes the automatically created ones
2560 from 23 above.
2561
2562 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2563 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2564 the first matched character to be a colon. This applied both to named and
2565 numbered groups.
2566
2567 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2568
2569 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2570 building PCRE on Windows easier, so I have included it in the distribution.
2571
2572 31. There is now a check in pcretest against a ridiculously large number being
2573 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2574 loop, the loop is abandoned.
2575
2576 32. Forward references to subpatterns in conditions such as (?(2)...) where
2577 subpattern 2 is defined later cause pcre_compile() to search forwards in
2578 the pattern for the relevant set of parentheses. This search went wrong
2579 when there were unescaped parentheses in a character class, parentheses
2580 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2581
2582 33. "Subroutine" calls and backreferences were previously restricted to
2583 referencing subpatterns earlier in the regex. This restriction has now
2584 been removed.
2585
2586 34. Added a number of extra features that are going to be in Perl 5.10. On the
2587 whole, these are just syntactic alternatives for features that PCRE had
2588 previously implemented using the Python syntax or my own invention. The
2589 other formats are all retained for compatibility.
2590
2591 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2592 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2593 also .NET compatible.
2594
2595 (b) A recursion or subroutine call to a named group can now be defined as
2596 (?&name) as well as (?P>name).
2597
2598 (c) A backreference to a named group can now be defined as \k<name> or
2599 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2600 5.10, are also .NET compatible.
2601
2602 (d) A conditional reference to a named group can now use the syntax
2603 (?(<name>) or (?('name') as well as (?(name).
2604
2605 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2606 groups (named and numbered) that are never evaluated inline, but can be
2607 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2608 is always false. There may be only one alternative in such a group.
2609
2610 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2611 as the simple (?(R). The condition is true only if the most recent
2612 recursion is that of the given number or name. It does not search out
2613 through the entire recursion stack.
2614
2615 (g) The escape \gN or \g{N} has been added, where N is a positive or
2616 negative number, specifying an absolute or relative reference.
2617
2618 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2619 some "unreachable code" warnings.
2620
2621 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2622 things, this adds five new scripts.
2623
2624 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2625 There were also incompatibilities regarding the handling of \Q..\E inside
2626 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2627 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2628
2629 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2630 matches an empty string, and forcibly breaks the loop. There were bugs in
2631 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2632 against aaaa the result was just "a" rather than "aaaa", for example. Two
2633 separate and independent bugs (that affected different cases) have been
2634 fixed.
2635
2636 39. Refactored the code to abolish the use of different opcodes for small
2637 capturing bracket numbers. This is a tidy that I avoided doing when I
2638 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2639 The new approach is not only tidier, it makes it possible to reduce the
2640 memory needed to fix the previous bug (38).
2641
2642 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2643 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2644 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2645 mode.
2646
2647 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2648 report.
2649
2650 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2651 copy construction and assignment in the C++ wrapper.
2652
2653 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2654 couple of bugs that could have given wrong results in the "--newline=crlf"
2655 case.
2656
2657 44. Added a number of casts and did some reorganization of signed/unsigned int
2658 variables following suggestions from Dair Grant. Also renamed the variable
2659 "this" as "item" because it is a C++ keyword.
2660
2661 45. Arranged for dftables to add
2662
2663 #include "pcre_internal.h"
2664
2665 to pcre_chartables.c because without it, gcc 4.x may remove the array
2666 definition from the final binary if PCRE is built into a static library and
2667 dead code stripping is activated.
2668
2669 46. For an unanchored pattern, if a match attempt fails at the start of a
2670 newline sequence, and the newline setting is CRLF or ANY, and the next two
2671 characters are CRLF, advance by two characters instead of one.
2672
2673
2674 Version 6.7 04-Jul-06
2675 ---------------------
2676
2677 1. In order to handle tests when input lines are enormously long, pcretest has
2678 been re-factored so that it automatically extends its buffers when
2679 necessary. The code is crude, but this _is_ just a test program. The
2680 default size has been increased from 32K to 50K.
2681
2682 2. The code in pcre_study() was using the value of the re argument before
2683 testing it for NULL. (Of course, in any sensible call of the function, it
2684 won't be NULL.)
2685
2686 3. The memmove() emulation function in pcre_internal.h, which is used on
2687 systems that lack both memmove() and bcopy() - that is, hardly ever -
2688 was missing a "static" storage class specifier.
2689
2690 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2691 containing an extended class (one that cannot be represented by a bitmap
2692 because it contains high-valued characters or Unicode property items, e.g.
2693 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2694 pattern, but PCRE should not loop if you do not (it no longer does).
2695 [Detail: two cases were found: (a) a repeated subpattern containing an
2696 extended class; (b) a recursive reference to a subpattern that followed a
2697 previous extended class. It wasn't skipping over the extended class
2698 correctly when UTF-8 mode was not set.]
2699
2700 5. A negated single-character class was not being recognized as fixed-length
2701 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2702 compile error "lookbehind assertion is not fixed length".
2703
2704 6. The RunPerlTest auxiliary script was showing an unexpected difference
2705 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2706 write a Perl script that can interpret lines of an input file either as
2707 byte characters or as UTF-8, which is what "perltest" was being required to
2708 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2709 can't do is switch easily at run time between having the "use utf8;" pragma
2710 or not. In the end, I fudged it by using the RunPerlTest script to insert
2711 "use utf8;" explicitly for the UTF-8 tests.
2712
2713 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2714 the end of the subject string, contrary to the documentation and to what
2715 Perl does. This was true of both matching functions. Now it matches only at
2716 the start of the subject and immediately after *internal* newlines.
2717
2718 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2719 a pointer to an int instead of a pointer to an unsigned long int. This
2720 caused problems on 64-bit systems.
2721
2722 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2723 instance of the 'standard' template library not being so standard".
2724
2725 10. There was no check on the number of named subpatterns nor the maximum
2726 length of a subpattern name. The product of these values is used to compute
2727 the size of the memory block for a compiled pattern. By supplying a very
2728 long subpattern name and a large number of named subpatterns, the size
2729 computation could be caused to overflow. This is now prevented by limiting
2730 the length of names to 32 characters, and the number of named subpatterns
2731 to 10,000.
2732
2733 11. Subpatterns that are repeated with specific counts have to be replicated in
2734 the compiled pattern. The size of memory for this was computed from the
2735 length of the subpattern and the repeat count. The latter is limited to
2736 65535, but there was no limit on the former, meaning that integer overflow
2737 could in principle occur. The compiled length of a repeated subpattern is
2738 now limited to 30,000 bytes in order to prevent this.
2739
2740 12. Added the optional facility to have named substrings with the same name.
2741
2742 13. Added the ability to use a named substring as a condition, using the
2743 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2744 are numbers (not recommended). Forward references are permitted.
2745
2746 14. Added forward references in named backreferences (if you see what I mean).
2747
2748 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2749 pattern could run off the end of the subject. For example, the pattern
2750 "(?s)(.{1,5})"8 did this with the subject "ab".
2751
2752 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2753 PCRE_CASELESS was set when matching characters that were quantified with ?
2754 or *.
2755
2756 17. A character class other than a single negated character that had a minimum
2757 but no maximum quantifier - for example [ab]{6,} - was not handled
2758 correctly by pce_dfa_exec(). It would match only one character.
2759
2760 18. A valid (though odd) pattern that looked like a POSIX character
2761 class but used an invalid character after [ (for example [[,abc,]]) caused
2762 pcre_compile() to give the error "Failed: internal error: code overflow" or
2763 in some cases to crash with a glibc free() error. This could even happen if
2764 the pattern terminated after [[ but there just happened to be a sequence of
2765 letters, a binary zero, and a closing ] in the memory that followed.
2766
2767 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2768 over the years. Originally (before any Unicode support), just the bottom 8
2769 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2770 output from "man perlunicode" includes this:
2771
2772 The regular expression compiler produces polymorphic opcodes. That
2773 is, the pattern adapts to the data and automatically switches to
2774 the Unicode character scheme when presented with Unicode data--or
2775 instead uses a traditional byte scheme when presented with byte
2776 data.
2777
2778 Sadly, a wide octal escape does not cause a switch, and in a string with
2779 no other multibyte characters, these octal escapes are treated as before.
2780 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2781 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2782 Unicode string.
2783
2784 I have not perpetrated such confusion in PCRE. Up till now, it took just
2785 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2786 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2787 translate to the appropriate multibyte character.
2788
2789 29. Applied some refactoring to reduce the number of warnings from Microsoft
2790 and Borland compilers. This has included removing the fudge introduced
2791 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2792 a warning about an unused variable.
2793
2794 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2795 characters since release 4.0, because Perl (from release 5.004) does not.
2796 [Or at least, is documented not to: some releases seem to be in conflict
2797 with the documentation.] However, when a pattern was studied with
2798 pcre_study() and all its branches started with \s, PCRE still included VT
2799 as a possible starting character. Of course, this did no harm; it just
2800 caused an unnecessary match attempt.
2801
2802 22. Removed a now-redundant internal flag bit that recorded the fact that case
2803 dependency changed within the pattern. This was once needed for "required
2804 byte" processing, but is no longer used. This recovers a now-scarce options
2805 bit. Also moved the least significant internal flag bit to the most-
2806 significant bit of the word, which was not previously used (hangover from
2807 the days when it was an int rather than a uint) to free up another bit for
2808 the future.
2809
2810 23. Added support for CRLF line endings as well as CR and LF. As well as the
2811 default being selectable at build time, it can now be changed at runtime
2812 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2813 specify that it is scanning data with non-default line endings.
2814
2815 24. Changed the definition of CXXLINK to make it agree with the definition of
2816 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2817
2818 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2819 recursions. This makes a big different to stack usage for some patterns.
2820
2821 26. If a subpattern containing a named recursion or subroutine reference such
2822 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2823 the space required for the compiled pattern went wrong and gave too small a
2824 value. Depending on the environment, this could lead to "Failed: internal
2825 error: code overflow at offset 49" or "glibc detected double free or
2826 corruption" errors.
2827
2828 27. Applied patches from Google (a) to support the new newline modes and (b) to
2829 advance over multibyte UTF-8 characters in GlobalReplace.
