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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1250 - (show annotations)
Wed Feb 20 09:59:29 2013 UTC (6 years, 9 months ago) by zherczeg
File size: 242406 byte(s)
JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.33 xx-xxxx-201x
5 -------------------------
6
7 1. Added 'U' to some constants that are compared to unsigned integers, to
8 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9 variables that are added to signed variables, to ensure the result is
10 signed and can be negated.
11
12 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
13 CMake config files.
14
15 3. Revise the creation of config.h.generic so that all boolean macros are
16 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17 overriding via -D on the command line possible.
18
19 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20 to unsigned int is reported to make a quite noticeable speed difference in
21 a specific Windows environment. Testing on Linux did also appear to show
22 some benefit (and it is clearly not harmful). Also fixed the definition of
23 Xop which should be unsigned.
24
25 5. Related to (4), changing the definition of the intermediate variable cc
26 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27 improvements.
28
29 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
30 unnecessary spaces.
31
32 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33 and later.
34
35 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36
37 9. Optimizing fast_forward_start_bits in JIT.
38
39 10. Adding experimental support for callouts in JIT, and fixing some
40 issues revealed during this work. Namely:
41
42 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43
44 (b) Minimum length was not checked before the matching is started.
45
46 11. The value of capture_last that is passed to callouts was incorrect in some
47 cases when there was a capture on one path that was subsequently abandoned
48 after a backtrack. Also, the capture_last value is now reset after a
49 recursion, since all captures are also reset in this case.
50
51 12. The interpreter no longer returns the "too many substrings" error in the
52 case when an overflowing capture is in a branch that is subsequently
53 abandoned after a backtrack.
54
55 13. In the pathological case when an offset vector of size 2 is used, pcretest
56 now prints out the matched string after a yield of 0 or 1.
57
58 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59 Only supported by the JIT compiler at the moment.
60
61 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
62
63
64 Version 8.32 30-November-2012
65 -----------------------------
66
67 1. Improved JIT compiler optimizations for first character search and single
68 character iterators.
69
70 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
71 Patch by Daniel Richard G.
72
73 3. Single character iterator optimizations in the JIT compiler.
74
75 4. Improved JIT compiler optimizations for character ranges.
76
77 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
78 Reported by Giuseppe D'Angelo.
79
80 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
81 of a line, was being set incorrectly in cases where .* appeared inside
82 atomic brackets at the start of a pattern, or where there was a subsequent
83 *PRUNE or *SKIP.
84
85 7. Improved instruction cache flush for POWER/PowerPC.
86 Patch by Daniel Richard G.
87
88 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
89 grep:
90
91 (a) There is now no limit to the number of patterns to be matched.
92
93 (b) An error is given if a pattern is too long.
94
95 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
96 are now supported.
97
98 (d) --exclude-from and --include-from (multiple use) have been added.
99
100 (e) Exclusions and inclusions now apply to all files and directories, not
101 just to those obtained from scanning a directory recursively.
102
103 (f) Multiple uses of -f and --file-list are now supported.
104
105 (g) In a Windows environment, the default for -d has been changed from
106 "read" (the GNU grep default) to "skip", because otherwise the presence
107 of a directory in the file list provokes an error.
108
109 (h) The documentation has been revised and clarified in places.
110
111 9. Improve the matching speed of capturing brackets.
112
113 10. Changed the meaning of \X so that it now matches a Unicode extended
114 grapheme cluster.
115
116 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
117 out POSIX threads when JIT support is configured.
118
119 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
120
121 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
122 structure could go wrong in environments where size_t is not the same size
123 as int.
124
125 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
126 set.
127
128 15. The EBCDIC support had decayed; later updates to the code had included
129 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
130 general tidy up of EBCDIC-related issues, and the documentation was also
131 not quite right. There is now a test that can be run on ASCII systems to
132 check some of the EBCDIC-related things (but is it not a full test).
133
134 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
135 in a small tidy to the code.
136
137 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
138
139 18. If the --only-matching (-o) option in pcregrep is specified multiple
140 times, each one causes appropriate output. For example, -o1 -o2 outputs the
141 substrings matched by the 1st and 2nd capturing parentheses. A separating
142 string can be specified by --om-separator (default empty).
143
144 19. Improving the first n character searches.
145
146 20. Turn case lists for horizontal and vertical white space into macros so that
147 they are defined only once.
148
149 21. This set of changes together give more compatible Unicode case-folding
150 behaviour for characters that have more than one other case when UCP
151 support is available.
152
153 (a) The Unicode property table now has offsets into a new table of sets of
154 three or more characters that are case-equivalent. The MultiStage2.py
155 script that generates these tables (the pcre_ucd.c file) now scans
156 CaseFolding.txt instead of UnicodeData.txt for character case
157 information.
158
159 (b) The code for adding characters or ranges of characters to a character
160 class has been abstracted into a generalized function that also handles
161 case-independence. In UTF-mode with UCP support, this uses the new data
162 to handle characters with more than one other case.
163
164 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
165 whose other case is greater than 256 are now correctly matched
166 caselessly. Previously, the high codepoint matched the low one, but not
167 vice versa.
168
169 (d) The processing of \h, \H, \v, and \ in character classes now makes use
170 of the new class addition function, using character lists defined as
171 macros alongside the case definitions of 20 above.
172
173 (e) Caseless back references now work with characters that have more than
174 one other case.
175
176 (f) General caseless matching of characters with more than one other case
177 is supported.
178
179 22. Unicode character properties were updated from Unicode 6.2.0
180
181 23. Improved CMake support under Windows. Patch by Daniel Richard G.
182
183 24. Add support for 32-bit character strings, and UTF-32
184
185 25. Major JIT compiler update (code refactoring and bugfixing).
186 Experimental Sparc 32 support is added.
187
188 26. Applied a modified version of Daniel Richard G's patch to create
189 pcre.h.generic and config.h.generic by "make" instead of in the
190 PrepareRelease script.
191
192 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
193 pcre_compile.c when checking for a zero character.
194
195 28. Introducing a native interface for JIT. Through this interface, the compiled
196 machine code can be directly executed. The purpose of this interface is to
197 provide fast pattern matching, so several sanity checks are not performed.
198 However, feature tests are still performed. The new interface provides
199 1.4x speedup compared to the old one.
200
201 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
202 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
203 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
204
205 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
206 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
207 the "old" RFC 2279). Instead, it ended up passing a negative length to
208 pcre_exec().
209
210 31. Add support for GCC's visibility feature to hide internal functions.
211
212 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
213 "unknown -C option" after outputting 0 or 1.
214
215 33. There is now support for generating a code coverage report for the test
216 suite in environments where gcc is the compiler and lcov is installed. This
217 is mainly for the benefit of the developers.
218
219 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
220 unaddressable using valgrind annotations, allowing valgrind to detect
221 invalid memory accesses. This is mainly for the benefit of the developers.
222
223 25. (*UTF) can now be used to start a pattern in any of the three libraries.
224
225 26. Give configure error if --enable-cpp but no C++ compiler found.
226
227
228 Version 8.31 06-July-2012
229 -------------------------
230
231 1. Fixing a wrong JIT test case and some compiler warnings.
232
233 2. Removed a bashism from the RunTest script.
234
235 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
236 to unsigned type, result still unsigned" that was given by an MS compiler
237 on encountering the code "-sizeof(xxx)".
238
239 4. Partial matching support is added to the JIT compiler.
240
241 5. Fixed several bugs concerned with partial matching of items that consist
242 of more than one character:
243
244 (a) /^(..)\1/ did not partially match "aba" because checking references was
245 done on an "all or nothing" basis. This also applied to repeated
246 references.
247
248 (b) \R did not give a hard partial match if \r was found at the end of the
249 subject.
250
251 (c) \X did not give a hard partial match after matching one or more
252 characters at the end of the subject.
253
254 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
255 a partial match for the string "\r".
256
257 (e) When newline was set to CRLF, the metacharacter "." did not recognize
258 a partial match for a CR character at the end of the subject string.
259
260 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
261 running pcretest, the text "(JIT)" added to the output whenever JIT is
262 actually used to run the match.
263
264 7. Individual JIT compile options can be set in pcretest by following -s+[+]
265 or /S+[+] with a digit between 1 and 7.
266
267 8. OP_NOT now supports any UTF character not just single-byte ones.
268
269 9. (*MARK) control verb is now supported by the JIT compiler.
270
271 10. The command "./RunTest list" lists the available tests without actually
272 running any of them. (Because I keep forgetting what they all are.)
273
274 11. Add PCRE_INFO_MAXLOOKBEHIND.
275
276 12. Applied a (slightly modified) user-supplied patch that improves performance
277 when the heap is used for recursion (compiled with --disable-stack-for-
278 recursion). Instead of malloc and free for each heap frame each time a
279 logical recursion happens, frames are retained on a chain and re-used where
280 possible. This sometimes gives as much as 30% improvement.
281
282 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
283 call.
284
285 14. As documented, (*COMMIT) is now confined to within a positive assertion.
286
287 15. It is now possible to link pcretest with libedit as an alternative to
288 libreadline.
289
290 16. (*COMMIT) control verb is now supported by the JIT compiler.
291
292 17. The Unicode data tables have been updated to Unicode 6.1.0.
293
294 18. Added --file-list option to pcregrep.
295
296 19. Added binary file support to pcregrep, including the -a, --binary-files,
297 -I, and --text options.
298
299 20. The madvise function is renamed for posix_madvise for QNX compatibility
300 reasons. Fixed by Giuseppe D'Angelo.
301
302 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
303
304 22. Changed the option for creating symbolic links for 16-bit man pages from
305 -s to -sf so that re-installing does not cause issues.
306
307 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
308
309 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
310 matches in certain environments (the workspace was not being correctly
311 retained). Also added to pcre_dfa_exec() a simple plausibility check on
312 some of the workspace data at the beginning of a restart.
313
314 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
315 was not doing so when it should - probably a typo introduced by SVN 528
316 (change 8.10/14).
317
318 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
319 \w+ when the character tables indicated that \x{c4} was a word character.
320 There were several related cases, all because the tests for doing a table
321 lookup were testing for characters less than 127 instead of 255.
322
323 27. If a pattern contains capturing parentheses that are not used in a match,
324 their slots in the ovector are set to -1. For those that are higher than
325 any matched groups, this happens at the end of processing. In the case when
326 there were back references that the ovector was too small to contain
327 (causing temporary malloc'd memory to be used during matching), and the
328 highest capturing number was not used, memory off the end of the ovector
329 was incorrectly being set to -1. (It was using the size of the temporary
330 memory instead of the true size.)
331
332 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
333 ovector size, it uses memory at the end of the block that it has got.
334
335 29. Check for an overlong MARK name and give an error at compile time. The
336 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
337
338 30. JIT compiler update.
339
340 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
341 Rill for the patch.
342
343 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
344
345 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
346
347 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
348 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
349 was enabled.
350
351 35. Improve JIT code generation for greedy plus quantifier.
352
353 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
354 1 to "aa" instead of to an empty string. The bug affected repeated groups
355 that could potentially match an empty string.
356
357 37. Optimizing single character iterators in JIT.
358
359 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
360 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
361 codepoints that are too big for the mode are faulted, and in a UTF mode,
362 disallowed codepoints are also faulted.
363
364 39. If PCRE was compiled with UTF support, in three places in the DFA
365 matcher there was code that should only have been obeyed in UTF mode, but
366 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
367 processing when bytes with values greater than 127 were present. In 16-bit
368 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
369 both cases the values are those that cannot be the first data item in a UTF
370 character. The three items that might have provoked this were recursions,
371 possessively repeated groups, and atomic groups.
372
373 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
374 and pcregrep, because some OS require shared objects to be explicitly
375 passed to ld, causing the link step to fail if they are not.
376
377 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
378 mode, patterns that started with \h* or \R* might be incorrectly matched.
379
380
381 Version 8.30 04-February-2012
382 -----------------------------
383
384 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
385 name is defined in ctype.h.