2830
2831 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2832 difference for some implementation of PCRE in some Windows version.
2833
2834 29. Added some extra testing facilities to pcretest:
2835
2836 \q<number> in a data line sets the "match limit" value
2837 \Q<number> in a data line sets the "match recursion limt" value
2838 -S <number> sets the stack size, where <number> is in megabytes
2839
2840 The -S option isn't available for Windows.
2841
2842
2843 Version 6.6 06-Feb-06
2844 ---------------------
2845
2846 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2847 in pcreposix.h. I have copied the definition from pcre.h.
2848
2849 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2850 because pcre.h is no longer a built file.
2851
2852 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2853 not normally included in the compiled code.
2854
2855
2856 Version 6.5 01-Feb-06
2857 ---------------------
2858
2859 1. When using the partial match feature with pcre_dfa_exec(), it was not
2860 anchoring the second and subsequent partial matches at the new starting
2861 point. This could lead to incorrect results. For example, with the pattern
2862 /1234/, partially matching against "123" and then "a4" gave a match.
2863
2864 2. Changes to pcregrep:
2865
2866 (a) All non-match returns from pcre_exec() were being treated as failures
2867 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2868 error message is output. Some extra information is given for the
2869 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2870 probably the only errors that are likely to be caused by users (by
2871 specifying a regex that has nested indefinite repeats, for instance).
2872 If there are more than 20 of these errors, pcregrep is abandoned.
2873
2874 (b) A binary zero was treated as data while matching, but terminated the
2875 output line if it was written out. This has been fixed: binary zeroes
2876 are now no different to any other data bytes.
2877
2878 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2879 used to set a locale for matching. The --locale=xxxx long option has
2880 been added (no short equivalent) to specify a locale explicitly on the
2881 pcregrep command, overriding the environment variables.
2882
2883 (d) When -B was used with -n, some line numbers in the output were one less
2884 than they should have been.
2885
2886 (e) Added the -o (--only-matching) option.
2887
2888 (f) If -A or -C was used with -c (count only), some lines of context were
2889 accidentally printed for the final match.
2890
2891 (g) Added the -H (--with-filename) option.
2892
2893 (h) The combination of options -rh failed to suppress file names for files
2894 that were found from directory arguments.
2895
2896 (i) Added the -D (--devices) and -d (--directories) options.
2897
2898 (j) Added the -F (--fixed-strings) option.
2899
2900 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2901
2902 (l) Added the --colo(u)r option.
2903
2904 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2905 is not present by default.
2906
2907 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2908 items such as (?R) or (?1), when the recursion could match a number of
2909 alternatives. If it matched one of the alternatives, but subsequently,
2910 outside the recursion, there was a failure, the code tried to back up into
2911 the recursion. However, because of the way PCRE is implemented, this is not
2912 possible, and the result was an incorrect result from the match.
2913
2914 In order to prevent this happening, the specification of recursion has
2915 been changed so that all such subpatterns are automatically treated as
2916 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2917
2918 4. I had overlooked the fact that, in some locales, there are characters for
2919 which isalpha() is true but neither isupper() nor islower() are true. In
2920 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2921 and ordfeminine) are like this. This affected the treatment of \w and \W
2922 when they appeared in character classes, but not when they appeared outside
2923 a character class. The bit map for "word" characters is now created
2924 separately from the results of isalnum() instead of just taking it from the
2925 upper, lower, and digit maps. (Plus the underscore character, of course.)
2926
2927 5. The above bug also affected the handling of POSIX character classes such as
2928 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2929 permanent tables. Instead, the bit maps for such a class were previously
2930 created as the appropriate unions of the upper, lower, and digit bitmaps.
2931 Now they are created by subtraction from the [[:word:]] class, which has
2932 its own bitmap.
2933
2934 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2935 It is created by subtracting the vertical space characters (\x09, \x0a,
2936 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2937 subtraction was done in the overall bitmap for a character class, meaning
2938 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2939 be recognized. This bug has been fixed.
2940
2941 7. Patches from the folks at Google:
2942
2943 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2944 real life, but is still worth protecting against".
2945
2946 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2947 regular expressions".
2948
2949 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2950 have it.
2951
2952 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2953 "configure" and the latter not, in order to fix a problem somebody had
2954 with compiling the Arg class on HP-UX.
2955
2956 (e) Improve the error-handling of the C++ wrapper a little bit.
2957
2958 (f) New tests for checking recursion limiting.
2959
2960 8. The pcre_memmove() function, which is used only if the environment does not
2961 have a standard memmove() function (and is therefore rarely compiled),
2962 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2963 returning a result (though PCRE never actually uses the result).
2964
2965 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2966 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2967 returned instead of calling malloc() with an overflowing number that would
2968 most likely cause subsequent chaos.
2969
2970 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2971
2972 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2973 with this option is matched, the nmatch and pmatch options of regexec() are
2974 ignored.
2975
2976 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2977 provided in case anyone wants to the the POSIX interface with UTF-8
2978 strings.
2979
2980 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2981 C++ linking (needed for some HP-UX environments).
2982
2983 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2984 (unused parameter) and in the pcre_printint() function (omitted "default"
2985 switch label when the default is to do nothing).
2986
2987 15. Added some code to make it possible, when PCRE is compiled as a C++
2988 library, to replace subject pointers for pcre_exec() with a smart pointer
2989 class, thus making it possible to process discontinuous strings.
2990
2991 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2992 much the same function. They were added by different people who were trying
2993 to make PCRE easy to compile on non-Unix systems. It has been suggested
2994 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2995 for compiling on Windows systems. I have therefore replaced it with
2996 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2997 defaults to "extern" for C or "extern C" for C++, which works fine on
2998 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2999 SCOPE with something explicit in config.h. In addition:
3000
3001 (a) pcreposix.h still had just "extern" instead of either of these macros;
3002 I have replaced it with PCRE_DATA_SCOPE.
3003
3004 (b) Functions such as _pcre_xclass(), which are internal to the library,
3005 but external in the C sense, all had PCRE_EXPORT in their definitions.
3006 This is apparently wrong for the Windows case, so I have removed it.
3007 (It makes no difference on Unix-like systems.)
3008
3009 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3010 of recursive calls to match(). This is different to MATCH_LIMIT because
3011 that limits the total number of calls to match(), not all of which increase
3012 the depth of recursion. Limiting the recursion depth limits the amount of
3013 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3014 when PCRE is compiled, and changed at run time. A patch from Google adds
3015 this functionality to the C++ interface.
3016
3017 18. Changes to the handling of Unicode character properties:
3018
3019 (a) Updated the table to Unicode 4.1.0.
3020
3021 (b) Recognize characters that are not in the table as "Cn" (undefined).
3022
3023 (c) I revised the way the table is implemented to a much improved format
3024 which includes recognition of ranges. It now supports the ranges that
3025 are defined in UnicodeData.txt, and it also amalgamates other
3026 characters into ranges. This has reduced the number of entries in the
3027 table from around 16,000 to around 3,000, thus reducing its size
3028 considerably. I realized I did not need to use a tree structure after
3029 all - a binary chop search is just as efficient. Having reduced the
3030 number of entries, I extended their size from 6 bytes to 8 bytes to
3031 allow for more data.
3032
3033 (d) Added support for Unicode script names via properties such as \p{Han}.
3034
3035 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3036 matching that character.
3037
3038 20. When matching a repeated Unicode property with a minimum greater than zero,
3039 (for example \pL{2,}), PCRE could look past the end of the subject if it
3040 reached it while seeking the minimum number of characters. This could
3041 happen only if some of the characters were more than one byte long, because
3042 there is a check for at least the minimum number of bytes.
3043
3044 21. Refactored the implementation of \p and \P so as to be more general, to
3045 allow for more different types of property in future. This has changed the
3046 compiled form incompatibly. Anybody with saved compiled patterns that use
3047 \p or \P will have to recompile them.
3048
3049 22. Added "Any" and "L&" to the supported property types.
3050
3051 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3052 but give a compile time error if the value is greater than 0xff.
3053
3054 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3055 accidentally not being installed or uninstalled.
3056
3057 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3058 made were to insert the current release number. This seemed silly, because
3059 it made things harder for people building PCRE on systems that don't run
3060 "configure". I have turned pcre.h into a distributed file, no longer built
3061 by "configure", with the version identification directly included. There is
3062 no longer a pcre.h.in file.
3063
3064 However, this change necessitated a change to the pcre-config script as
3065 well. It is built from pcre-config.in, and one of the substitutions was the
3066 release number. I have updated configure.ac so that ./configure now finds
3067 the release number by grepping pcre.h.
3068
3069 26. Added the ability to run the tests under valgrind.
3070
3071
3072 Version 6.4 05-Sep-05
3073 ---------------------
3074
3075 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3076 "--" to be printed when multiple files were scanned, even when none of the
3077 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3078 consider it to be a bug, and have restored the previous behaviour.
3079
3080 2. A couple of code tidies to get rid of compiler warnings.
3081
3082 3. The pcretest program used to cheat by referring to symbols in the library
3083 whose names begin with _pcre_. These are internal symbols that are not
3084 really supposed to be visible externally, and in some environments it is
3085 possible to suppress them. The cheating is now confined to including
3086 certain files from the library's source, which is a bit cleaner.
3087
3088 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3089 file's purpose clearer.
3090
3091 5. Reorganized pcre_ucp_findchar().
3092
3093
3094 Version 6.3 15-Aug-05
3095 ---------------------
3096
3097 1. The file libpcre.pc.in did not have general read permission in the tarball.
3098
3099 2. There were some problems when building without C++ support:
3100
3101 (a) If C++ support was not built, "make install" and "make test" still
3102 tried to test it.
3103
3104 (b) There were problems when the value of CXX was explicitly set. Some
3105 changes have been made to try to fix these, and ...
3106
3107 (c) --disable-cpp can now be used to explicitly disable C++ support.
3108
3109 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3110 backslash in a target when C++ was disabled. This confuses some
3111 versions of "make", apparently. Using an intermediate variable solves
3112 this. (Same for CPP_LOBJ.)
3113
3114 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3115 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3116 necessary on certain architectures.
3117
3118 4. Added a setting of -export-symbols-regex to the link command to remove
3119 those symbols that are exported in the C sense, but actually are local
3120 within the library, and not documented. Their names all begin with
3121 "_pcre_". This is not a perfect job, because (a) we have to except some
3122 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3123 available (and never for static libraries). I have made a note to try to
3124 find a way round (a) in the future.
3125
3126
3127 Version 6.2 01-Aug-05
3128 ---------------------
3129
3130 1. There was no test for integer overflow of quantifier values. A construction
3131 such as {1111111111111111} would give undefined results. What is worse, if
3132 a minimum quantifier for a parenthesized subpattern overflowed and became
3133 negative, the calculation of the memory size went wrong. This could have
3134 led to memory overwriting.
3135
3136 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3137
3138 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3139 operating environments where this matters.
3140
3141 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3142 PCRE options from within the C++ wrapper.
3143
3144 5. Named capturing subpatterns were not being correctly counted when a pattern
3145 was compiled. This caused two problems: (a) If there were more than 100
3146 such subpatterns, the calculation of the memory needed for the whole
3147 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3148 back references of the form \12, where the number was greater than 9, were
3149 not recognized as back references, even though there were sufficient
3150 previous subpatterns.
3151
3152 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3153 versions of gcc, e.g. 2.95.4.
3154
3155
3156 Version 6.1 21-Jun-05
3157 ---------------------
3158
3159 1. There was one reference to the variable "posix" in pcretest.c that was not
3160 surrounded by "#if !defined NOPOSIX".
3161
3162 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3163 the cross-check on the old pcre_info() function, for the benefit of the
3164 cut-down version of PCRE that is currently imported into Exim.
3165
3166 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3167 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3168 patterns that start (?i) but I don't think that matters. The use of (?i) is
3169 just an example; this all applies to the other options as well.
3170
3171 4. Since libtool seems to echo the compile commands it is issuing, the output
3172 from "make" can be reduced a bit by putting "@" in front of each libtool
3173 compile command.
3174
3175 5. Patch from the folks at Google for configure.in to be a bit more thorough
3176 in checking for a suitable C++ installation before trying to compile the
3177 C++ stuff. This should fix a reported problem when a compiler was present,
3178 but no suitable headers.
3179
3180 6. The man pages all had just "PCRE" as their title. I have changed them to
3181 be the relevant file name. I have also arranged that these names are
3182 retained in the file doc/pcre.txt, which is a concatenation in text format
3183 of all the man pages except the little individual ones for each function.
3184
3185 7. The NON-UNIX-USE file had not been updated for the different set of source
3186 files that come with release 6. I also added a few comments about the C++
3187 wrapper.
3188
3189
3190 Version 6.0 07-Jun-05
3191 ---------------------
3192
3193 1. Some minor internal re-organization to help with my DFA experiments.
3194
3195 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3196 didn't matter for the library itself when fully configured, but did matter
3197 when compiling without UCP support, or within Exim, where the ucp files are
3198 not imported.
3199
3200 3. Refactoring of the library code to split up the various functions into
3201 different source modules. The addition of the new DFA matching code (see
3202 below) to a single monolithic source would have made it really too
3203 unwieldy, quite apart from causing all the code to be include in a
3204 statically linked application, when only some functions are used. This is
3205 relevant even without the DFA addition now that patterns can be compiled in
3206 one application and matched in another.
3207
3208 The downside of splitting up is that there have to be some external
3209 functions and data tables that are used internally in different modules of
3210 the library but which are not part of the API. These have all had their
3211 names changed to start with "_pcre_" so that they are unlikely to clash
3212 with other external names.
3213
3214 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3215 a different (DFA) algorithm. Although it is slower than the original
3216 function, it does have some advantages for certain types of matching
3217 problem.
3218
3219 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3220 including restarting after a partial match.
3221
3222 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3223 defined when compiling for Windows was sent to me. I have put it into the
3224 code, though I have no means of testing or verifying it.
3225
3226 7. Added the pcre_refcount() auxiliary function.
3227
3228 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3229 match before or at the first newline in the subject string. In pcretest,
3230 the /f option on a pattern can be used to set this.
3231
3232 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3233 would behave wrongly. This has been present in PCRE since release 4.0.
3234
3235 10. A number of changes to the pcregrep command:
3236
3237 (a) Refactored how -x works; insert ^(...)$ instead of setting
3238 PCRE_ANCHORED and checking the length, in preparation for adding
3239 something similar for -w.
3240
3241 (b) Added the -w (match as a word) option.
3242
3243 (c) Refactored the way lines are read and buffered so as to have more
3244 than one at a time available.
3245
3246 (d) Implemented a pcregrep test script.
3247
3248 (e) Added the -M (multiline match) option. This allows patterns to match
3249 over several lines of the subject. The buffering ensures that at least
3250 8K, or the rest of the document (whichever is the shorter) is available
3251 for matching (and similarly the previous 8K for lookbehind assertions).
3252
3253 (f) Changed the --help output so that it now says
3254
3255 -w, --word-regex(p)
3256
3257 instead of two lines, one with "regex" and the other with "regexp"
3258 because that confused at least one person since the short forms are the
3259 same. (This required a bit of code, as the output is generated
3260 automatically from a table. It wasn't just a text change.)
3261
3262 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3263 option but starts with a hyphen. Could be a pattern or a path name
3264 starting with a hyphen, for instance.
3265
3266 (h) "-" can be given as a file name to represent stdin.
3267
3268 (i) When file names are being printed, "(standard input)" is used for
3269 the standard input, for compatibility with GNU grep. Previously
3270 "<stdin>" was used.
3271
3272 (j) The option --label=xxx can be used to supply a name to be used for
3273 stdin when file names are being printed. There is no short form.
3274
3275 (k) Re-factored the options decoding logic because we are going to add
3276 two more options that take data. Such options can now be given in four
3277 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3278
3279 (l) Added the -A, -B, and -C options for requesting that lines of context
3280 around matches be printed.
3281
3282 (m) Added the -L option to print the names of files that do not contain
3283 any matching lines, that is, the complement of -l.
3284
3285 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3286 continue to scan other files.
3287
3288 (o) The -s option was incorrectly implemented. For compatibility with other
3289 greps, it now suppresses the error message for a non-existent or non-
3290 accessible file (but not the return code). There is a new option called
3291 -q that suppresses the output of matching lines, which was what -s was
3292 previously doing.
3293
3294 (p) Added --include and --exclude options to specify files for inclusion
3295 and exclusion when recursing.
3296
3297 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3298 Hopefully, it now does.
3299
3300 12. Missing cast in pcre_study().
3301
3302 13. Added an "uninstall" target to the makefile.
3303
3304 14. Replaced "extern" in the function prototypes in Makefile.in with
3305 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3306 world, but is set differently for Windows.
3307
3308 15. Added a second compiling function called pcre_compile2(). The only
3309 difference is that it has an extra argument, which is a pointer to an
3310 integer error code. When there is a compile-time failure, this is set
3311 non-zero, in addition to the error test pointer being set to point to an
3312 error message. The new argument may be NULL if no error number is required
3313 (but then you may as well call pcre_compile(), which is now just a
3314 wrapper). This facility is provided because some applications need a
3315 numeric error indication, but it has also enabled me to tidy up the way
3316 compile-time errors are handled in the POSIX wrapper.
3317
3318 16. Added VPATH=.libs to the makefile; this should help when building with one
3319 prefix path and installing with another. (Or so I'm told by someone who
3320 knows more about this stuff than I do.)
3321
3322 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3323 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3324 match everything, including newlines. This is not POSIX-compatible, but
3325 somebody wanted the feature. From pcretest it can be activated by using
3326 both the P and the s flags.
3327
3328 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3329
3330 19. libpcre.pc was being incorrectly installed as executable.
3331
3332 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3333 it now also looks for '\r' so that it will work unmodified on Windows.
3334
3335 21. Added Google's contributed C++ wrapper to the distribution.
3336
3337 22. Added some untidy missing memory free() calls in pcretest, to keep
3338 Electric Fence happy when testing.
3339
3340
3341
3342 Version 5.0 13-Sep-04
3343 ---------------------
3344
3345 1. Internal change: literal characters are no longer packed up into items
3346 containing multiple characters in a single byte-string. Each character
3347 is now matched using a separate opcode. However, there may be more than one
3348 byte in the character in UTF-8 mode.
3349
3350 2. The pcre_callout_block structure has two new fields: pattern_position and
3351 next_item_length. These contain the offset in the pattern to the next match
3352 item, and its length, respectively.
3353
3354 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3355 insertion of callouts before each pattern item. Added the /C option to
3356 pcretest to make use of this.
3357
3358 4. On the advice of a Windows user, the lines
3359
3360 #if defined(_WIN32) || defined(WIN32)
3361 _setmode( _fileno( stdout ), 0x8000 );
3362 #endif /* defined(_WIN32) || defined(WIN32) */
3363
3364 have been added to the source of pcretest. This apparently does useful
3365 magic in relation to line terminators.
3366
3367 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3368 for the benefit of those environments where the "b" makes a difference.
3369
3370 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3371 to know about it. I have put a hack into configure.in that adds in code
3372 to set GCC=yes if CC=icc. This seems to end up at a point in the
3373 generated configure script that is early enough to affect the setting of
3374 compiler options, which is what is needed, but I have no means of testing
3375 whether it really works. (The user who reported this had patched the
3376 generated configure script, which of course I cannot do.)
3377
3378 LATER: After change 22 below (new libtool files), the configure script
3379 seems to know about icc (and also ecc). Therefore, I have commented out
3380 this hack in configure.in.
3381
3382 7. Added support for pkg-config (2 patches were sent in).
3383
3384 8. Negated POSIX character classes that used a combination of internal tables
3385 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3386 [[:^ascii]]. Typically, they would match almost any characters. The other
3387 POSIX classes were not broken in this way.
3388
3389 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3390 to find the match, as PCRE was deluded into thinking that the match had to
3391 start at the start point or following a newline. The same bug applied to
3392 patterns with negative forward assertions or any backward assertions
3393 preceding ".*" at the start, unless the pattern required a fixed first
3394 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3395
3396 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3397 starting at the last subject character, bytes beyond the end of the subject
3398 string were read.