386
387 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
388 only in quite long subpatterns.
389
390 3. Removed the function pcre_info(), which has been obsolete and deprecated
391 since it was replaced by pcre_fullinfo() in February 2000.
392
393 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
394 match a (*MARK), and the match failed at the start of the subject, a
395 reference to memory before the start of the subject could occur. This bug
396 was introduced by fix 17 of release 8.21.
397
398 5. A reference to an unset group with zero minimum repetition was giving
399 totally wrong answers (in non-JavaScript-compatibility mode). For example,
400 /(another)?(\1?)test/ matched against "hello world test". This bug was
401 introduced in release 8.13.
402
403 6. Add support for 16-bit character strings (a large amount of work involving
404 many changes and refactorings).
405
406 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
407 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
408 from a file.
409
410 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
411 rounding is not applied in this particular case).
412
413 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
414 if they appear, or are escaped, in patterns.
415
416 10. Get rid of a number of -Wunused-but-set-variable warnings.
417
418 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
419 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
420 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
421 also returns the mark "x". This bug applied to capturing parentheses,
422 non-capturing parentheses, and atomic parentheses. It also applied to some
423 assertions.
424
425 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
426 information out of configure.ac instead of relying on pcre.h.generic, which
427 is not stored in the repository.
428
429 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
430 -lreadline.
431
432 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
433
434 15. Applied Graycode's patch to put the top-level frame on the stack rather
435 than the heap when not using the stack for recursion. This gives a
436 performance improvement in many cases when recursion is not deep.
437
438 16. Experimental code added to "pcretest -C" to output the stack frame size.
439
440
441 Version 8.21 12-Dec-2011
442 ------------------------
443
444 1. Updating the JIT compiler.
445
446 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
447 are added as well.
448
449 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
450 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
451 calling _pcre_jit_exec. Some extra comments are added.
452
453 4. (*MARK) settings inside atomic groups that do not contain any capturing
454 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
455 was introduced by change 18 for 8.20.
456
457 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
458 ECMA-262 standard.
459
460 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
461 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
462 This bug was probably introduced by change 9 of 8.13.
463
464 7. While fixing 6 above, I noticed that a number of other items were being
465 incorrectly rejected as "not fixed length". This arose partly because newer
466 opcodes had not been added to the fixed-length checking code. I have (a)
467 corrected the bug and added tests for these items, and (b) arranged for an
468 error to occur if an unknown opcode is encountered while checking for fixed
469 length instead of just assuming "not fixed length". The items that were
470 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
471 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
472 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
473
474 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
475 being incorrectly compiled and would have given unpredicatble results.
476
477 9. A possessively repeated subpattern with minimum repeat count greater than
478 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
479 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
480 the first (A) could occur when it should not.
481
482 10. Add a cast and remove a redundant test from the code.
483
484 11. JIT should use pcre_malloc/pcre_free for allocation.
485
486 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
487 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
488 is anything other than /usr, -L is still shown).
489
490 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
491
492 14. Perl does not support \N without a following name in a [] class; PCRE now
493 also gives an error.
494
495 15. If a forward reference was repeated with an upper limit of around 2000,
496 it caused the error "internal error: overran compiling workspace". The
497 maximum number of forward references (including repeats) was limited by the
498 internal workspace, and dependent on the LINK_SIZE. The code has been
499 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
500 the default depends on LINK_SIZE. There is a new upper limit (for safety)
501 of around 200,000 forward references. While doing this, I also speeded up
502 the filling in of repeated forward references.
503
504 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
505 incorrectly expecting the subject to contain another "a" after the start.
506
507 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
508 in the match, the SKIP should be ignored. This was not happening; instead
509 the SKIP was being treated as NOMATCH. For patterns such as
510 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
511 tested.
512
513 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
514 now much more compatible with Perl, in particular in cases where the result
515 is a non-match for a non-anchored pattern. For example, if
516 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
517 "m", where previously it did not return a name. A side effect of this
518 change is that for partial matches, the last encountered mark name is
519 returned, as for non matches. A number of tests that were previously not
520 Perl-compatible have been moved into the Perl-compatible test files. The
521 refactoring has had the pleasing side effect of removing one argument from
522 the match() function, thus reducing its stack requirements.
523
524 19. If the /S+ option was used in pcretest to study a pattern using JIT,
525 subsequent uses of /S (without +) incorrectly behaved like /S+.
526
527 21. Retrieve executable code size support for the JIT compiler and fixing
528 some warnings.
529
530 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
531 not work when the shorter character appeared right at the end of the
532 subject string.
533
534 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
535 systems.
536
537 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
538 output it when the /M option is used in pcretest.
539
540 25. The CheckMan script was not being included in the distribution. Also, added
541 an explicit "perl" to run Perl scripts from the PrepareRelease script
542 because this is reportedly needed in Windows.
543
544 26. If study data was being save in a file and studying had not found a set of
545 "starts with" bytes for the pattern, the data written to the file (though
546 never used) was taken from uninitialized memory and so caused valgrind to
547 complain.
548
549 27. Updated RunTest.bat as provided by Sheri Pierce.
550
551 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
552
553 29. Computation of memory usage for the table of capturing group names was
554 giving an unnecessarily large value.
555
556
557 Version 8.20 21-Oct-2011
558 ------------------------
559
560 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
561 a POSIX class. After further experiments with Perl, which convinced me that
562 Perl has bugs and confusions, a closing square bracket is no longer allowed
563 in a POSIX name. This bug also affected patterns with classes that started
564 with full stops.
565
566 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
567 captured substring, but while checking the failing first alternative,
568 substring 1 is temporarily captured. If the output vector supplied to
569 pcre_exec() was not big enough for this capture, the yield of the function
570 was still zero ("insufficient space for captured substrings"). This cannot
571 be totally fixed without adding another stack variable, which seems a lot
572 of expense for a edge case. However, I have improved the situation in cases
573 such as /(a)(b)x|abc/ matched against "abc", where the return code
574 indicates that fewer than the maximum number of slots in the ovector have
575 been set.
576
577 3. Related to (2) above: when there are more back references in a pattern than
578 slots in the output vector, pcre_exec() uses temporary memory during
579 matching, and copies in the captures as far as possible afterwards. It was
580 using the entire output vector, but this conflicts with the specification
581 that only 2/3 is used for passing back captured substrings. Now it uses
582 only the first 2/3, for compatibility. This is, of course, another edge
583 case.
584
585 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
586 main code base, and can be used by building with --enable-jit. When this is
587 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
588 runtime --no-jit option is given.
589
590 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
591 ovector, the return from the function was zero, implying that there were
592 other matches that did not fit. The correct "exactly full" value is now
593 returned.
594
595 6. If a subpattern that was called recursively or as a subroutine contained
596 (*PRUNE) or any other control that caused it to give a non-standard return,
597 invalid errors such as "Error -26 (nested recursion at the same subject
598 position)" or even infinite loops could occur.
599
600 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
601 computing the minimum length on reaching *ACCEPT, and so ended up with the
602 wrong value of 1 rather than 0. Further investigation indicates that
603 computing a minimum subject length in the presence of *ACCEPT is difficult
604 (think back references, subroutine calls), and so I have changed the code
605 so that no minimum is registered for a pattern that contains *ACCEPT.
606
607 8. If (*THEN) was present in the first (true) branch of a conditional group,
608 it was not handled as intended. [But see 16 below.]
609
610 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
611 Sheri Pierce.
612
613 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
614 the first byte in a match must be "a".
615
616 11. Change 17 for 8.13 increased the recursion depth for patterns like
617 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
618 pattern contains any instances of (*THEN). If it does not, the old
619 optimizations are restored. It would be nice to do this on a per-group
620 basis, but at the moment that is not feasible.
621
622 12. In some environments, the output of pcretest -C is CRLF terminated. This
623 broke RunTest's code that checks for the link size. A single white space
624 character after the value is now allowed for.
625
626 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
627 For "fr", it uses the Windows-specific input and output files.
628
629 14. If (*THEN) appeared in a group that was called recursively or as a
630 subroutine, it did not work as intended. [But see next item.]
631
632 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
633 pattern fragments (but not containing any | characters). If A and B are
634 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
635 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
636 D. In other words, Perl considers parentheses that do not contain any |
637 characters to be part of a surrounding alternative, whereas PCRE was
638 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
639 differently. PCRE now behaves in the same way as Perl, except in the case
640 of subroutine/recursion calls such as (?1) which have in any case always
641 been different (but PCRE had them first :-).
642
643 16. Related to 15 above: Perl does not treat the | in a conditional group as
644 creating alternatives. Such a group is treated in the same way as an
645 ordinary group without any | characters when processing (*THEN). PCRE has
646 been changed to match Perl's behaviour.
647
648 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
649 RunGrepTest script failed.
650
651 18. Change 22 for version 13 caused atomic groups to use more stack. This is
652 inevitable for groups that contain captures, but it can lead to a lot of
653 stack use in large patterns. The old behaviour has been restored for atomic
654 groups that do not contain any capturing parentheses.
655
656 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
657 suppress the check for a minimum subject length at run time. (If it was
658 given to pcre_exec() or pcre_dfa_exec() it did work.)
659
660 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
661 fail to work when decoding hex characters in data strings in EBCDIC
662 environments.
663
664 21. It appears that in at least one Mac OS environment, the isxdigit() function
665 is implemented as a macro that evaluates to its argument more than once,
666 contravening the C 90 Standard (I haven't checked a later standard). There
667 was an instance in pcretest which caused it to go wrong when processing
668 \x{...} escapes in subject strings. The has been rewritten to avoid using
669 things like p++ in the argument of isxdigit().
670
671
672 Version 8.13 16-Aug-2011
673 ------------------------
674
675 1. The Unicode data tables have been updated to Unicode 6.0.0.
676
677 2. Two minor typos in pcre_internal.h have been fixed.
678
679 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
680 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
681 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
682
683 4. There were a number of related bugs in the code for matching backrefences
684 caselessly in UTF-8 mode when codes for the characters concerned were
685 different numbers of bytes. For example, U+023A and U+2C65 are an upper
686 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
687 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
688 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
689 2-byte code at the end of the subject (it thought there wasn't enough data
690 left).
691
692 5. Comprehensive information about what went wrong is now returned by
693 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
694 as the output vector has at least 2 elements. The offset of the start of
695 the failing character and a reason code are placed in the vector.
696
697 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
698 now returned is for the first byte of the failing character, instead of the
699 last byte inspected. This is an incompatible change, but I hope it is small
700 enough not to be a problem. It makes the returned offset consistent with
701 pcre_exec() and pcre_dfa_exec().
702
703 7. pcretest now gives a text phrase as well as the error number when
704 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
705 failure, the offset and reason code are output.
706
707 8. When \R was used with a maximizing quantifier it failed to skip backwards
708 over a \r\n pair if the subsequent match failed. Instead, it just skipped
709 back over a single character (\n). This seems wrong (because it treated the
710 two characters as a single entity when going forwards), conflicts with the
711 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
712 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
713 has been changed.
714
715 9. Some internal refactoring has changed the processing so that the handling
716 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
717 time (the PCRE_DOTALL option was changed this way some time ago: version
718 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
719 which was always a bit of a fudge. It also means that there is one less
720 argument for the match() function, which reduces its stack requirements
721 slightly. This change also fixes an incompatibility with Perl: the pattern
722 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
723
724 10. More internal refactoring has drastically reduced the number of recursive
725 calls to match() for possessively repeated groups such as (abc)++ when
726 using pcre_exec().
727
728 11. While implementing 10, a number of bugs in the handling of groups were
729 discovered and fixed:
730
731 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
732 (a|)*(?1) gave a compile-time internal error.
733 ((a|)+)+ did not notice that the outer group could match an empty string.
734 (^a|^)+ was not marked as anchored.
735 (.*a|.*)+ was not marked as matching at start or after a newline.