3399
3400 11. Renamed the variable "class" as "classbits" to make life easier for C++
3401 users. (Previously there was a macro definition, but it apparently wasn't
3402 enough.)
3403
3404 12. Added the new field "tables" to the extra data so that tables can be passed
3405 in at exec time, or the internal tables can be re-selected. This allows
3406 a compiled regex to be saved and re-used at a later time by a different
3407 program that might have everything at different addresses.
3408
3409 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3410 -R library as well as a -L library.
3411
3412 14. The debugging options of pcretest (-d on the command line or D on a
3413 pattern) showed incorrect output for anything following an extended class
3414 that contained multibyte characters and which was followed by a quantifier.
3415
3416 15. Added optional support for general category Unicode character properties
3417 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3418 support. It adds about 90K to the size of the library. The meanings of the
3419 inbuilt class escapes such as \d and \s have NOT been changed.
3420
3421 16. Updated pcredemo.c to include calls to free() to release the memory for the
3422 compiled pattern.
3423
3424 17. The generated file chartables.c was being created in the source directory
3425 instead of in the building directory. This caused the build to fail if the
3426 source directory was different from the building directory, and was
3427 read-only.
3428
3429 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3430 file. No doubt somebody will tell me if they don't make sense... Also added
3431 Dan Mooney's comments about building on OpenVMS.
3432
3433 19. Added support for partial matching via the PCRE_PARTIAL option for
3434 pcre_exec() and the \P data escape in pcretest.
3435
3436 20. Extended pcretest with 3 new pattern features:
3437
3438 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3439 write the compiled pattern to the file whose name is "rest-of-line".
3440 This is a straight binary dump of the data, with the saved pointer to
3441 the character tables forced to be NULL. The study data, if any, is
3442 written too. After writing, pcretest reads a new pattern.
3443
3444 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3445 compiled pattern from the given file. There must not be any
3446 occurrences of "<" in the file name (pretty unlikely); if there are,
3447 pcretest will instead treat the initial "<" as a pattern delimiter.
3448 After reading in the pattern, pcretest goes on to read data lines as
3449 usual.
3450
3451 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3452 and 16-bit fields in a compiled pattern, to simulate a pattern that
3453 was compiled on a host of opposite endianness.
3454
3455 21. The pcre-exec() function can now cope with patterns that were compiled on
3456 hosts of opposite endianness, with this restriction:
3457
3458 As for any compiled expression that is saved and used later, the tables
3459 pointer field cannot be preserved; the extra_data field in the arguments
3460 to pcre_exec() should be used to pass in a tables address if a value
3461 other than the default internal tables were used at compile time.
3462
3463 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3464 now diagnosed as an error. Previously, most of the time, a negative number
3465 would have been treated as zero, but if in addition "ovector" was passed as
3466 NULL, a crash could occur.
3467
3468 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3469 new versions from the libtool 1.5 distribution (the last one is a copy of
3470 a file called libtool.m4). This seems to have fixed the need to patch
3471 "configure" to support Darwin 1.3 (which I used to do). However, I still
3472 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3473 workstation).
3474
3475 24. Changed the PCRE licence to be the more standard "BSD" licence.
3476
3477
3478 Version 4.5 01-Dec-03
3479 ---------------------
3480
3481 1. There has been some re-arrangement of the code for the match() function so
3482 that it can be compiled in a version that does not call itself recursively.
3483 Instead, it keeps those local variables that need separate instances for
3484 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3485 needs to "recurse". Keeping track of where control must go is done by means
3486 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3487 hide most of the details from the main code, and operates only if
3488 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3489 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3490 operating.
3491
3492 To make it easier for callers to provide specially tailored get/free
3493 functions for this usage, two new functions, pcre_stack_malloc, and
3494 pcre_stack_free, are used. They are always called in strict stacking order,
3495 and the size of block requested is always the same.
3496
3497 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3498 PCRE has been compiled to use the stack or the heap for recursion. The
3499 -C option of pcretest uses this to show which version is compiled.
3500
3501 A new data escape \S, is added to pcretest; it causes the amounts of store
3502 obtained and freed by both kinds of malloc/free at match time to be added
3503 to the output.
3504
3505 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3506 what's available on my current Linux desktop machine.
3507
3508 3. When matching a UTF-8 string, the test for a valid string at the start has
3509 been extended. If start_offset is not zero, PCRE now checks that it points
3510 to a byte that is the start of a UTF-8 character. If not, it returns
3511 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3512 this is necessary because there may be backward assertions in the pattern.
3513 When matching the same subject several times, it may save resources to use
3514 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3515
3516 4. The code for checking the validity of UTF-8 strings has been tightened so
3517 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3518 containing "overlong sequences".
3519
3520 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3521 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3522 should have been "&", but it just so happened that all the cases this let
3523 through by mistake were picked up later in the function.
3524
3525 6. I had used a variable called "isblank" - this is a C99 function, causing
3526 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3527
3528 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3529 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3530 so the version is shown; (c) stop "make test" echoing "./RunTest".
3531
3532 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3533
3534 9. The prototype for memmove() for systems that don't have it was using
3535 size_t, but the inclusion of the header that defines size_t was later. I've
3536 moved the #includes for the C headers earlier to avoid this.
3537
3538 10. Added some adjustments to the code to make it easier to compiler on certain
3539 special systems:
3540
3541 (a) Some "const" qualifiers were missing.
3542 (b) Added the macro EXPORT before all exported functions; by default this
3543 is defined to be empty.
3544 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3545 that it reads its output file name as an argument instead of writing
3546 to the standard output and assuming this can be redirected.
3547
3548 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3549 class containing characters with values greater than 255, PCRE compilation
3550 went into a loop.
3551
3552 12. A recursive reference to a subpattern that was within another subpattern
3553 that had a minimum quantifier of zero caused PCRE to crash. For example,
3554 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3555 recursion. If the recursively-called subpattern itself had a zero repeat,
3556 that was OK.
3557
3558 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3559 buffer into which it was copied (for escape processing) was still set at
3560 1024, so long lines caused crashes.
3561
3562 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3563 "internal error: code overflow...". This applied to any character class
3564 that was followed by a possessive quantifier.
3565
3566 15. Modified the Makefile to add libpcre.la as a prerequisite for
3567 libpcreposix.la because I was told this is needed for a parallel build to
3568 work.
3569
3570 16. If a pattern that contained .* following optional items at the start was
3571 studied, the wrong optimizing data was generated, leading to matching
3572 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3573 matching string must start with a or b or c. The correct conclusion for
3574 this pattern is that a match can start with any character.
3575
3576
3577 Version 4.4 13-Aug-03
3578 ---------------------
3579
3580 1. In UTF-8 mode, a character class containing characters with values between
3581 127 and 255 was not handled correctly if the compiled pattern was studied.
3582 In fixing this, I have also improved the studying algorithm for such
3583 classes (slightly).
3584
3585 2. Three internal functions had redundant arguments passed to them. Removal
3586 might give a very teeny performance improvement.
3587
3588 3. Documentation bug: the value of the capture_top field in a callout is *one
3589 more than* the number of the hightest numbered captured substring.
3590
3591 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3592 in incorrectly linking with a previously installed version. They now link
3593 explicitly with libpcre.la.
3594
3595 5. configure.in no longer needs to recognize Cygwin specially.
3596
3597 6. A problem in pcre.in for Windows platforms is fixed.
3598
3599 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3600 pcretest, it used to include the size of the study block as part of its
3601 output. Unfortunately, the structure contains a field that has a different
3602 size on different hardware architectures. This meant that the tests that
3603 showed this size failed. As the block is currently always of a fixed size,
3604 this information isn't actually particularly useful in pcretest output, so
3605 I have just removed it.
3606
3607 8. Three pre-processor statements accidentally did not start in column 1.
3608 Sadly, there are *still* compilers around that complain, even though
3609 standard C has not required this for well over a decade. Sigh.
3610
3611 9. In pcretest, the code for checking callouts passed small integers in the
3612 callout_data field, which is a void * field. However, some picky compilers
3613 complained about the casts involved for this on 64-bit systems. Now
3614 pcretest passes the address of the small integer instead, which should get
3615 rid of the warnings.
3616
3617 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3618 both compile and run time, and gives an error if an invalid UTF-8 sequence
3619 is found. There is a option for disabling this check in cases where the
3620 string is known to be correct and/or the maximum performance is wanted.
3621
3622 11. In response to a bug report, I changed one line in Makefile.in from
3623
3624 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3625 to
3626 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3627
3628 to look similar to other lines, but I have no way of telling whether this
3629 is the right thing to do, as I do not use Windows. No doubt I'll get told
3630 if it's wrong...
3631
3632
3633 Version 4.3 21-May-03
3634 ---------------------
3635
3636 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3637 Makefile.
3638
3639 2. Some refactoring to improve the quality of the code:
3640
3641 (i) The utf8_table... variables are now declared "const".
3642
3643 (ii) The code for \cx, which used the "case flipping" table to upper case
3644 lower case letters, now just substracts 32. This is ASCII-specific,
3645 but the whole concept of \cx is ASCII-specific, so it seems
3646 reasonable.
3647
3648 (iii) PCRE was using its character types table to recognize decimal and
3649 hexadecimal digits in the pattern. This is silly, because it handles
3650 only 0-9, a-f, and A-F, but the character types table is locale-
3651 specific, which means strange things might happen. A private
3652 table is now used for this - though it costs 256 bytes, a table is
3653 much faster than multiple explicit tests. Of course, the standard
3654 character types table is still used for matching digits in subject
3655 strings against \d.
3656
3657 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3658 ending in _t are). So I've renamed it as ESC_tee.
3659
3660 3. The first argument for regexec() in the POSIX wrapper should have been
3661 defined as "const".
3662
3663 4. Changed pcretest to use malloc() for its buffers so that they can be
3664 Electric Fenced for debugging.
3665
3666 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3667 to read one or more bytes before the start of the subject string. Often this
3668 had no effect on PCRE's behaviour, but in some circumstances it could
3669 provoke a segmentation fault.
3670
3671 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3672 to try to read one or more bytes before the start of the subject string.
3673
3674 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3675 UTF-8 support could misbehave in various ways if the subject string
3676 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3677 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3678 back over UTF-8 characters.)
3679
3680
3681 Version 4.2 14-Apr-03
3682 ---------------------
3683
3684 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3685
3686 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3687 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3688 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3689 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3690 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3691 and BUILD_EXEEXT
3692 Note: automatic setting of the BUILD variables is not yet working
3693 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3694 compile-time but not at link-time
3695 [LINK]: use for linking executables only
3696 make different versions for Windows and non-Windows
3697 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3698 libraries
3699 [LINK_FOR_BUILD]: new variable
3700 [OBJEXT]: use throughout
3701 [EXEEXT]: use throughout
3702 <winshared>: new target
3703 <wininstall>: new target
3704 <dftables.o>: use native compiler
3705 <dftables>: use native linker
3706 <install>: handle Windows platform correctly
3707 <clean>: ditto
3708 <check>: ditto
3709 copy DLL to top builddir before testing
3710
3711 As part of these changes, -no-undefined was removed again. This was reported
3712 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3713 in any case.
3714
3715 3. Some tidies to get rid of compiler warnings:
3716
3717 . In the match_data structure, match_limit was an unsigned long int, whereas
3718 match_call_count was an int. I've made them both unsigned long ints.
3719
3720 . In pcretest the fact that a const uschar * doesn't automatically cast to
3721 a void * provoked a warning.
3722
3723 . Turning on some more compiler warnings threw up some "shadow" variables
3724 and a few more missing casts.
3725
3726 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3727 option, a class that contained a single character with a value between 128
3728 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3729
3730 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3731 option, a class that contained several characters, but with at least one
3732 whose value was between 128 and 255 caused PCRE to crash.
3733
3734
3735 Version 4.1 12-Mar-03
3736 ---------------------
3737
3738 1. Compiling with gcc -pedantic found a couple of places where casts were
3739 needed, and a string in dftables.c that was longer than standard compilers are
3740 required to support.
3741
3742 2. Compiling with Sun's compiler found a few more places where the code could
3743 be tidied up in order to avoid warnings.
3744
3745 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3746 first of these names is deprecated in the latest Autoconf in favour of the name
3747 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3748 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3749 analogy I have changed it to CFLAGS_FOR_BUILD.
3750
3751 4. Added -no-undefined to the linking command in the Makefile, because this is
3752 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3753 linking step for the pcreposix library.
3754
3755 5. PCRE was failing to diagnose the case of two named groups with the same
3756 name.
3757
3758 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3759 literal character that is needed in the subject for a match, and scans along to
3760 ensure that it is present before embarking on the full matching process. This
3761 saves time in cases of nested unlimited repeats that are never going to match.
3762 Problem: the scan can take a lot of time if the subject is very long (e.g.
3763 megabytes), thus penalizing straightforward matches. It is now done only if the
3764 amount of subject to be scanned is less than 1000 bytes.
3765
3766 7. A lesser problem with the same optimization is that it was recording the
3767 first character of an anchored pattern as "needed", thus provoking a search
3768 right along the subject, even when the first match of the pattern was going to
3769 fail. The "needed" character is now not set for anchored patterns, unless it
3770 follows something in the pattern that is of non-fixed length. Thus, it still
3771 fulfils its original purpose of finding quick non-matches in cases of nested
3772 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3773
3774
3775 Version 4.0 17-Feb-03
3776 ---------------------
3777
3778 1. If a comment in an extended regex that started immediately after a meta-item
3779 extended to the end of string, PCRE compiled incorrect data. This could lead to
3780 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3781
3782 2. Moved to autoconf 2.53 and libtool 1.4.2.
3783
3784 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3785 the special perltest8 script is no longer needed - all the tests can be run
3786 from a single perltest script.
3787
3788 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3789 by \s. It has now been removed in PCRE. This means it isn't recognized as
3790 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3791 class [:space:] *does* include VT, thereby creating a mess.
3792
3793 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3794 space and tab.
3795
3796 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3797 its new features into the main test script, reducing the number of scripts.
3798
3799 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3800 were backward compatible, and made the (?i) apply to the whole pattern, as if
3801 /i were given. Now it behaves more logically, and applies the option setting
3802 only to what follows. PCRE has been changed to follow suit. However, if it
3803 finds options settings right at the start of the pattern, it extracts them into
3804 the global options, as before. Thus, they show up in the info data.
3805
3806 8. Added support for the \Q...\E escape sequence. Characters in between are
3807 treated as literals. This is slightly different from Perl in that $ and @ are
3808 also handled as literals inside the quotes. In Perl, they will cause variable
3809 interpolation. Note the following examples:
3810
3811 Pattern PCRE matches Perl matches
3812
3813 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3814 \Qabc\$xyz\E abc\$xyz abc\$xyz
3815 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3816
3817 For compatibility with Perl, \Q...\E sequences are recognized inside character
3818 classes as well as outside them.
3819
3820 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3821 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3822 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3823 signed/unsigned warnings.
3824
3825 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3826 option for pcretest, so I've replaced it by a simple function that does just
3827 that job.
3828
3829 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3830 "pcregrep -".
3831
3832 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3833 Java package. This provides some syntactic sugar for simple cases of what my
3834 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3835 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3836 item, you can use this simplified notation. Note that only makes sense with
3837 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3838 greediness, whatever the setting of the PCRE_UNGREEDY option.
3839
3840 13. A change of greediness default within a pattern was not taking effect at
3841 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3842 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3843 was abstracted outside.
3844
3845 14. PCRE now supports the \G assertion. It is true when the current matching
3846 position is at the start point of the match. This differs from \A when the
3847 starting offset is non-zero. Used with the /g option of pcretest (or similar
3848 code), it works in the same way as it does for Perl's /g option. If all
3849 alternatives of a regex begin with \G, the expression is anchored to the start
3850 match position, and the "anchored" flag is set in the compiled expression.
3851
3852 15. Some bugs concerning the handling of certain option changes within patterns
3853 have been fixed. These applied to options other than (?ims). For example,
3854 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3855 been the other way round. Some of this was related to change 7 above.
3856
3857 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3858 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3859 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3860 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3861
3862 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3863 mode. Unlike ".", it always matches newline, whatever the setting of
3864 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3865 assertions. Perl allows it, but it doesn't (in general) work because it can't
3866 calculate the length of the lookbehind. At least, that's the case for Perl
3867 5.8.0 - I've been told they are going to document that it doesn't work in
3868 future.
3869
3870 18. Added an error diagnosis for escapes that PCRE does not support: these are
3871 \L, \l, \N, \P, \p, \U, \u, and \X.
3872
3873 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3874 reading past the end of the pattern in cases such as /[abcd/.
3875
3876 20. PCRE was getting more memory than necessary for patterns with classes that
3877 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3878
3879 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3880 compiling PCRE for use with Virtual Pascal.
3881
3882 22. Small fix to the Makefile to make it work properly if the build is done
3883 outside the source tree.
3884
3885 23. Added a new extension: a condition to go with recursion. If a conditional
3886 subpattern starts with (?(R) the "true" branch is used if recursion has
3887 happened, whereas the "false" branch is used only at the top level.
3888
3889 24. When there was a very long string of literal characters (over 255 bytes
3890 without UTF support, over 250 bytes with UTF support), the computation of how
3891 much memory was required could be incorrect, leading to segfaults or other
3892 strange effects.
3893
3894 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3895 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3896 there was a subsequent back reference to those brackets. This meant that, for
3897 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3898 possible to check for precisely this case. All we can do is abandon the
3899 optimization if .* occurs inside capturing brackets when there are any back
3900 references whatsoever. (See below for a better fix that came later.)
3901
3902 26. The handling of the optimization for finding the first character of a
3903 non-anchored pattern, and for finding a character that is required later in the
3904 match were failing in some cases. This didn't break the matching; it just
3905 failed to optimize when it could. The way this is done has been re-implemented.
3906
3907 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3908
3909 28. Added a new feature that provides some of the functionality that Perl
3910 provides with (?{...}). The facility is termed a "callout". The way it is done
3911 in PCRE is for the caller to provide an optional function, by setting
3912 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3913 global variable. By default it is unset, which disables all calling out. To get
3914 the function called, the regex must include (?C) at appropriate points. This
3915 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3916 This provides a means of identifying different callout points. When PCRE
3917 reaches such a point in the regex, if pcre_callout has been set, the external
3918 function is called. It is provided with data in a structure called
3919 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3920 matching continues; if it returns a non-zero value, the match at the current
3921 point fails. However, backtracking will occur if possible. [This was changed
3922 later and other features added - see item 49 below.]
3923
3924 29. pcretest is upgraded to test the callout functionality. It provides a
3925 callout function that displays information. By default, it shows the start of
3926 the match and the current position in the text. There are some new data escapes
3927 to vary what happens:
3928
3929 \C+ in addition, show current contents of captured substrings
3930 \C- do not supply a callout function
3931 \C!n return 1 when callout number n is reached
3932 \C!n!m return 1 when callout number n is reached for the mth time
3933
3934 30. If pcregrep was called with the -l option and just a single file name, it
3935 output "<stdin>" if a match was found, instead of the file name.
3936
3937 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3938 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3939 pcre_exec(). This saves a malloc/free per call. The default value of
3940 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3941 when configuring.
3942
3943 32. The default maximum size of a compiled pattern is 64K. There have been a
3944 few cases of people hitting this limit. The code now uses macros to handle the
3945 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3946 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3947 configuring. Tests 2 and 5 work only with 2-byte links because they output
3948 debugging information about compiled patterns.
3949
3950 33. Internal code re-arrangements:
3951
3952 (a) Moved the debugging function for printing out a compiled regex into
3953 its own source file (printint.c) and used #include to pull it into
3954 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3955 separate copies.
3956
3957 (b) Defined the list of op-code names for debugging as a macro in
3958 internal.h so that it is next to the definition of the opcodes.
3959
3960 (c) Defined a table of op-code lengths for simpler skipping along compiled
3961 code. This is again a macro in internal.h so that it is next to the
3962 definition of the opcodes.
3963
3964 34. Added support for recursive calls to individual subpatterns, along the
3965 lines of Robin Houston's patch (but implemented somewhat differently).