736
737 12. Yet more internal refactoring has removed another argument from the match()
738 function. Special calls to this function are now indicated by setting a
739 value in a variable in the "match data" data block.
740
741 13. Be more explicit in pcre_study() instead of relying on "default" for
742 opcodes that mean there is no starting character; this means that when new
743 ones are added and accidentally left out of pcre_study(), testing should
744 pick them up.
745
746 14. The -s option of pcretest has been documented for ages as being an old
747 synonym of -m (show memory usage). I have changed it to mean "force study
748 for every regex", that is, assume /S for every regex. This is similar to -i
749 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
750 using it. It makes it easier to run collections of tests with and without
751 study enabled, and thereby test pcre_study() more easily. All the standard
752 tests are now run with and without -s (but some patterns can be marked as
753 "never study" - see 20 below).
754
755 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
756 restoration of the capturing data to the outer values was not happening
757 correctly.
758
759 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
760 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
761 pattern had matched an empty string, and so incorrectly returned a no
762 match.
763
764 17. There was optimizing code for the last branch of non-capturing parentheses,
765 and also for the obeyed branch of a conditional subexpression, which used
766 tail recursion to cut down on stack usage. Unfortunately, now that there is
767 the possibility of (*THEN) occurring in these branches, tail recursion is
768 no longer possible because the return has to be checked for (*THEN). These
769 two optimizations have therefore been removed. [But see 8.20/11 above.]
770
771 18. If a pattern containing \R was studied, it was assumed that \R always
772 matched two bytes, thus causing the minimum subject length to be
773 incorrectly computed because \R can also match just one byte.
774
775 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
776 was incorrectly computed.
777
778 20. If /S is present twice on a test pattern in pcretest input, it now
779 *disables* studying, thereby overriding the use of -s on the command line
780 (see 14 above). This is necessary for one or two tests to keep the output
781 identical in both cases.
782
783 21. When (*ACCEPT) was used in an assertion that matched an empty string and
784 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
785
786 22. When an atomic group that contained a capturing parenthesis was
787 successfully matched, but the branch in which it appeared failed, the
788 capturing was not being forgotten if a higher numbered group was later
789 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
790 group 1 to "a", when in fact it should be unset. This applied to multi-
791 branched capturing and non-capturing groups, repeated or not, and also to
792 positive assertions (capturing in negative assertions does not happen
793 in PCRE) and also to nested atomic groups.
794
795 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
796 subject after a captured substring, to make it easier to tell which of a
797 number of identical substrings has been captured.
798
799 24. The way atomic groups are processed by pcre_exec() has been changed so that
800 if they are repeated, backtracking one repetition now resets captured
801 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
802 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
803 "aaa". Previously, it would have been "a". As part of this code
804 refactoring, the way recursive calls are handled has also been changed.
805
806 25. If an assertion condition captured any substrings, they were not passed
807 back unless some other capturing happened later. For example, if
808 (?(?=(a))a) was matched against "a", no capturing was returned.
809
810 26. When studying a pattern that contained subroutine calls or assertions,
811 the code for finding the minimum length of a possible match was handling
812 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
813 group 1 called group 2 while simultaneously a separate group 2 called group
814 1). A stack overflow occurred in this case. I have fixed this by limiting
815 the recursion depth to 10.
816
817 27. Updated RunTest.bat in the distribution to the version supplied by Tom
818 Fortmann. This supports explicit test numbers on the command line, and has
819 argument validation and error reporting.
820
821 28. An instance of \X with an unlimited repeat could fail if at any point the
822 first character it looked at was a mark character.
823
824 29. Some minor code refactoring concerning Unicode properties and scripts
825 should reduce the stack requirement of match() slightly.
826
827 30. Added the '=' option to pcretest to check the setting of unused capturing
828 slots at the end of the pattern, which are documented as being -1, but are
829 not included in the return count.
830
831 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
832 compiled something random. Now it gives a compile-time error (as does
833 Perl).
834
835 32. A *MARK encountered during the processing of a positive assertion is now
836 recorded and passed back (compatible with Perl).
837
838 33. If --only-matching or --colour was set on a pcregrep call whose pattern
839 had alternative anchored branches, the search for a second match in a line
840 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
841 matched the line "0102" twice. The same bug affected patterns that started
842 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
843 twice.
844
845 34. Previously, PCRE did not allow quantification of assertions. However, Perl
846 does, and because of capturing effects, quantifying parenthesized
847 assertions may at times be useful. Quantifiers are now allowed for
848 parenthesized assertions.
849
850 35. A minor code tidy in pcre_compile() when checking options for \R usage.
851
852 36. \g was being checked for fancy things in a character class, when it should
853 just be a literal "g".
854
855 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
856 appearance of a nested POSIX class supersedes an apparent external class.
857 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
858 unescaped square brackets may also appear as part of class names. For
859 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
860 more like Perl. (But see 8.20/1 above.)
861
862 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
863 was because it thought it was \N{name}, which is not supported).
864
865 39. Add minix to OS list not supporting the -S option in pcretest.
866
867 40. PCRE tries to detect cases of infinite recursion at compile time, but it
868 cannot analyze patterns in sufficient detail to catch mutual recursions
869 such as ((?1))((?2)). There is now a runtime test that gives an error if a
870 subgroup is called recursively as a subpattern for a second time at the
871 same position in the subject string. In previous releases this might have
872 been caught by the recursion limit, or it might have run out of stack.
873
874 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
875 happen only once. PCRE was, however incorrectly giving a compile time error
876 "recursive call could loop indefinitely" because it cannot analyze the
877 pattern in sufficient detail. The compile time test no longer happens when
878 PCRE is compiling a conditional subpattern, but actual runaway loops are
879 now caught at runtime (see 40 above).
880
881 42. It seems that Perl allows any characters other than a closing parenthesis
882 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
883 has been changed to be the same.
884
885 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
886 as not to get warnings when autogen.sh is called. Also changed
887 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
888
889 44. To help people who use pcregrep to scan files containing exceedingly long
890 lines, the following changes have been made:
891
892 (a) The default value of the buffer size parameter has been increased from
893 8K to 20K. (The actual buffer used is three times this size.)
894
895 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
896 PCRE is built.
897
898 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
899 to be set at run time.
900
901 (d) Numerical values in pcregrep options can be followed by K or M, for
902 example --buffer-size=50K.
903
904 (e) If a line being scanned overflows pcregrep's buffer, an error is now
905 given and the return code is set to 2.
906
907 45. Add a pointer to the latest mark to the callout data block.
908
909 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
910 partial match of an empty string instead of no match. This was specific to
911 the use of ".".
912
913 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
914 complete match instead of a partial match. This bug was dependent on both
915 the PCRE_UTF8 and PCRE_DOTALL options being set.
916
917 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
918 starting byte set, because \b was not being ignored.
919
920
921 Version 8.12 15-Jan-2011
922 ------------------------
923
924 1. Fixed some typos in the markup of the man pages, and wrote a script that
925 checks for such things as part of the documentation building process.
926
927 2. On a big-endian 64-bit system, pcregrep did not correctly process the
928 --match-limit and --recursion-limit options (added for 8.11). In
929 particular, this made one of the standard tests fail. (The integer value
930 went into the wrong half of a long int.)
931
932 3. If the --colour option was given to pcregrep with -v (invert match), it
933 did strange things, either producing crazy output, or crashing. It should,
934 of course, ignore a request for colour when reporting lines that do not
935 match.
936
937 4. Another pcregrep bug caused similar problems if --colour was specified with
938 -M (multiline) and the pattern match finished with a line ending.
939
940 5. In pcregrep, when a pattern that ended with a literal newline sequence was
941 matched in multiline mode, the following line was shown as part of the
942 match. This seems wrong, so I have changed it.
943
944 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
945 the check for further matches in the same line (so they could be coloured)
946 to overrun the end of the current line. If another match was found, it was
947 incorrectly shown (and then shown again when found in the next line).
948
949 7. If pcregrep was compiled under Windows, there was a reference to the
950 function pcregrep_exit() before it was defined. I am assuming this was
951 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
952 reported by a user. I've moved the definition above the reference.
953
954
955 Version 8.11 10-Dec-2010
956 ------------------------
957
958 1. (*THEN) was not working properly if there were untried alternatives prior
959 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
960 backtracked to try for "b" instead of moving to the next alternative branch
961 at the same level (in this case, to look for "c"). The Perl documentation
962 is clear that when (*THEN) is backtracked onto, it goes to the "next
963 alternative in the innermost enclosing group".
964
965 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
966 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
967 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
968 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
969 (*THEN).
970
971 3. If \s appeared in a character class, it removed the VT character from
972 the class, even if it had been included by some previous item, for example
973 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
974 of \s, but is part of the POSIX "space" class.)
975
976 4. A partial match never returns an empty string (because you can always
977 match an empty string at the end of the subject); however the checking for
978 an empty string was starting at the "start of match" point. This has been
979 changed to the "earliest inspected character" point, because the returned
980 data for a partial match starts at this character. This means that, for
981 example, /(?<=abc)def/ gives a partial match for the subject "abc"
982 (previously it gave "no match").
983
984 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
985 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
986 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
987 has an implication that the given string is incomplete (because a partial
988 match is preferred over a full match). For this reason, these items now
989 give a partial match in this situation. [Aside: previously, the one case
990 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
991 match rather than a full match, which was wrong by the old rules, but is
992 now correct.]
993
994 6. There was a bug in the handling of #-introduced comments, recognized when
995 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
996 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
997 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
998 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
999 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1000 places in pcre_compile().
1001
1002 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1003 comments when looking ahead for named forward references to subpatterns,
1004 the only newline sequence it recognized was NL. It now handles newlines
1005 according to the set newline convention.
1006
1007 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1008 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1009 cater for a lack of strerror(). These oversights have been fixed.
1010
1011 9. Added --match-limit and --recursion-limit to pcregrep.
1012
1013 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1014
1015 11. When the -o option was used, pcregrep was setting a return code of 1, even
1016 when matches were found, and --line-buffered was not being honoured.
1017
1018 12. Added an optional parentheses number to the -o and --only-matching options
1019 of pcregrep.
1020
1021 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1022 can match an empty string. The code to do it in pcretest and pcredemo
1023 needed fixing:
1024
1025 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1026 only one byte after an empty string match just before CRLF (this case
1027 just got forgotten; "any" and "anycrlf" were OK).
1028
1029 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1030 mode when an empty string match preceded an ASCII character followed by
1031 a non-ASCII character. (The code for advancing by one character rather
1032 than one byte was nonsense.)
1033
1034 (c) The pcredemo.c sample program did not have any code at all to handle
1035 the cases when CRLF is a valid newline sequence.
1036
1037 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1038 as a starting offset was within the subject string. There is now a new
1039 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1040 negative or greater than the length of the string. In order to test this,
1041 pcretest is extended to allow the setting of negative starting offsets.
1042
1043 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1044 starting offset points to the beginning of a UTF-8 character was
1045 unnecessarily clumsy. I tidied it up.
1046
1047 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1048 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1049
1050 17. Nobody had reported that the --include_dir option, which was added in
1051 release 7.7 should have been called --include-dir (hyphen, not underscore)
1052 for compatibility with GNU grep. I have changed it to --include-dir, but
1053 left --include_dir as an undocumented synonym, and the same for
1054 --exclude-dir, though that is not available in GNU grep, at least as of
1055 release 2.5.4.
1056
1057 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1058 characters from a string of bytes) have been redefined so as not to use
1059 loops, in order to improve performance in some environments. At the same
1060 time, I abstracted some of the common code into auxiliary macros to save
1061 repetition (this should not affect the compiled code).
1062
1063 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1064 compile-time error is now given if \c is not followed by an ASCII
1065 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1066 different, and any byte value is allowed.)
1067
1068 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1069 START_OPTIMIZE option, which is now allowed at compile time - but just
1070 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1071 to pcregrep and other applications that have no direct access to PCRE
1072 options. The new /Y option in pcretest sets this option when calling
1073 pcre_compile().