3966
3967 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3968 allow it to read and process whole directories in Win32. This code was
3969 contributed by Lionel Fourquaux; it has not been tested by me.
3970
3971 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3972 used to name a group. Names consist of alphanumerics and underscores, and must
3973 be unique. Back references use the syntax (?P=name) and recursive calls use
3974 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3975 numbers. The function pcre_fullinfo() can be used after compilation to extract
3976 a name/number map. There are three relevant calls:
3977
3978 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3979 PCRE_INFO_NAMECOUNT yields the number of entries
3980 PCRE_INFO_NAMETABLE yields a pointer to the map.
3981
3982 The map is a vector of fixed-size entries. The size of each entry depends on
3983 the length of the longest name used. The first two bytes of each entry are the
3984 group number, most significant byte first. There follows the corresponding
3985 name, zero terminated. The names are in alphabetical order.
3986
3987 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3988 case instead of 255. Making it the same both with and without UTF-8 support
3989 means that the same test output works with both.
3990
3991 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3992 calling malloc() with a zero argument.
3993
3994 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3995 optimization. I've improved things by keeping a bitmap of backreferences with
3996 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3997 fact referenced, the optimization can be applied. It is unlikely that a
3998 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3999 the match to follow \n) will appear inside brackets with a number greater than
4000 31, but if it does, any back reference > 31 suppresses the optimization.
4001
4002 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4003 of disabling numbered capturing parentheses. Any opening parenthesis that is
4004 not followed by ? behaves as if it were followed by ?: but named parentheses
4005 can still be used for capturing (and they will acquire numbers in the usual
4006 way).
4007
4008 41. Redesigned the return codes from the match() function into yes/no/error so
4009 that errors can be passed back from deep inside the nested calls. A malloc
4010 failure while inside a recursive subpattern call now causes the
4011 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4012
4013 42. It is now possible to set a limit on the number of times the match()
4014 function is called in a call to pcre_exec(). This facility makes it possible to
4015 limit the amount of recursion and backtracking, though not in a directly
4016 obvious way, because the match() function is used in a number of different
4017 circumstances. The count starts from zero for each position in the subject
4018 string (for non-anchored patterns). The default limit is, for compatibility, a
4019 large number, namely 10 000 000. You can change this in two ways:
4020
4021 (a) When configuring PCRE before making, you can use --with-match-limit=n
4022 to set a default value for the compiled library.
4023
4024 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4025 a different value is set. See 45 below.
4026
4027 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4028
4029 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4030 of things that can be changed at compile time. The first argument specifies
4031 what is wanted and the second points to where the information is to be placed.
4032 The current list of available information is:
4033
4034 PCRE_CONFIG_UTF8
4035
4036 The output is an integer that is set to one if UTF-8 support is available;
4037 otherwise it is set to zero.
4038
4039 PCRE_CONFIG_NEWLINE
4040
4041 The output is an integer that it set to the value of the code that is used for
4042 newline. It is either LF (10) or CR (13).
4043
4044 PCRE_CONFIG_LINK_SIZE
4045
4046 The output is an integer that contains the number of bytes used for internal
4047 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4048
4049 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4050
4051 The output is an integer that contains the threshold above which the POSIX
4052 interface uses malloc() for output vectors. See item 31 above.
4053
4054 PCRE_CONFIG_MATCH_LIMIT
4055
4056 The output is an unsigned integer that contains the default limit of the number
4057 of match() calls in a pcre_exec() execution. See 42 above.
4058
4059 44. pcretest has been upgraded by the addition of the -C option. This causes it
4060 to extract all the available output from the new pcre_config() function, and to
4061 output it. The program then exits immediately.
4062
4063 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4064 order to support additional features. One way would have been to define
4065 pcre_exec2() (for example) with extra arguments, but this would not have been
4066 extensible, and would also have required all calls to the original function to
4067 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4068 is used for passing in "extra" data from pcre_study().
4069
4070 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4071 contains the following fields:
4072
4073 flags a bitmap indicating which of the following fields are set
4074 study_data opaque data from pcre_study()
4075 match_limit a way of specifying a limit on match() calls for a specific
4076 call to pcre_exec()
4077 callout_data data for callouts (see 49 below)
4078
4079 The flag bits are also defined in pcre.h, and are
4080
4081 PCRE_EXTRA_STUDY_DATA
4082 PCRE_EXTRA_MATCH_LIMIT
4083 PCRE_EXTRA_CALLOUT_DATA
4084
4085 The pcre_study() function now returns one of these new pcre_extra blocks, with
4086 the actual study data pointed to by the study_data field, and the
4087 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4088 before. That is, this change is entirely upwards-compatible and requires no
4089 change to existing code.
4090
4091 If you want to pass in additional data to pcre_exec(), you can either place it
4092 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4093 block.
4094
4095 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4096 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4097 times with different match limits, until it finds the minimum value needed for
4098 pcre_exec() to complete. The value is then output. This can be instructive; for
4099 most simple matches the number is quite small, but for pathological cases it
4100 gets very large very quickly.
4101
4102 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4103 returns the size of the data block pointed to by the study_data field in a
4104 pcre_extra block, that is, the value that was passed as the argument to
4105 pcre_malloc() when PCRE was getting memory in which to place the information
4106 created by pcre_study(). The fourth argument should point to a size_t variable.
4107 pcretest has been extended so that this information is shown after a successful
4108 pcre_study() call when information about the compiled regex is being displayed.
4109
4110 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4111 because what follows is always an absolute path. (Later: it turns out that this
4112 is more than cosmetic for MinGW, because it doesn't like empty path
4113 components.)
4114
4115 49. Some changes have been made to the callout feature (see 28 above):
4116
4117 (i) A callout function now has three choices for what it returns:
4118
4119 0 => success, carry on matching
4120 > 0 => failure at this point, but backtrack if possible
4121 < 0 => serious error, return this value from pcre_exec()
4122
4123 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4124 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4125 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4126 use by callout functions. It will never be used by PCRE itself.
4127
4128 (ii) The pcre_extra structure (see 45 above) has a void * field called
4129 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4130 pcre_callout_block structure has a field of the same name. The contents of
4131 the field passed in the pcre_extra structure are passed to the callout
4132 function in the corresponding field in the callout block. This makes it
4133 easier to use the same callout-containing regex from multiple threads. For
4134 testing, the pcretest program has a new data escape
4135
4136 \C*n pass the number n (may be negative) as callout_data
4137
4138 If the callout function in pcretest receives a non-zero value as
4139 callout_data, it returns that value.
4140
4141 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4142 there were some redundant $(CFLAGS) in commands that are now specified as
4143 $(LINK), which already includes $(CFLAGS).
4144
4145 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4146 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4147 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4148 one-byte characters throughout. Note that case-insensitive matching applies
4149 only to characters whose values are less than 256. PCRE doesn't support the
4150 notion of cases for higher-valued characters.
4151
4152 (i) A character class whose characters are all within 0-255 is handled as
4153 a bit map, and the map is inverted for negative classes. Previously, a
4154 character > 255 always failed to match such a class; however it should
4155 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4156
4157 (ii) A negated character class with a single character < 255 is coded as
4158 "not this character" (OP_NOT). This wasn't working properly when the test
4159 character was multibyte, either singly or repeated.
4160
4161 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4162 mode, for example: \x{100}{2,3}.
4163
4164 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4165 singly or repeated) now correctly test multibyte characters. However,
4166 PCRE doesn't recognize any characters with values greater than 255 as
4167 digits, spaces, or word characters. Such characters always match \D, \S,
4168 and \W, and never match \d, \s, or \w.
4169
4170 (v) Classes may now contain characters and character ranges with values
4171 greater than 255. For example: [ab\x{100}-\x{400}].
4172
4173 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4174 PCRE in UTF-8 mode.
4175
4176 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4177 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4178 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4179 value.)
4180
4181 53. The single man page has become too large. I have therefore split it up into
4182 a number of separate man pages. These also give rise to individual HTML pages;
4183 these are now put in a separate directory, and there is an index.html page that
4184 lists them all. Some hyperlinking between the pages has been installed.
4185
4186 54. Added convenience functions for handling named capturing parentheses.
4187
4188 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4189 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4190 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4191 are faulted.
4192
4193 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4194 calling configure. These values are used when compiling the dftables.c program
4195 which is run to generate the source of the default character tables. They
4196 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4197 you will need to set these values.
4198
4199 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4200
4201
4202 Version 3.9 02-Jan-02
4203 ---------------------
4204
4205 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4206
4207 2. If --disable-static was given, the building process failed when trying to
4208 build pcretest and pcregrep. (For some reason it was using libtool to compile
4209 them, which is not right, as they aren't part of the library.)
4210
4211
4212 Version 3.8 18-Dec-01
4213 ---------------------
4214
4215 1. The experimental UTF-8 code was completely screwed up. It was packing the
4216 bytes in the wrong order. How dumb can you get?
4217
4218
4219 Version 3.7 29-Oct-01
4220 ---------------------
4221
4222 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4223 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4224 this didn't happen under Solaris 8, where I normally test things.
4225
4226 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4227 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4228 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4229 this fix an hour or so after the initial 3.7 release.)
4230
4231
4232 Version 3.6 23-Oct-01
4233 ---------------------
4234
4235 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4236 offsets passed as NULL with zero offset count.
4237
4238 2. The config.guess and config.sub files had not been updated when I moved to
4239 the latest autoconf.
4240
4241
4242 Version 3.5 15-Aug-01
4243 ---------------------
4244
4245 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4246 had been forgotten.
4247
4248 2. By using declared but undefined structures, we can avoid using "void"
4249 definitions in pcre.h while keeping the internal definitions of the structures
4250 private.
4251
4252 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4253 user point of view, this means that both static and shared libraries are built
4254 by default, but this can be individually controlled. More of the work of
4255 handling this static/shared cases is now inside libtool instead of PCRE's make
4256 file.
4257
4258 4. The pcretest utility is now installed along with pcregrep because it is
4259 useful for users (to test regexs) and by doing this, it automatically gets
4260 relinked by libtool. The documentation has been turned into a man page, so
4261 there are now .1, .txt, and .html versions in /doc.
4262
4263 5. Upgrades to pcregrep:
4264 (i) Added long-form option names like gnu grep.