1074
1075 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1076 back references. Groups containing recursive back references were forced to
1077 be atomic by that change, but in the case of named groups, the amount of
1078 memory required was incorrectly computed, leading to "Failed: internal
1079 error: code overflow". This has been fixed.
1080
1081 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1082 pcretest.c, to avoid build problems in some Borland environments.
1083
1084
1085 Version 8.10 25-Jun-2010
1086 ------------------------
1087
1088 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1089 THEN.
1090
1091 2. (*ACCEPT) was not working when inside an atomic group.
1092
1093 3. Inside a character class, \B is treated as a literal by default, but
1094 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1095 causes the error). The code is unchanged, but I tidied the documentation.
1096
1097 4. Inside a character class, PCRE always treated \R and \X as literals,
1098 whereas Perl faults them if its -w option is set. I have changed PCRE so
1099 that it faults them when PCRE_EXTRA is set.
1100
1101 5. Added support for \N, which always matches any character other than
1102 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1103
1104 6. When compiling pcregrep with newer versions of gcc which may have
1105 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1106 declared with attribute warn_unused_result" were given. Just casting the
1107 result to (void) does not stop the warnings; a more elaborate fudge is
1108 needed. I've used a macro to implement this.
1109
1110 7. Minor change to pcretest.c to avoid a compiler warning.
1111
1112 8. Added four artifical Unicode properties to help with an option to make
1113 \s etc use properties (see next item). The new properties are: Xan
1114 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1115
1116 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1117 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1118 this option. Modified pcretest to add /W to test this facility. Added
1119 REG_UCP to make it available via the POSIX interface.
1120
1121 10. Added --line-buffered to pcregrep.
1122
1123 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1124 studied, and the match started with a letter with a code point greater than
1125 127 whose first byte was different to the first byte of the other case of
1126 the letter, the other case of this starting letter was not recognized
1127 (#976).
1128
1129 12. If a pattern that was studied started with a repeated Unicode property
1130 test, for example, \p{Nd}+, there was the theoretical possibility of
1131 setting up an incorrect bitmap of starting bytes, but fortunately it could
1132 not have actually happened in practice until change 8 above was made (it
1133 added property types that matched character-matching opcodes).
1134
1135 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1136 possible starting bytes for non-anchored patterns.
1137
1138 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1139 \R, and also a number of cases that involve Unicode properties, both
1140 explicit and implicit when PCRE_UCP is set.
1141
1142 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1143 input, it could crash or give wrong results if characters with values
1144 greater than 0xc0 were present in the subject string. (Detail: it assumed
1145 UTF-8 input when processing these items.)
1146
1147 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1148 size_t is 64-bit (#991).
1149
1150 17. Added a check for running out of memory when PCRE is compiled with
1151 --disable-stack-for-recursion (#990).
1152
1153 18. If the last data line in a file for pcretest does not have a newline on
1154 the end, a newline was missing in the output.
1155
1156 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1157 less than 128) in its various bitmaps. However, there is a facility for
1158 generating tables according to the current locale when PCRE is compiled. It
1159 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1160 characters, are recognized by isspace() and therefore were getting set in
1161 these tables, and indeed these tables seem to approximate to ISO 8859. This
1162 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1163 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1164 which of course cannot start UTF-8 characters. I have changed the code so
1165 that only real ASCII characters (less than 128) and the correct starting
1166 bytes for UTF-8 encodings are set for characters greater than 127 when in
1167 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1168 altogether.)
1169
1170 20. Added the /T option to pcretest so as to be able to run tests with non-
1171 standard character tables, thus making it possible to include the tests
1172 used for 19 above in the standard set of tests.
1173
1174 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1175 reference to a subpattern the other side of a comment that contains an
1176 opening parenthesis caused either an internal compiling error, or a
1177 reference to the wrong subpattern.
1178
1179
1180 Version 8.02 19-Mar-2010
1181 ------------------------
1182
1183 1. The Unicode data tables have been updated to Unicode 5.2.0.
1184
1185 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1186 configured.
1187
1188 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1189 original author of that file, following a query about its status.
1190
1191 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1192 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1193
1194 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1195 quantifier applied to a forward-referencing subroutine call, could compile
1196 incorrect code or give the error "internal error: previously-checked
1197 referenced subpattern not found".
1198
1199 6. Both MS Visual Studio and Symbian OS have problems with initializing
1200 variables to point to external functions. For these systems, therefore,
1201 pcre_malloc etc. are now initialized to local functions that call the
1202 relevant global functions.
1203
1204 7. There were two entries missing in the vectors called coptable and poptable
1205 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1206 I've fixed the data, and added a kludgy way of testing at compile time that
1207 the lengths are correct (equal to the number of opcodes).
1208
1209 8. Following on from 7, I added a similar kludge to check the length of the
1210 eint vector in pcreposix.c.
1211
1212 9. Error texts for pcre_compile() are held as one long string to avoid too
1213 much relocation at load time. To find a text, the string is searched,
1214 counting zeros. There was no check for running off the end of the string,
1215 which could happen if a new error number was added without updating the
1216 string.
1217
1218 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1219
1220 11. \K was not working if it appeared in an atomic group or in a group that
1221 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1222 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1223 the assertion is positive, but not if it is negative.
1224
1225 12. Change 11 fortuitously reduced the size of the stack frame used in the
1226 "match()" function of pcre_exec.c by one pointer. Forthcoming
1227 implementation of support for (*MARK) will need an extra pointer on the
1228 stack; I have reserved it now, so that the stack frame size does not
1229 decrease.
1230
1231 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1232 item in branch that calls a recursion is a subroutine call - as in the
1233 second branch in the above example - was incorrectly given the compile-
1234 time error "recursive call could loop indefinitely" because pcre_compile()
1235 was not correctly checking the subroutine for matching a non-empty string.
1236
1237 14. The checks for overrunning compiling workspace could trigger after an
1238 overrun had occurred. This is a "should never occur" error, but it can be
1239 triggered by pathological patterns such as hundreds of nested parentheses.
1240 The checks now trigger 100 bytes before the end of the workspace.
1241
1242 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1243
1244
1245 Version 8.01 19-Jan-2010
1246 ------------------------
1247
1248 1. If a pattern contained a conditional subpattern with only one branch (in
1249 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1250 computed the wrong minimum data length (which is of course zero for such
1251 subpatterns). This could cause incorrect "no match" results.
1252
1253 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1254 the pattern is reset in the first branch, pcre_compile() failed with
1255 "internal error: code overflow at offset...". This happened only when
1256 the reset was to the original external option setting. (An optimization
1257 abstracts leading options settings into an external setting, which was the
1258 cause of this.)
1259
1260 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1261 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1262 assertion pattern did not match (meaning that the assertion was true), it
1263 was incorrectly treated as false if the SKIP had been reached during the
1264 matching. This also applied to assertions used as conditions.
1265
1266 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1267 assertion subpattern, including such a pattern used as a condition,
1268 unpredictable results occurred, instead of the error return
1269 PCRE_ERROR_DFA_UITEM.
1270
1271 5. The C++ GlobalReplace function was not working like Perl for the special
1272 situation when an empty string is matched. It now does the fancy magic
1273 stuff that is necessary.
1274
1275 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1276 removed. (These were left over from very, very early versions of PCRE.)
1277
1278 7. Some cosmetic changes to the code to make life easier when compiling it
1279 as part of something else:
1280
1281 (a) Change DEBUG to PCRE_DEBUG.
1282
1283 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1284 called "current" as "current_branch", to prevent a collision with the
1285 Linux macro when compiled as a kernel module.
1286
1287 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1288 prevent a collision with the Linux macro when compiled as a kernel
1289 module.
1290
1291 8. In pcre_compile() there are some checks for integer overflows that used to
1292 cast potentially large values to (double). This has been changed to that
1293 when building, a check for int64_t is made, and if it is found, it is used
1294 instead, thus avoiding the use of floating point arithmetic. (There is no
1295 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1296 double.
1297
1298 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1299 2005 (difference between two addresses compared to an unsigned value).
1300
1301 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1302 custom one, because of the following reported problem in Windows:
1303
1304 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1305 under Win32.
1306 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1307 therefore missing the function definition.
1308 - The compiler thus generates a "C" signature for the test function.
1309 - The linker fails to find the "C" function.
1310 - PCRE fails to configure if asked to do so against libbz2.
1311
1312 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1313 messages were output:
1314
1315 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1316 rerunning libtoolize, to keep the correct libtool macros in-tree.
1317 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1318
1319 I have done both of these things.
1320
1321 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1322 most of the time, it *can* run out if it is given a pattern that contains a
1323 runaway infinite recursion. I updated the discussion in the pcrestack man
1324 page.
1325
1326 13. Now that we have gone to the x.xx style of version numbers, the minor
1327 version may start with zero. Using 08 or 09 is a bad idea because users
1328 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1329 interpreted as invalid octal numbers. I've updated the previous comment in
1330 configure.ac, and also added a check that gives an error if 08 or 09 are
1331 used.
1332
1333 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1334 causing partial matching to fail when the end of the subject matched \W
1335 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1336
1337 15. There were some discrepancies between the declarations in pcre_internal.h
1338 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1339 their definitions. The declarations used "const uschar *" and the
1340 definitions used USPTR. Even though USPTR is normally defined as "const
1341 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1342 reported that: "This difference in casting confuses some C++ compilers, for
1343 example, SunCC recognizes above declarations as different functions and
1344 generates broken code for hbpcre." I have changed the declarations to use
1345 USPTR.
1346
1347 16. GNU libtool is named differently on some systems. The autogen.sh script now
1348 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1349 (FreeBSD).
1350
1351 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1352 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1353 comment: "Figure out how to create a longlong from a string: strtoll and
1354 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1355 instance, but it only takes 2 args instead of 3!"
1356
1357 18. A subtle bug concerned with back references has been fixed by a change of
1358 specification, with a corresponding code fix. A pattern such as
1359 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1360 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1361 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1362 same bug. Such groups have to be quantified to be useful, or contained
1363 inside another quantified group. (If there's no repetition, the reference
1364 can never match.) The problem arises because, having left the group and
1365 moved on to the rest of the pattern, a later failure that backtracks into
1366 the group uses the captured value from the final iteration of the group
1367 rather than the correct earlier one. I have fixed this in PCRE by forcing
1368 any group that contains a reference to itself to be an atomic group; that
1369 is, there cannot be any backtracking into it once it has completed. This is
1370 similar to recursive and subroutine calls.
1371
1372
1373 Version 8.00 19-Oct-09
1374 ----------------------
1375
1376 1. The table for translating pcre_compile() error codes into POSIX error codes
1377 was out-of-date, and there was no check on the pcre_compile() error code
1378 being within the table. This could lead to an OK return being given in
1379 error.
1380
1381 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1382 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1383 in a Windows environment.
1384
1385 3. The pcregrep --count option prints the count for each file even when it is
1386 zero, as does GNU grep. However, pcregrep was also printing all files when
1387 --files-with-matches was added. Now, when both options are given, it prints
1388 counts only for those files that have at least one match. (GNU grep just
1389 prints the file name in this circumstance, but including the count seems
1390 more useful - otherwise, why use --count?) Also ensured that the
1391 combination -clh just lists non-zero counts, with no names.
1392
1393 4. The long form of the pcregrep -F option was incorrectly implemented as
1394 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1395 but it seems right to fix it, and I didn't think it was worth preserving
1396 the old behaviour.
1397
1398 5. The command line items --regex=pattern and --regexp=pattern were not
1399 recognized by pcregrep, which required --regex pattern or --regexp pattern
1400 (with a space rather than an '='). The man page documented the '=' forms,
1401 which are compatible with GNU grep; these now work.
1402
1403 6. No libpcreposix.pc file was created for pkg-config; there was just
1404 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1405
1406 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1407 when UCP support is not needed, by modifying the Python script that
1408 generates it from Unicode data files. This should not matter if the module
1409 is correctly used as a library, but I received one complaint about 50K of
1410 unwanted data. My guess is that the person linked everything into his
1411 program rather than using a library. Anyway, it does no harm.
1412
1413 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1414 was a minimum greater than 1 for a wide character in a possessive
1415 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1416 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1417 character. Chaos in the form of incorrect output or a compiling loop could
1418 result.
1419
1420 9. The restrictions on what a pattern can contain when partial matching is
1421 requested for pcre_exec() have been removed. All patterns can now be
1422 partially matched by this function. In addition, if there are at least two
1423 slots in the offset vector, the offset of the earliest inspected character
1424 for the match and the offset of the end of the subject are set in them when
1425 PCRE_ERROR_PARTIAL is returned.
1426
1427 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1428 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1429 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1430 and may be more useful for multi-segment matching.
1431
1432 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1433 used to be given if ever the end of the subject was reached; now it is
1434 given only if matching could not proceed because another character was
1435 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1436 string "Z", which now yields "no match" instead of "partial match". In the
1437 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1438 final character ended with (*FAIL).
1439
1440 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1441 if the pattern had a "must contain" character that was already found in the
1442 earlier partial match, unless partial matching was again requested. For
1443 example, with the pattern /dog.(body)?/, the "must contain" character is
1444 "g". If the first part-match was for the string "dog", restarting with
1445 "sbody" failed. This bug has been fixed.
1446
1447 13. The string returned by pcre_dfa_exec() after a partial match has been
1448 changed so that it starts at the first inspected character rather than the
1449 first character of the match. This makes a difference only if the pattern
1450 starts with a lookbehind assertion or \b or \B (\K is not supported by
1451 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1452 matching functions compatible, and I think it's the right thing to do.
1453
1454 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1455 so that the demonstration program is easily available in environments where
1456 PCRE has not been installed from source.
1457
1458 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1459 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1460 library.
1461
1462 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1463 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1464 is not the first non-POSIX option to be added. Clearly some people find
1465 these options useful.
1466
1467 17. If a caller to the POSIX matching function regexec() passes a non-zero
1468 value for nmatch with a NULL value for pmatch, the value of
1469 nmatch is forced to zero.
1470
1471 18. RunGrepTest did not have a test for the availability of the -u option of
1472 the diff command, as RunTest does. It now checks in the same way as
1473 RunTest, and also checks for the -b option.
1474
1475 19. If an odd number of negated classes containing just a single character
1476 interposed, within parentheses, between a forward reference to a named
1477 subpattern and the definition of the subpattern, compilation crashed with
1478 an internal error, complaining that it could not find the referenced
1479 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1480 [The bug was that it was starting one character too far in when skipping
1481 over the character class, thus treating the ] as data rather than
1482 terminating the class. This meant it could skip too much.]
1483
1484 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1485 /g option in pcretest when the pattern contains \K, which makes it possible
1486 to have an empty string match not at the start, even when the pattern is
1487 anchored. Updated pcretest and pcredemo to use this option.
1488
1489 21. If the maximum number of capturing subpatterns in a recursion was greater
1490 than the maximum at the outer level, the higher number was returned, but
1491 with unset values at the outer level. The correct (outer level) value is
1492 now given.
1493
1494 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1495 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1496 make it do so. The string so far is captured, making this feature
1497 compatible with Perl.
1498
1499 23. The tests have been re-organized, adding tests 11 and 12, to make it
1500 possible to check the Perl 5.10 features against Perl 5.10.
1501
1502 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1503 pattern matches a fixed length string. PCRE did not allow this; now it
1504 does. Neither allows recursion.
1505
1506 25. I finally figured out how to implement a request to provide the minimum
1507 length of subject string that was needed in order to match a given pattern.
1508 (It was back references and recursion that I had previously got hung up
1509 on.) This code has now been added to pcre_study(); it finds a lower bound
1510 to the length of subject needed. It is not necessarily the greatest lower
1511 bound, but using it to avoid searching strings that are too short does give
1512 some useful speed-ups. The value is available to calling programs via
1513 pcre_fullinfo().
1514
1515 26. While implementing 25, I discovered to my embarrassment that pcretest had
1516 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1517 study optimizations had never been tested with that matching function.
1518 Oops. What is worse, even when it was passed study data, there was a bug in
1519 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1520 were also very few tests of studied patterns with pcre_dfa_exec().
1521
1522 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1523 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1524 on the other side of the coin, they are no longer allowed to have different
1525 names, because these cannot be distinguished in PCRE, and this has caused
1526 confusion. (This is a difference from Perl.)
1527
1528 28. When duplicate subpattern names are present (necessarily with different
1529 numbers, as required by 27 above), and a test is made by name in a
1530 conditional pattern, either for a subpattern having been matched, or for
1531 recursion in such a pattern, all the associated numbered subpatterns are
1532 tested, and the overall condition is true if the condition is true for any
1533 one of them. This is the way Perl works, and is also more like the way
1534 testing by number works.
1535
1536
1537 Version 7.9 11-Apr-09
1538 ---------------------
1539
1540 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1541 (pcretest), all targets were linked against these libraries. This included
1542 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1543 libraries. This caused unwanted dependencies to be created. This problem
1544 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1545 pcretest is linked with readline.
1546
1547 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1548 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1549 moved outside it again, because FALSE and TRUE are already defined in AIX,
1550 but BOOL is not.
1551
1552 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1553 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1554
1555 4. The pcregrep documentation said spaces were inserted as well as colons (or
1556 hyphens) following file names and line numbers when outputting matching
1557 lines. This is not true; no spaces are inserted. I have also clarified the
1558 wording for the --colour (or --color) option.
1559
1560 5. In pcregrep, when --colour was used with -o, the list of matching strings
1561 was not coloured; this is different to GNU grep, so I have changed it to be
1562 the same.
1563
1564 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1565 each matching line was coloured. Now it goes on to look for further matches
1566 of any of the test patterns, which is the same behaviour as GNU grep.
1567
1568 7. A pattern that could match an empty string could cause pcregrep to loop; it
1569 doesn't make sense to accept an empty string match in pcregrep, so I have
1570 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1571 seems to be how GNU grep behaves.
1572
1573 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1574 start or after a newline", because the conditional assertion was not being
1575 correctly handled. The rule now is that both the assertion and what follows
1576 in the first alternative must satisfy the test.
1577
1578 9. If auto-callout was enabled in a pattern with a conditional group whose
1579 condition was an assertion, PCRE could crash during matching, both with
1580 pcre_exec() and pcre_dfa_exec().
1581
1582 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1583 used for matching.
1584
1585 11. Unicode property support in character classes was not working for
1586 characters (bytes) greater than 127 when not in UTF-8 mode.
1587
1588 12. Added the -M command line option to pcretest.
1589
1590 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1591
1592 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1593
1594 16. Added comments and documentation about mis-use of no_arg in the C++
1595 wrapper.
1596
1597 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1598 from Martin Jerabek that uses macro names for all relevant character and
1599 string constants.
1600
1601 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1602 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1603 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1604 these, but not everybody uses configure.
1605
1606 19. A conditional group that had only one branch was not being correctly
1607 recognized as an item that could match an empty string. This meant that an
1608 enclosing group might also not be so recognized, causing infinite looping
1609 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1610 with the subject "ab", where knowledge that the repeated group can match
1611 nothing is needed in order to break the loop.
1612
1613 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1614 exec(), but without supplying a callout function, matching went wrong.
1615
1616 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1617 leak if the size of the offset vector was greater than 30. When the vector
1618 is smaller, the saved offsets during recursion go onto a local stack
1619 vector, but for larger vectors malloc() is used. It was failing to free
1620 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1621 error, in fact).
1622
1623 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1624 heapframe that is used only when UTF-8 support is enabled. This caused no
1625 problem, but was untidy.
1626
1627 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1628 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1629 included within another project.
1630
1631 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1632 slightly modified by me:
1633
1634 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1635 not building pcregrep.
1636
1637 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1638 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1639
1640 25. Forward references, both numeric and by name, in patterns that made use of
1641 duplicate group numbers, could behave incorrectly or give incorrect errors,
1642 because when scanning forward to find the reference group, PCRE was not
1643 taking into account the duplicate group numbers. A pattern such as
1644 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1645
1646 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1647 the feature of a custom pointer more persuasive (as requested by a user).
1648
1649 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1650 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1651 pre-defined.
1652
1653 28. Added support for (*UTF8) at the start of a pattern.
1654
1655 29. Arrange for flags added by the "release type" setting in CMake to be shown
1656 in the configuration summary.
1657
1658
1659 Version 7.8 05-Sep-08
1660 ---------------------
1661
1662 1. Replaced UCP searching code with optimized version as implemented for Ad
1663 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1664 stage table and inline lookup instead of a function, giving speed ups of 2
1665 to 5 times on some simple patterns that I tested. Permission was given to
1666 distribute the MultiStage2.py script that generates the tables (it's not in
1667 the tarball, but is in the Subversion repository).
1668
1669 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1670 scripts.
1671
1672 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1673 a group with a zero qualifier. The result of the study could be incorrect,
1674 or the function might crash, depending on the pattern.
1675
1676 4. Caseless matching was not working for non-ASCII characters in back
1677 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1678 It now works when Unicode Property Support is available.
1679
1680 5. In pcretest, an escape such as \x{de} in the data was always generating
1681 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1682 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1683 truncation.
1684
1685 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1686
1687 7. Added two (int) casts to pcregrep when printing the difference of two
1688 pointers, in case they are 64-bit values.
1689
1690 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1691 test 2 if it fails.
1692
1693 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1694 and a #define of that name to empty if it is not externally set. This is to
1695 allow users of MSVC to set it if necessary.
1696
1697 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1698 the convenience functions in the pcre_get.c source file.
1699
1700 11. An option change at the start of a pattern that had top-level alternatives
1701 could cause overwriting and/or a crash. This command provoked a crash in
1702 some environments:
1703
1704 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1705
1706 This potential security problem was recorded as CVE-2008-2371.
1707
1708 12. For a pattern where the match had to start at the beginning or immediately
1709 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1710 pcre_dfa_exec() could read past the end of the passed subject if there was
1711 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1712 pcretest so that it places the subject at the end of its malloc-ed buffer.
1713
1714 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1715 exec() might read past the end of the data buffer in UTF-8 mode.
1716
1717 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1718 the data contained the byte 0x85 as part of a UTF-8 character within its
1719 first line. This applied both to normal and DFA matching.
1720
1721 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1722 /^[^d]*?$/8 failed to match "abc".
1723
1724 16. Added a missing copyright notice to pcrecpp_internal.h.
1725
1726 17. Make it more clear in the documentation that values returned from
1727 pcre_exec() in ovector are byte offsets, not character counts.
1728
1729 18. Tidied a few places to stop certain compilers from issuing warnings.
1730
1731 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1732 supplied by Stefan Weber. I made a further small update for 7.8 because
1733 there is a change of source arrangements: the pcre_searchfuncs.c module is
1734 replaced by pcre_ucd.c.
1735
1736
1737 Version 7.7 07-May-08
1738 ---------------------
1739
1740 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1741 a string to a long long, pretend we don't even have a long long." This is
1742 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1743
1744 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1745 pre-7.6 versions, which defined a global no_arg variable instead of putting
1746 it in the RE class. (See also #8 below.)
1747
1748 3. Remove a line of dead code, identified by coverity and reported by Nuno
1749 Lopes.
1750
1751 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1752
1753 (1) The include/exclude patterns were being applied to the whole pathnames
1754 of files, instead of just to the final components.
1755
1756 (2) If there was more than one level of directory, the subdirectories were
1757 skipped unless they satisfied the include/exclude conditions. This is
1758 inconsistent with GNU grep (and could even be seen as contrary to the
1759 pcregrep specification - which I improved to make it absolutely clear).
1760 The action now is always to scan all levels of directory, and just
1761 apply the include/exclude patterns to regular files.