4265 (ii) Added --help to list all options with an explanatory phrase.
4266 (iii) Added -r, --recursive to recurse into sub-directories.
4267 (iv) Added -f, --file to read patterns from a file.
4268
4269 6. pcre_exec() was referring to its "code" argument before testing that
4270 argument for NULL (and giving an error if it was NULL).
4271
4272 7. Upgraded Makefile.in to allow for compiling in a different directory from
4273 the source directory.
4274
4275 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4276 options bits, the pointer it was passed was to an int instead of to an unsigned
4277 long int. This mattered only on 64-bit systems.
4278
4279 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4280 generated) instead of pcre.in, which it its source. Also made the same change
4281 in several of the .c files.
4282
4283 10. A new release of gcc defines printf() as a macro, which broke pcretest
4284 because it had an ifdef in the middle of a string argument for printf(). Fixed
4285 by using separate calls to printf().
4286
4287 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4288 script, to force use of CR or LF instead of \n in the source. On non-Unix
4289 systems, the value can be set in config.h.
4290
4291 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4292 absolute limit. Changed the text of the error message to make this clear, and
4293 likewise updated the man page.
4294
4295 13. The limit of 99 on the number of capturing subpatterns has been removed.
4296 The new limit is 65535, which I hope will not be a "real" limit.
4297
4298
4299 Version 3.4 22-Aug-00
4300 ---------------------
4301
4302 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4303
4304 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4305
4306
4307 Version 3.3 01-Aug-00
4308 ---------------------
4309
4310 1. If an octal character was given, but the value was greater than \377, it
4311 was not getting masked to the least significant bits, as documented. This could
4312 lead to crashes in some systems.
4313
4314 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4315 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4316
4317 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4318 These just pass their arguments on to (pcre_free)(), but they are provided
4319 because some uses of PCRE bind it to non-C systems that can call its functions,
4320 but cannot call free() or pcre_free() directly.
4321
4322 4. Add "make test" as a synonym for "make check". Corrected some comments in
4323 the Makefile.
4324
4325 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4326 Makefile.
4327
4328 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4329 command called pgrep for grepping around the active processes.
4330
4331 7. Added the beginnings of support for UTF-8 character strings.
4332
4333 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4334 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4335 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4336 out for the ar command.)
4337
4338
4339 Version 3.2 12-May-00
4340 ---------------------
4341
4342 This is purely a bug fixing release.
4343
4344 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4345 of ZA. This was just one example of several cases that could provoke this bug,
4346 which was introduced by change 9 of version 2.00. The code for breaking
4347 infinite loops after an iteration that matches an empty string was't working
4348 correctly.
4349
4350 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4351 when matched against abbab (for example). After matching an empty string, it
4352 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4353 caused it to match further down the string than it should.
4354
4355 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4356 was there because it doesn't seem to be needed, and it causes trouble on some
4357 systems, as it is not a Standard C header. It has been removed.
4358
4359 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4360 were reported on the Macintosh. The changes were from
4361
4362 while ((c = *(++ptr)) != 0 && c != '\n');
4363 to
4364 while ((c = *(++ptr)) != 0 && c != '\n') ;
4365
4366 Totally extraordinary, but if that's what it takes...
4367
4368 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4369 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4370 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4371 assumes the way PCRE uses memmove() (always moving upwards).
4372
4373 6. PCRE is being used in one environment where strchr() is not available. There
4374 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4375 faster code anyway.
4376
4377
4378 Version 3.1 09-Feb-00
4379 ---------------------
4380
4381 The only change in this release is the fixing of some bugs in Makefile.in for
4382 the "install" target:
4383
4384 (1) It was failing to install pcreposix.h.
4385
4386 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4387
4388
4389 Version 3.0 01-Feb-00
4390 ---------------------
4391
4392 1. Add support for the /+ modifier to perltest (to output $` like it does in
4393 pcretest).
4394
4395 2. Add support for the /g modifier to perltest.
4396
4397 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4398 matches null strings.
4399
4400 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4401 pattern. Perl treats empty patterns specially - it reuses the most recent
4402 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4403 effect.
4404
4405 5. The POSIX interface was broken in that it was just handing over the POSIX
4406 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4407 required a bigger vector, with some working space on the end. This means that
4408 the POSIX wrapper now has to get and free some memory, and copy the results.
4409
4410 6. Added some simple autoconf support, placing the test data and the
4411 documentation in separate directories, re-organizing some of the
4412 information files, and making it build pcre-config (a GNU standard). Also added
4413 libtool support for building PCRE as a shared library, which is now the
4414 default.
4415
4416 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4417 09 are not valid octal constants. Single digits will be used for minor values
4418 less than 10.
4419
4420 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4421 existing programs that set these in the POSIX interface can use PCRE without
4422 modification.
4423
4424 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4425 return all that pcre_info() returns, plus additional data. The pcre_info()
4426 function is retained for compatibility, but is considered to be obsolete.
4427
4428 10. Added experimental recursion feature (?R) to handle one common case that
4429 Perl 5.6 will be able to do with (?p{...}).
4430
4431 11. Added support for POSIX character classes like [:alpha:], which Perl is
4432 adopting.
4433
4434
4435 Version 2.08 31-Aug-99
4436 ----------------------
4437
4438 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4439 trying to match at the startoffset position, but instead was moving forward to
4440 the next newline as if a previous match had failed.
4441
4442 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4443 and could get into a loop if a null string was matched other than at the start
4444 of the subject.
4445
4446 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4447 be distinguished at compile time, and for completeness also added PCRE_DATE.
4448
4449 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4450 in GnuWin32 environments.
4451
4452
4453 Version 2.07 29-Jul-99
4454 ----------------------
4455
4456 1. The documentation is now supplied in plain text form and HTML as well as in
4457 the form of man page sources.
4458
4459 2. C++ compilers don't like assigning (void *) values to other pointer types.
4460 In particular this affects malloc(). Although there is no problem in Standard
4461 C, I've put in casts to keep C++ compilers happy.
4462
4463 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4464 should be (const char *).
4465
4466 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4467 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4468 However, I haven't made this a standard facility. The documentation doesn't
4469 mention it, and the Makefile doesn't support it.
4470
4471 5. The Makefile now contains an "install" target, with editable destinations at
4472 the top of the file. The pcretest program is not installed.
4473
4474 6. pgrep -V now gives the PCRE version number and date.
4475
4476 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4477 causing the entire string to be ignored, instead of just the last character.
4478
4479 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4480 non-matching string, it can take a very, very long time, even for strings of
4481 quite modest length, because of the nested recursion. PCRE now does better in
4482 some of these cases. It does this by remembering the last required literal
4483 character in the pattern, and pre-searching the subject to ensure it is present
4484 before running the real match. In other words, it applies a heuristic to detect
4485 some types of certain failure quickly, and in the above example, if presented
4486 with a string that has no trailing " it gives "no match" very quickly.
4487
4488 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4489 other alternatives are tried instead.
4490
4491
4492 Version 2.06 09-Jun-99
4493 ----------------------
4494
4495 1. Change pcretest's output for amount of store used to show just the code
4496 space, because the remainder (the data block) varies in size between 32-bit and
4497 64-bit systems.
4498
4499 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4500 start matching at. This allows lookbehinds to work when searching for multiple
4501 occurrences in a string.
4502
4503 3. Added additional options to pcretest for testing multiple occurrences:
4504
4505 /+ outputs the rest of the string that follows a match
4506 /g loops for multiple occurrences, using the new startoffset argument
4507 /G loops for multiple occurrences by passing an incremented pointer
4508
4509 4. PCRE wasn't doing the "first character" optimization for patterns starting
4510 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4511 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4512 the letter 'x'. On long subject strings, this gives a significant speed-up.
4513
4514
4515 Version 2.05 21-Apr-99
4516 ----------------------
4517
4518 1. Changed the type of magic_number from int to long int so that it works
4519 properly on 16-bit systems.
4520
4521 2. Fixed a bug which caused patterns starting with .* not to work correctly
4522 when the subject string contained newline characters. PCRE was assuming
4523 anchoring for such patterns in all cases, which is not correct because .* will
4524 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4525 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4526 must be retried after every newline in the subject.
4527
4528
4529 Version 2.04 18-Feb-99
4530 ----------------------
4531
4532 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4533 computation of the store needed to hold the pattern was incorrect (too large).
4534 If such patterns were nested a few deep, this could multiply and become a real
4535 problem.
4536
4537 2. Added /M option to pcretest to show the memory requirement of a specific
4538 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4539
4540 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4541 compiled in such a way that the backtracking after subsequent failure was
4542 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4543 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4544
4545
4546 Version 2.03 02-Feb-99
4547 ----------------------
4548
4549 1. Fixed typo and small mistake in man page.
4550
4551 2. Added 4th condition (GPL supersedes if conflict) and created separate
4552 LICENCE file containing the conditions.
4553
4554 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4555 Perl, that is the internal \ allows the delimiter to be included in the
4556 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4557 the final delimiter, add \ to the end of the pattern (to test the error).
4558
4559 4. Added the convenience functions for extracting substrings after a successful
4560 match. Updated pcretest to make it able to test these functions.
4561
4562
4563 Version 2.02 14-Jan-99
4564 ----------------------
4565
4566 1. Initialized the working variables associated with each extraction so that
4567 their saving and restoring doesn't refer to uninitialized store.
4568
4569 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4570 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4571 fix the problem.
4572
4573 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4574 calls, and wasn't printing the correct value for compiling calls. Increased the
4575 default value of LOOPREPEAT, and the number of significant figures in the
4576 times.
4577
4578 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4579
4580 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4581 a building problem on Windows NT with a FAT file system.
4582
4583
4584 Version 2.01 21-Oct-98
4585 ----------------------
4586
4587 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4588 to character tables built by pcre_maketables() in the current locale. If NULL
4589 is passed, the default tables are used.
4590
4591
4592 Version 2.00 24-Sep-98
4593 ----------------------
4594
4595 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4596 it any more.
4597
4598 2. Allow quantification of (?>) groups, and make it work correctly.
4599
4600 3. The first character computation wasn't working for (?>) groups.
4601
4602 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4603 end of the subject) and add 5.005's \z, which really does match only at the
4604 very end of the subject.