1762
1763 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1764 --exclude_dir in the tests to avoid scanning .svn directories.
1765
1766 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1767 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1768 doesn't support NULs in patterns.
1769
1770 7. Added some missing "const"s to declarations of static tables in
1771 pcre_compile.c and pcre_dfa_exec.c.
1772
1773 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1774 caused by fix #2 above. (Subsequently also a second patch to fix the
1775 first patch. And a third patch - this was a messy problem.)
1776
1777 9. Applied Craig's patch to remove the use of push_back().
1778
1779 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1780 matching function regexec().
1781
1782 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1783 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1784 references. PCRE supports relative numbers with this syntax (I don't think
1785 Oniguruma does).
1786
1787 12. Previously, a group with a zero repeat such as (...){0} was completely
1788 omitted from the compiled regex. However, this means that if the group
1789 was called as a subroutine from elsewhere in the pattern, things went wrong
1790 (an internal error was given). Such groups are now left in the compiled
1791 pattern, with a new opcode that causes them to be skipped at execution
1792 time.
1793
1794 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1795 to the way PCRE behaves:
1796
1797 (a) A lone ] character is dis-allowed (Perl treats it as data).
1798
1799 (b) A back reference to an unmatched subpattern matches an empty string
1800 (Perl fails the current match path).
1801
1802 (c) A data ] in a character class must be notated as \] because if the
1803 first data character in a class is ], it defines an empty class. (In
1804 Perl it is not possible to have an empty class.) The empty class []
1805 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1806 The negative empty class [^] matches any one character, independently
1807 of the DOTALL setting.
1808
1809 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1810 non-existent subpattern following a character class starting with ']' and
1811 containing () gave an internal compiling error instead of "reference to
1812 non-existent subpattern". Fortunately, when the pattern did exist, the
1813 compiled code was correct. (When scanning forwards to check for the
1814 existencd of the subpattern, it was treating the data ']' as terminating
1815 the class, so got the count wrong. When actually compiling, the reference
1816 was subsequently set up correctly.)
1817
1818 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1819 it was being rejected as not supported by pcre_dfa_exec(), even though
1820 other assertions are supported. I have made pcre_dfa_exec() support
1821 (*FAIL).
1822
1823 16. The implementation of 13c above involved the invention of a new opcode,
1824 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1825 cannot be changed at match time, I realized I could make a small
1826 improvement to matching performance by compiling OP_ALLANY instead of
1827 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1828 on the OP_ANY path.
1829
1830 17. Compiling pcretest on Windows with readline support failed without the
1831 following two fixes: (1) Make the unistd.h include conditional on
1832 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1833
1834 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1835 ncurses library to be included for pcretest when ReadLine support is
1836 requested, but also to allow for it to be overridden. This patch came from
1837 Daniel Bergström.
1838
1839 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1840 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1841 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1842 spotting this.
1843
1844
1845 Version 7.6 28-Jan-08
1846 ---------------------
1847
1848 1. A character class containing a very large number of characters with
1849 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1850 overflow.
1851
1852 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1853 HAVE_LONG_LONG is not defined.
1854
1855 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1856 bring them up to date and include new features. This patch includes:
1857
1858 - Fixed PH's badly added libz and libbz2 support.
1859 - Fixed a problem with static linking.
1860 - Added pcredemo. [But later removed - see 7 below.]
1861 - Fixed dftables problem and added an option.
1862 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1863 HAVE_LONG_LONG.
1864 - Added readline support for pcretest.
1865 - Added an listing of the option settings after cmake has run.
1866
1867 4. A user submitted a patch to Makefile that makes it easy to create
1868 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1869 Makefile.am that cause it to include this special target, without
1870 affecting anything else. Note that the same mingw target plus all
1871 the other distribution libraries and programs are now supported
1872 when configuring with CMake (see 6 below) instead of with
1873 Configure/Make.
1874
1875 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1876 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1877 exported in the Windows port". It has not yet been confirmed that the patch
1878 solves the problem, but it does no harm.
1879
1880 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1881 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1882 with CMake, and also correct the comment about stack recursion.
1883
1884 7. Remove the automatic building of pcredemo from the ./configure system and
1885 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1886 of a program that users should build themselves after PCRE is installed, so
1887 building it automatically is not really right. What is more, it gave
1888 trouble in some build environments.
1889
1890 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1891
1892
1893 Version 7.5 10-Jan-08
1894 ---------------------
1895
1896 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1897 values in parens when parsing an RE using the C++ wrapper."
1898
1899 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1900 Characters greater than 255 were excluded from the class instead of being
1901 included.
1902
1903 3. The same bug as (2) above applied to negated POSIX classes such as
1904 [:^space:].
1905
1906 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1907 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1908 I have changed it.
1909
1910 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1911 first named subpattern) and a construct such as (?&a) would reference the
1912 first named subpattern whose name started with "a" (in other words, the
1913 length check was missing). Both these problems are fixed. "Subpattern name
1914 expected" is now given for (?&) (a zero-length name), and this patch also
1915 makes it give the same error for \k'' (previously it complained that that
1916 was a reference to a non-existent subpattern).
1917
1918 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1919 this is right because (?- can be followed by option settings as well as by
1920 digits. I have, however, made the messages clearer.
1921
1922 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1923 than the number used in the conditional) now cause a compile-time error.
1924 This is actually not compatible with Perl, which accepts such patterns, but
1925 treats the conditional as always being FALSE (as PCRE used to), but it
1926 seems to me that giving a diagnostic is better.
1927
1928 8. Change "alphameric" to the more common word "alphanumeric" in comments
1929 and messages.
1930
1931 9. Fix two occurrences of "backslash" in comments that should have been
1932 "backspace".
1933
1934 10. Remove two redundant lines of code that can never be obeyed (their function
1935 was moved elsewhere).
1936
1937 11. The program that makes PCRE's Unicode character property table had a bug
1938 which caused it to generate incorrect table entries for sequences of
1939 characters that have the same character type, but are in different scripts.
1940 It amalgamated them into a single range, with the script of the first of
1941 them. In other words, some characters were in the wrong script. There were
1942 thirteen such cases, affecting characters in the following ranges:
1943
1944 U+002b0 - U+002c1
1945 U+0060c - U+0060d
1946 U+0061e - U+00612
1947 U+0064b - U+0065e
1948 U+0074d - U+0076d
1949 U+01800 - U+01805
1950 U+01d00 - U+01d77
1951 U+01d9b - U+01dbf
1952 U+0200b - U+0200f
1953 U+030fc - U+030fe
1954 U+03260 - U+0327f
1955 U+0fb46 - U+0fbb1
1956 U+10450 - U+1049d
1957
1958 12. The -o option (show only the matching part of a line) for pcregrep was not
1959 compatible with GNU grep in that, if there was more than one match in a
1960 line, it showed only the first of them. It now behaves in the same way as
1961 GNU grep.
1962
1963 13. If the -o and -v options were combined for pcregrep, it printed a blank
1964 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1965 does the same. The return code can be used to tell if there were any
1966 non-matching lines.
1967
1968 14. Added --file-offsets and --line-offsets to pcregrep.
1969
1970 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1971 infinitely looping recursion. The bug was that positive lookaheads were not
1972 being skipped when checking for a possible empty match (negative lookaheads
1973 and both kinds of lookbehind were skipped).
1974
1975 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1976 inclusion of <windows.h> to before rather than after the definition of
1977 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1978
1979 17. Specifying a possessive quantifier with a specific limit for a Unicode
1980 character property caused pcre_compile() to compile bad code, which led at
1981 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1982 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1983 caused the error; without that there was no problem.
1984
1985 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1986
1987 19. Added --enable-pcretest-libreadline.
1988
1989 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1990 RE::GlobalReplace(). As a result, the number of replacements returned was
1991 double what it should be. I removed one of the increments, but Craig sent a
1992 later patch that removed the other one (the right fix) and added unit tests
1993 that check the return values (which was not done before).
1994
1995 21. Several CMake things:
1996
1997 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1998 the names libpcre and libpcreposix, not just pcre and pcreposix.
1999
2000 (2) The above change means that pcretest and pcregrep are now correctly
2001 linked with the newly-built libraries, not previously installed ones.
2002
2003 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2004
2005 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2006 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2007 UTF-8 newline character). The key issue is that the pattern starts .*;
2008 this means that the match must be either at the beginning, or after a
2009 newline. The bug was in the code for advancing after a failed match and
2010 checking that the new position followed a newline. It was not taking
2011 account of UTF-8 characters correctly.
2012
2013 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2014 character classes. PCRE was not treating the sequence [:...:] as a
2015 character class unless the ... were all letters. Perl, however, seems to
2016 allow any characters between [: and :], though of course it rejects as
2017 unknown any "names" that contain non-letters, because all the known class
2018 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2019 for example, whereas PCRE did not - it did not recognize a POSIX character
2020 class. This seemed a bit dangerous, so the code has been changed to be
2021 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2022 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2023 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2024 Perl does, and where it didn't before.
2025
2026 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2027 Windows environments %n is disabled by default.
2028
2029
2030 Version 7.4 21-Sep-07
2031 ---------------------
2032
2033 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2034 means that a class such as [\s] counted as "explicit reference to CR or
2035 LF". That isn't really right - the whole point of the change was to try to
2036 help when there was an actual mention of one of the two characters. So now
2037 the change happens only if \r or \n (or a literal CR or LF) character is
2038 encountered.
2039
2040 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2041 of both had grown to the point where there were only 3 bits left.
2042 Fortunately, there was spare space in the data structure, and so I have
2043 moved the internal flags into a new 16-bit field to free up more option
2044 bits.
2045
2046 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2047 but did not set the internal JCHANGED flag - either of these is enough to
2048 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2049 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2050 start sets both bits.
2051
2052 4. Added options (at build time, compile time, exec time) to change \R from
2053 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2054
2055 5. doc/pcresyntax.html was missing from the distribution.
2056
2057 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2058 compatibility, even though it is no longer used.
2059
2060 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2061 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2062 windows.h file is present (where different names are used). [This was
2063 reversed later after testing - see 16 below.]
2064
2065 8. Changed all #include <config.h> to #include "config.h". There were also
2066 some further <pcre.h> cases that I changed to "pcre.h".
2067
2068 9. When pcregrep was used with the --colour option, it missed the line ending
2069 sequence off the lines that it output.
2070
2071 10. It was pointed out to me that arrays of string pointers cause lots of
2072 relocations when a shared library is dynamically loaded. A technique of
2073 using a single long string with a table of offsets can drastically reduce
2074 these. I have refactored PCRE in four places to do this. The result is
2075 dramatic:
2076
2077 Originally: 290
2078 After changing UCP table: 187
2079 After changing error message table: 43
2080 After changing table of "verbs" 36
2081 After changing table of Posix names 22
2082
2083 Thanks to the folks working on Gregex for glib for this insight.
2084
2085 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2086 unicode-properties was also set.
2087
2088 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2089
2090 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2091 checked only for CRLF.
2092
2093 14. Added casts to pcretest.c to avoid compiler warnings.
2094
2095 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2096
2097 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2098 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2099 entirely. This removes changes made in 7 above.
2100
2101 17. The CMake files have been updated, and there is now more information about
2102 building with CMake in the NON-UNIX-USE document.
2103
2104
2105 Version 7.3 28-Aug-07
2106 ---------------------
2107
2108 1. In the rejigging of the build system that eventually resulted in 7.1, the
2109 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2110 brackets there is not right, since it causes compilers to look for an
2111 installed pcre.h, not the version that is in the source that is being
2112 compiled (which of course may be different). I have changed it back to:
2113
2114 #include "pcre.h"
2115
2116 I have a vague recollection that the change was concerned with compiling in
2117 different directories, but in the new build system, that is taken care of
2118 by the VPATH setting the Makefile.
2119
2120 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2121 when the subject happened to end in the byte 0x85 (e.g. if the last
2122 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2123 characters but of course it shouldn't be taken as a newline when it is part
2124 of another character. The bug was that, for an unlimited repeat of . in
2125 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2126 characters when looking for a newline.
2127
2128 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2129
2130 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2131 in debug output.
2132
2133 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2134 long printing in the pcrecpp unittest when running under MinGW.
2135
2136 6. ESC_K was left out of the EBCDIC table.
2137
2138 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2139 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2140 limit also applies to "virtual nesting" when a pattern is recursive, and in
2141 this case 1000 isn't so big. I have been able to remove this limit at the
2142 expense of backing off one optimization in certain circumstances. Normally,
2143 when pcre_exec() would call its internal match() function recursively and
2144 immediately return the result unconditionally, it uses a "tail recursion"
2145 feature to save stack. However, when a subpattern that can match an empty
2146 string has an unlimited repetition quantifier, it no longer makes this
2147 optimization. That gives it a stack frame in which to save the data for
2148 checking that an empty string has been matched. Previously this was taken
2149 from the 1000-entry workspace that had been reserved. So now there is no
2150 explicit limit, but more stack is used.
2151
2152 8. Applied Daniel's patches to solve problems with the import/export magic
2153 syntax that is required for Windows, and which was going wrong for the
2154 pcreposix and pcrecpp parts of the library. These were overlooked when this
2155 problem was solved for the main library.
2156
2157 9. There were some crude static tests to avoid integer overflow when computing
2158 the size of patterns that contain repeated groups with explicit upper
2159 limits. As the maximum quantifier is 65535, the maximum group length was
2160 set at 30,000 so that the product of these two numbers did not overflow a
2161 32-bit integer. However, it turns out that people want to use groups that
2162 are longer than 30,000 bytes (though not repeat them that many times).
2163 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2164 made it possible to implement the integer overflow checks in a much more
2165 dynamic way, which I have now done. The artificial limitation on group
2166 length has been removed - we now have only the limit on the total length of
2167 the compiled pattern, which depends on the LINK_SIZE setting.
2168
2169 10. Fixed a bug in the documentation for get/copy named substring when
2170 duplicate names are permitted. If none of the named substrings are set, the
2171 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2172 empty string.
2173
2174 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2175 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2176 because the ] is interpreted as the first data character and the
2177 terminating ] is not found. PCRE has been made compatible with Perl in this
2178 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2179 cause memory overwriting.
2180
2181 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2182 string has been matched (to stop an infinite loop). It was not recognizing
2183 a conditional subpattern that could match an empty string if that
2184 subpattern was within another subpattern. For example, it looped when
2185 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2186 condition was not nested. This bug has been fixed.
2187
2188 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2189 past the start of the subject in the presence of bytes with the top bit
2190 set, for example "\x8aBCD".
2191
2192 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2193 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2194
2195 14. Optimized (?!) to (*FAIL).
2196
2197 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2198 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2199 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2200 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2201 does: it's just the validity check that is more restrictive.
2202
2203 16. Inserted checks for integer overflows during escape sequence (backslash)
2204 processing, and also fixed erroneous offset values for syntax errors during
2205 backslash processing.
2206
2207 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2208 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2209
2210 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2211 caused an overrun.
2212
2213 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2214 something other than just ASCII characters) inside a group that had an
2215 unlimited repeat caused a loop at compile time (while checking to see
2216 whether the group could match an empty string).
2217
2218 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2219 [\P{Any}] did so. (Error in the code for printing property names.)
2220
2221 21. An orphan \E inside a character class could cause a crash.
2222
2223 22. A repeated capturing bracket such as (A)? could cause a wild memory
2224 reference during compilation.
2225
2226 23. There are several functions in pcre_compile() that scan along a compiled
2227 expression for various reasons (e.g. to see if it's fixed length for look
2228 behind). There were bugs in these functions when a repeated \p or \P was
2229 present in the pattern. These operators have additional parameters compared
2230 with \d, etc, and these were not being taken into account when moving along
2231 the compiled data. Specifically:
2232
2233 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2234 length.
2235
2236 (b) An item such as \pL+ within a repeated group could cause crashes or
2237 loops.
2238
2239 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2240 "reference to non-existent subpattern" error.
2241
2242 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2243
2244 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2245 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2246
2247 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2248
2249 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2250 character were causing crashes (broken optimization).
2251
2252 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2253 \p or \P) caused a compile-time loop.
2254
2255 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2256 break. For example, the unstudied pattern [\r\n]A does not match the string
2257 "\r\nA" because change 7.0/46 below moves the current point on by two
2258 characters after failing to match at the start. However, the pattern \nA
2259 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2260 the same is true. There doesn't seem any very clean way out of this, but
2261 what I have chosen to do makes the common cases work: PCRE now takes note
2262 of whether there can be an explicit match for \r or \n anywhere in the
2263 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2264 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2265 pattern has explicit CR or LF references.
2266
2267 29. Added (*CR) etc for changing newline setting at start of pattern.
2268
2269
2270 Version 7.2 19-Jun-07
2271 ---------------------
2272
2273 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2274 which is apparently normally available under Windows.
2275
2276 2. Re-jig the pcregrep tests with different newline settings in an attempt
2277 to make them independent of the local environment's newline setting.
2278
2279 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2280
2281 4. Some of the "internals" tests were previously cut out when the link size
2282 was not 2, because the output contained actual offsets. The recent new
2283 "Z" feature of pcretest means that these can be cut out, making the tests
2284 usable with all link sizes.
2285
2286 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2287 stack recursion. This gives a massive performance boost under BSD, but just
2288 a small improvement under Linux. However, it saves one field in the frame
2289 in all cases.
2290
2291 6. Added more features from the forthcoming Perl 5.10:
2292
2293 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2294 recursion call. It refers to the nth most recently opened parentheses.
2295
2296 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2297 to be opened parentheses.
2298
2299 (c) Conditions that refer to capturing parentheses can be specified
2300 relatively, for example, (?(-2)... or (?(+3)...
2301
2302 (d) \K resets the start of the current match so that everything before
2303 is not part of it.
2304
2305 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2306
2307 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2308 reference syntax.
2309
2310 (g) (?| introduces a group in which the numbering of parentheses in each
2311 alternative starts with the same number.
2312
2313 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2314
2315 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2316 PCRE_INFO_JCHANGED.
2317
2318 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2319 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2320 for detecting groups that can match an empty string.
2321
2322 9. A pattern with a very large number of alternatives (more than several
2323 hundred) was running out of internal workspace during the pre-compile
2324 phase, where pcre_compile() figures out how much memory will be needed. A
2325 bit of new cunning has reduced the workspace needed for groups with
2326 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2327 workspace instead of running out of the 4096 that are available.
2328
2329 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2330
2331 11. Applied patch from Google to remove an optimization that didn't quite work.
2332 The report of the bug said:
2333
2334 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2335 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2336 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2337
2338 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2339 it matched the wrong number of bytes.
2340
2341
2342 Version 7.1 24-Apr-07
2343 ---------------------
2344
2345 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2346 that is more "standard", making use of automake and other Autotools. There
2347 is some re-arrangement of the files and adjustment of comments consequent
2348 on this.
2349
2350 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2351 for recursive directory scanning broke on some systems because the files
2352 are not scanned in any specific order and on different systems the order
2353 was different. A call to "sort" has been inserted into RunGrepTest for the
2354 approprate test as a short-term fix. In the longer term there may be an
2355 alternative.
2356
2357 3. I had an email from Eric Raymond about problems translating some of PCRE's
2358 man pages to HTML (despite the fact that I distribute HTML pages, some
2359 people do their own conversions for various reasons). The problems
2360 concerned the use of low-level troff macros .br and .in. I have therefore
2361 removed all such uses from the man pages (some were redundant, some could
2362 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2363 HTML has been updated to handle .nf/.fi and to complain if it encounters
2364 .br or .in.
2365
2366 4. Updated comments in configure.ac that get placed in config.h.in and also
2367 arranged for config.h to be included in the distribution, with the name
2368 config.h.generic, for the benefit of those who have to compile without
2369 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2370
2371 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2372 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2373 makevp.bat for the current PCRE, using the additional files
2374 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2375
2376 6. A Windows user reported a minor discrepancy with test 2, which turned out
2377 to be caused by a trailing space on an input line that had got lost in his
2378 copy. The trailing space was an accident, so I've just removed it.
2379
2380 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2381 that is needed.
2382
2383 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2384 as "const" (a) because they are and (b) because it helps the PHP
2385 maintainers who have recently made a script to detect big data structures
2386 in the php code that should be moved to the .rodata section. I remembered
2387 to update Builducptable as well, so it won't revert if ucptable.h is ever
2388 re-created.
2389
2390 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2391 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2392 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2393 support is not required. This saves 1.5-2K of code, which is important in
2394 some applications.
2395
2396 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2397 so as not to refer to the tables, even though these functions will never be
2398 called when UTF-8 support is disabled. Otherwise there are problems with a
2399 shared library.
2400
2401 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2402
2403 (a) It was defining its arguments as char * instead of void *.
2404
2405 (b) It was assuming that all moves were upwards in memory; this was true
2406 a long time ago when I wrote it, but is no longer the case.
2407
2408 The emulated memove() is provided for those environments that have neither
2409 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2410 is clearly not the case, as these two bugs were recently reported.
2411
2412 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2413 and Detrail to create the HTML documentation, the .txt form of the man
2414 pages, and it removes trailing spaces from listed files. It also creates
2415 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2416 case, it wraps all the #defines with #ifndefs. This script should be run
2417 before "make dist".
2418
2419 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2420 with Unicode property support.
2421
2422 (a) For a maximizing quantifier, if the two different cases of the
2423 character were of different lengths in their UTF-8 codings (there are
2424 some cases like this - I found 11), and the matching function had to
2425 back up over a mixture of the two cases, it incorrectly assumed they
2426 were both the same length.
2427
2428 (b) When PCRE was configured to use the heap rather than the stack for
2429 recursion during matching, it was not correctly preserving the data for
2430 the other case of a UTF-8 character when checking ahead for a match
2431 while processing a minimizing repeat. If the check also involved
2432 matching a wide character, but failed, corruption could cause an
2433 erroneous result when trying to check for a repeat of the original
2434 character.
2435
2436 13. Some tidying changes to the testing mechanism:
2437
2438 (a) The RunTest script now detects the internal link size and whether there
2439 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2440 values substituted by "configure". (The RunGrepTest script already did
2441 this for UTF-8.) The configure.ac script no longer substitutes the
2442 relevant variables.
2443
2444 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2445 with length and offset values. This means that the output is different
2446 for different internal link sizes. Test 2 is skipped for link sizes
2447 other than 2 because of this, bypassing the problem. Unfortunately,
2448 there was also a test in test 3 (the locale tests) that used /B and
2449 failed for link sizes other than 2. Rather than cut the whole test out,
2450 I have added a new /Z option to pcretest that replaces the length and
2451 offset values with spaces. This is now used to make test 3 independent
2452 of link size. (Test 2 will be tidied up later.)
2453
2454 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2455 segmentation fault instead of returning the appropriate error message.
2456
2457 15. In multiline mode when the newline sequence was set to "any", the pattern
2458 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2459 This doesn't seem right; it now treats the CRLF combination as the line
2460 ending, and so does not match in that case. It's only a pattern such as ^$
2461 that would hit this one: something like ^ABC$ would have failed after \r
2462 and then tried again after \r\n.
2463
2464 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2465 in an attempt to make files that differ only in their line terminators
2466 compare equal. This works on Linux.
2467
2468 17. Under certain error circumstances pcregrep might try to free random memory
2469 as it exited. This is now fixed, thanks to valgrind.
2470
2471 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2472 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2473 was because its rules for how to advance for /g after matching an empty
2474 string at the end of a line did not allow for this case. They now check for
2475 it specially.
2476
2477 20. pcretest is supposed to handle patterns and data of any length, by
2478 extending its buffers when necessary. It was getting this wrong when the
2479 buffer for a data line had to be extended.