4605
4606 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4607
4608 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4609 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4610 localized options. All options to pcre_study() were also removed.
4611
4612 7. Add other new features from 5.005:
4613
4614 $(?<= positive lookbehind
4615 $(?<! negative lookbehind
4616 (?imsx-imsx) added the unsetting capability
4617 such a setting is global if at outer level; local otherwise
4618 (?imsx-imsx:) non-capturing groups with option setting
4619 (?(cond)re|re) conditional pattern matching
4620
4621 A backreference to itself in a repeated group matches the previous
4622 captured string.
4623
4624 8. General tidying up of studying (both automatic and via "study")
4625 consequential on the addition of new assertions.
4626
4627 9. As in 5.005, unlimited repeated groups that could match an empty substring
4628 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4629 runtime if any iteration does actually match an empty substring.
4630
4631 10. Include the RunTest script in the distribution.
4632
4633 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4634 discrepancies, some of which were old and were also with respect to 5.004. They
4635 have now been fixed.
4636
4637
4638 Version 1.09 28-Apr-98
4639 ----------------------
4640
4641 1. A negated single character class followed by a quantifier with a minimum
4642 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4643 program crashes, or just wrong answers. This did not apply to negated classes
4644 containing more than one character, or to minima other than one.
4645
4646
4647 Version 1.08 27-Mar-98
4648 ----------------------
4649
4650 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4651
4652 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4653 latter must appear before anything that relies on it in the pattern.
4654
4655
4656 Version 1.07 16-Feb-98
4657 ----------------------
4658
4659 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4660 repeat of a potentially empty string).
4661
4662
4663 Version 1.06 23-Jan-98
4664 ----------------------
4665
4666 1. Added Markus Oberhumer's little patches for C++.
4667
4668 2. Literal strings longer than 255 characters were broken.
4669
4670
4671 Version 1.05 23-Dec-97
4672 ----------------------
4673
4674 1. Negated character classes containing more than one character were failing if
4675 PCRE_CASELESS was set at run time.
4676
4677
4678 Version 1.04 19-Dec-97
4679 ----------------------
4680
4681 1. Corrected the man page, where some "const" qualifiers had been omitted.
4682
4683 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4684 input syntax.
4685
4686 3. Fixed memory leak which occurred when a regex with back references was
4687 matched with an offsets vector that wasn't big enough. The temporary memory
4688 that is used in this case wasn't being freed if the match failed.
4689
4690 4. Tidied pcretest to ensure it frees memory that it gets.
4691
4692 5. Temporary memory was being obtained in the case where the passed offsets
4693 vector was exactly big enough.
4694
4695 6. Corrected definition of offsetof() from change 5 below.
4696
4697 7. I had screwed up change 6 below and broken the rules for the use of
4698 setjmp(). Now fixed.
4699
4700
4701 Version 1.03 18-Dec-97
4702 ----------------------
4703
4704 1. A erroneous regex with a missing opening parenthesis was correctly
4705 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4706 on some systems.
4707
4708 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4709 it was reported that one broken compiler failed on the former because "code" is
4710 also an independent variable.
4711
4712 3. The erroneous regex a[]b caused an array overrun reference.
4713
4714 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4715 fail on data ending with that character. (It was going on too far, and checking
4716 the next character, typically a binary zero.) This was specific to the
4717 optimized code for single-character negative classes.
4718
4719 5. Added a contributed patch from the TIN world which does the following:
4720
4721 + Add an undef for memmove, in case the the system defines a macro for it.
4722
4723 + Add a definition of offsetof(), in case there isn't one. (I don't know
4724 the reason behind this - offsetof() is part of the ANSI standard - but
4725 it does no harm).
4726
4727 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4728 most of the places where whitespace preceded '#'. I have given up and
4729 allowed the remaining 2 cases to be at the margin.
4730
4731 + Rename some variables in pcre to eliminate shadowing. This seems very
4732 pedantic, but does no harm, of course.
4733
4734 6. Moved the call to setjmp() into its own function, to get rid of warnings
4735 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4736
4737 7. Constructs such as \d{8,} were compiling into the equivalent of
4738 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4739 outcome, but in this particular case used more store than had been allocated,
4740 which caused the bug to be discovered because it threw up an internal error.
4741
4742 8. The debugging code in both pcre and pcretest for outputting the compiled
4743 form of a regex was going wrong in the case of back references followed by
4744 curly-bracketed repeats.
4745
4746
4747 Version 1.02 12-Dec-97
4748 ----------------------
4749
4750 1. Typos in pcre.3 and comments in the source fixed.
4751
4752 2. Applied a contributed patch to get rid of places where it used to remove
4753 'const' from variables, and fixed some signed/unsigned and uninitialized
4754 variable warnings.
4755
4756 3. Added the "runtest" target to Makefile.
4757
4758 4. Set default compiler flag to -O2 rather than just -O.
4759
4760
4761 Version 1.01 19-Nov-97
4762 ----------------------
4763
4764 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4765 like /([ab]*)*/, that is, for classes with more than one character in them.
4766
4767 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4768 as /((?>a*))*/ (a PCRE_EXTRA facility).
4769
4770
4771 Version 1.00 18-Nov-97
4772 ----------------------
4773
4774 1. Added compile-time macros to support systems such as SunOS4 which don't have
4775 memmove() or strerror() but have other things that can be used instead.
4776
4777 2. Arranged that "make clean" removes the executables.
4778
4779
4780 Version 0.99 27-Oct-97
4781 ----------------------
4782
4783 1. Fixed bug in code for optimizing classes with only one character. It was
4784 initializing a 32-byte map regardless, which could cause it to run off the end
4785 of the memory it had got.
4786
4787 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4788
4789
4790 Version 0.98 22-Oct-97
4791 ----------------------
4792
4793 1. Fixed bug in code for handling temporary memory usage when there are more
4794 back references than supplied space in the ovector. This could cause segfaults.
4795
4796
4797 Version 0.97 21-Oct-97
4798 ----------------------
4799
4800 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4801
4802 2. Optimized negated single characters not to use a bit map.
4803
4804 3. Brought error texts together as macro definitions; clarified some of them;
4805 fixed one that was wrong - it said "range out of order" when it meant "invalid
4806 escape sequence".
4807
4808 4. Changed some char * arguments to const char *.
4809
4810 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4811
4812 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4813 pcretest.
4814
4815
4816 Version 0.96 16-Oct-97
4817 ----------------------
4818
4819 1. Added a simple "pgrep" utility to the distribution.
4820
4821 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4822 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4823 where "ddd" means "one or more decimal digits".
4824
4825 3. Fixed serious bug. If a pattern had a back reference, but the call to
4826 pcre_exec() didn't supply a large enough ovector to record the related
4827 identifying subpattern, the match always failed. PCRE now remembers the number
4828 of the largest back reference, and gets some temporary memory in which to save
4829 the offsets during matching if necessary, in order to ensure that
4830 backreferences always work.
4831
4832 4. Increased the compatibility with Perl in a number of ways:
4833
4834 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4835 to request this handling. The option can be set at compile or exec time.
4836
4837 (b) $ matches before a terminating newline by default; an option
4838 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4839 mode). The option can be set at compile or exec time.
4840
4841 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4842 the same as Perl's. If the decimal number it represents is less than 10
4843 or there aren't that many previous left capturing parentheses, an octal
4844 escape is read. Inside a character class, it's always an octal escape,
4845 even if it is a single digit.
4846
4847 (d) An escaped but undefined alphabetic character is taken as a literal,
4848 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4849 escapes.
4850
4851 (e) {0} is now permitted. (The previous item is removed from the compiled
4852 pattern).
4853
4854 5. Changed all the names of code files so that the basic parts are no longer
4855 than 10 characters, and abolished the teeny "globals.c" file.
4856
4857 6. Changed the handling of character classes; they are now done with a 32-byte
4858 bit map always.
4859
4860 7. Added the -d and /D options to pcretest to make it possible to look at the
4861 internals of compilation without having to recompile pcre.
4862
4863
4864 Version 0.95 23-Sep-97
4865 ----------------------
4866
4867 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4868 \x20 at the start of a run of normal characters. These were being treated as
4869 real characters, instead of the source characters being re-checked.
4870
4871
4872 Version 0.94 18-Sep-97
4873 ----------------------
4874
4875 1. The functions are now thread-safe, with the caveat that the global variables
4876 containing pointers to malloc() and free() or alternative functions are the
4877 same for all threads.
4878
4879 2. Get pcre_study() to generate a bitmap of initial characters for non-
4880 anchored patterns when this is possible, and use it if passed to pcre_exec().
4881
4882
4883 Version 0.93 15-Sep-97
4884 ----------------------
4885
4886 1. /(b)|(:+)/ was computing an incorrect first character.
4887
4888 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4889 but not actually doing anything yet.
4890
4891 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4892 as Perl does (e.g. [-az] or [az-]).
4893
4894 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4895 all possible positions.
4896
4897 5. Split up into different modules to avoid including unneeded functions in a
4898 compiled binary. However, compile and exec are still in one module. The "study"
4899 function is split off.
4900
4901 6. The character tables are now in a separate module whose source is generated
4902 by an auxiliary program - but can then be edited by hand if required. There are
4903 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4904 toupper() in the code.
4905
4906 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4907 make them global. Abolish the function for setting them, as the caller can now
4908 set them directly.
4909
4910
4911 Version 0.92 11-Sep-97
4912 ----------------------
4913
4914 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4915 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4916
4917 2. Caseless matching was not working in character classes if the characters in
4918 the pattern were in upper case.
4919
4920 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4921
4922 4. Make PCRE_ANCHORED public and accept as a compile option.
4923
4924 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4925 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4926 pass them.
4927
4928 6. Give an error if bad option bits passed at compile or run time.
4929
4930 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4931 pcretest to cause it to pass that flag.
4932
4933 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4934 options, and the first character, if set.
4935
4936 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4937
4938
4939 Version 0.91 10-Sep-97
4940 ----------------------
4941
4942 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4943 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4944
4945 2. PCRE was looping on encountering an indefinitely repeated back reference to
4946 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4947 Perl does - treats the match as successful.
4948
4949 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5