2480
2481 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2482 CRLF as a newline sequence.
2483
2484 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2485 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2486 I have nevertheless tidied it up.
2487
2488 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2489
2490 24. Added a man page for pcre-config.
2491
2492
2493 Version 7.0 19-Dec-06
2494 ---------------------
2495
2496 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2497 moving to gcc 4.1.1.
2498
2499 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2500 sys/time.h, which is documented as needed for this function. It doesn't
2501 seem to matter on Linux, but it showed up on some releases of OS X.
2502
2503 3. It seems that there are systems where bytes whose values are greater than
2504 127 match isprint() in the "C" locale. The "C" locale should be the
2505 default when a C program starts up. In most systems, only ASCII printing
2506 characters match isprint(). This difference caused the output from pcretest
2507 to vary, making some of the tests fail. I have changed pcretest so that:
2508
2509 (a) When it is outputting text in the compiled version of a pattern, bytes
2510 other than 32-126 are always shown as hex escapes.
2511
2512 (b) When it is outputting text that is a matched part of a subject string,
2513 it does the same, unless a different locale has been set for the match
2514 (using the /L modifier). In this case, it uses isprint() to decide.
2515
2516 4. Fixed a major bug that caused incorrect computation of the amount of memory
2517 required for a compiled pattern when options that changed within the
2518 pattern affected the logic of the preliminary scan that determines the
2519 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2520 that the computed length was too small. The symptoms of this bug were
2521 either the PCRE error "internal error: code overflow" from pcre_compile(),
2522 or a glibc crash with a message such as "pcretest: free(): invalid next
2523 size (fast)". Examples of patterns that provoked this bug (shown in
2524 pcretest format) are:
2525
2526 /(?-x: )/x
2527 /(?x)(?-x: \s*#\s*)/
2528 /((?i)[\x{c0}])/8
2529 /(?i:[\x{c0}])/8
2530
2531 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2532 is now done differently.
2533
2534 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2535 wrapper classes; (b) implement a new function in the C++ scanner that is
2536 more efficient than the old way of doing things because it avoids levels of
2537 recursion in the regex matching; (c) add a paragraph to the documentation
2538 for the FullMatch() function.
2539
2540 6. The escape sequence \n was being treated as whatever was defined as
2541 "newline". Not only was this contrary to the documentation, which states
2542 that \n is character 10 (hex 0A), but it also went horribly wrong when
2543 "newline" was defined as CRLF. This has been fixed.
2544
2545 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2546 was being set to -1 for the "end of line" case (supposedly a value that no
2547 character can have). Though this value is never used (the check for end of
2548 line is "zero bytes in current character"), it caused compiler complaints.
2549 I've changed it to 0xffffffff.
2550
2551 8. In pcre_version.c, the version string was being built by a sequence of
2552 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2553 string (as it is for production releases) called a macro with an empty
2554 argument. The C standard says the result of this is undefined. The gcc
2555 compiler treats it as an empty string (which was what was wanted) but it is
2556 reported that Visual C gives an error. The source has been hacked around to
2557 avoid this problem.
2558
2559 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2560 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2561 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2562 of them did).
2563
2564 10. Originally, pcretest opened its input and output without "b"; then I was
2565 told that "b" was needed in some environments, so it was added for release
2566 5.0 to both the input and output. (It makes no difference on Unix-like
2567 systems.) Later I was told that it is wrong for the input on Windows. I've
2568 now abstracted the modes into two macros, to make it easier to fiddle with
2569 them, and removed "b" from the input mode under Windows.
2570
2571 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2572
2573 12. Added -help and --help to pcretest as an official way of being reminded
2574 of the options.
2575
2576 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2577 and pcrecpp.cc because they annoy compilers at high warning levels.
2578
2579 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2580
2581 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2582 scripts are not C programs :-) and which was not noticed because it works
2583 on Linux.
2584
2585 16. pcretest is supposed to handle any length of pattern and data line (as one
2586 line or as a continued sequence of lines) by extending its input buffer if
2587 necessary. This feature was broken for very long pattern lines, leading to
2588 a string of junk being passed to pcre_compile() if the pattern was longer
2589 than about 50K.
2590
2591 17. I have done a major re-factoring of the way pcre_compile() computes the
2592 amount of memory needed for a compiled pattern. Previously, there was code
2593 that made a preliminary scan of the pattern in order to do this. That was
2594 OK when PCRE was new, but as the facilities have expanded, it has become
2595 harder and harder to keep it in step with the real compile phase, and there
2596 have been a number of bugs (see for example, 4 above). I have now found a
2597 cunning way of running the real compile function in a "fake" mode that
2598 enables it to compute how much memory it would need, while actually only
2599 ever using a few hundred bytes of working memory and without too many
2600 tests of the mode. This should make future maintenance and development
2601 easier. A side effect of this work is that the limit of 200 on the nesting
2602 depth of parentheses has been removed (though this was never a serious
2603 limitation, I suspect). However, there is a downside: pcre_compile() now
2604 runs more slowly than before (30% or more, depending on the pattern). I
2605 hope this isn't a big issue. There is no effect on runtime performance.
2606
2607 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2608 newline (only possible for the last line of a file) and it was a
2609 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2610
2611 19. Added additional timing features to pcretest. (1) The -tm option now times
2612 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2613 separate command line item, by a number that specifies the number of
2614 repeats to use when timing. The default is 50000; this gives better
2615 precision, but takes uncomfortably long for very large patterns.
2616
2617 20. Extended pcre_study() to be more clever in cases where a branch of a
2618 subpattern has no definite first character. For example, (a*|b*)[cd] would
2619 previously give no result from pcre_study(). Now it recognizes that the
2620 first character must be a, b, c, or d.
2621
2622 21. There was an incorrect error "recursive call could loop indefinitely" if
2623 a subpattern (or the entire pattern) that was being tested for matching an
2624 empty string contained only one non-empty item after a nested subpattern.
2625 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2626 incorrectly, because the \d was being skipped in the check.
2627
2628 22. The pcretest program now has a new pattern option /B and a command line
2629 option -b, which is equivalent to adding /B to every pattern. This causes
2630 it to show the compiled bytecode, without the additional information that
2631 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2632 is the same as /B/I).
2633
2634 23. A new optimization is now able automatically to treat some sequences such
2635 as a*b as a*+b. More specifically, if something simple (such as a character
2636 or a simple class like \d) has an unlimited quantifier, and is followed by
2637 something that cannot possibly match the quantified thing, the quantifier
2638 is automatically "possessified".
2639
2640 24. A recursive reference to a subpattern whose number was greater than 39
2641 went wrong under certain circumstances in UTF-8 mode. This bug could also
2642 have affected the operation of pcre_study().
2643
2644 25. Realized that a little bit of performance could be had by replacing
2645 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2646
2647 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2648
2649 27. Possessive quantifiers such as a++ were previously implemented by turning
2650 them into atomic groups such as ($>a+). Now they have their own opcodes,
2651 which improves performance. This includes the automatically created ones
2652 from 23 above.
2653
2654 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2655 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2656 the first matched character to be a colon. This applied both to named and
2657 numbered groups.
2658
2659 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2660
2661 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2662 building PCRE on Windows easier, so I have included it in the distribution.
2663
2664 31. There is now a check in pcretest against a ridiculously large number being
2665 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2666 loop, the loop is abandoned.
2667
2668 32. Forward references to subpatterns in conditions such as (?(2)...) where
2669 subpattern 2 is defined later cause pcre_compile() to search forwards in
2670 the pattern for the relevant set of parentheses. This search went wrong
2671 when there were unescaped parentheses in a character class, parentheses
2672 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2673
2674 33. "Subroutine" calls and backreferences were previously restricted to
2675 referencing subpatterns earlier in the regex. This restriction has now
2676 been removed.
2677
2678 34. Added a number of extra features that are going to be in Perl 5.10. On the
2679 whole, these are just syntactic alternatives for features that PCRE had
2680 previously implemented using the Python syntax or my own invention. The
2681 other formats are all retained for compatibility.
2682
2683 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2684 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2685 also .NET compatible.
2686
2687 (b) A recursion or subroutine call to a named group can now be defined as
2688 (?&name) as well as (?P>name).
2689
2690 (c) A backreference to a named group can now be defined as \k<name> or
2691 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2692 5.10, are also .NET compatible.
2693
2694 (d) A conditional reference to a named group can now use the syntax
2695 (?(<name>) or (?('name') as well as (?(name).
2696
2697 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2698 groups (named and numbered) that are never evaluated inline, but can be
2699 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2700 is always false. There may be only one alternative in such a group.
2701
2702 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2703 as the simple (?(R). The condition is true only if the most recent
2704 recursion is that of the given number or name. It does not search out
2705 through the entire recursion stack.
2706
2707 (g) The escape \gN or \g{N} has been added, where N is a positive or
2708 negative number, specifying an absolute or relative reference.
2709
2710 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2711 some "unreachable code" warnings.
2712
2713 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2714 things, this adds five new scripts.
2715
2716 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2717 There were also incompatibilities regarding the handling of \Q..\E inside
2718 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2719 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2720
2721 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2722 matches an empty string, and forcibly breaks the loop. There were bugs in
2723 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2724 against aaaa the result was just "a" rather than "aaaa", for example. Two
2725 separate and independent bugs (that affected different cases) have been
2726 fixed.
2727
2728 39. Refactored the code to abolish the use of different opcodes for small
2729 capturing bracket numbers. This is a tidy that I avoided doing when I
2730 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2731 The new approach is not only tidier, it makes it possible to reduce the
2732 memory needed to fix the previous bug (38).
2733
2734 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2735 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2736 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2737 mode.
2738
2739 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2740 report.
2741
2742 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2743 copy construction and assignment in the C++ wrapper.
2744
2745 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2746 couple of bugs that could have given wrong results in the "--newline=crlf"
2747 case.
2748
2749 44. Added a number of casts and did some reorganization of signed/unsigned int
2750 variables following suggestions from Dair Grant. Also renamed the variable
2751 "this" as "item" because it is a C++ keyword.
2752
2753 45. Arranged for dftables to add
2754
2755 #include "pcre_internal.h"
2756
2757 to pcre_chartables.c because without it, gcc 4.x may remove the array
2758 definition from the final binary if PCRE is built into a static library and
2759 dead code stripping is activated.
2760
2761 46. For an unanchored pattern, if a match attempt fails at the start of a
2762 newline sequence, and the newline setting is CRLF or ANY, and the next two
2763 characters are CRLF, advance by two characters instead of one.
2764
2765
2766 Version 6.7 04-Jul-06
2767 ---------------------
2768
2769 1. In order to handle tests when input lines are enormously long, pcretest has
2770 been re-factored so that it automatically extends its buffers when
2771 necessary. The code is crude, but this _is_ just a test program. The
2772 default size has been increased from 32K to 50K.
2773
2774 2. The code in pcre_study() was using the value of the re argument before
2775 testing it for NULL. (Of course, in any sensible call of the function, it
2776 won't be NULL.)
2777
2778 3. The memmove() emulation function in pcre_internal.h, which is used on
2779 systems that lack both memmove() and bcopy() - that is, hardly ever -
2780 was missing a "static" storage class specifier.
2781
2782 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2783 containing an extended class (one that cannot be represented by a bitmap
2784 because it contains high-valued characters or Unicode property items, e.g.
2785 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2786 pattern, but PCRE should not loop if you do not (it no longer does).
2787 [Detail: two cases were found: (a) a repeated subpattern containing an
2788 extended class; (b) a recursive reference to a subpattern that followed a
2789 previous extended class. It wasn't skipping over the extended class
2790 correctly when UTF-8 mode was not set.]
2791
2792 5. A negated single-character class was not being recognized as fixed-length
2793 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2794 compile error "lookbehind assertion is not fixed length".
2795
2796 6. The RunPerlTest auxiliary script was showing an unexpected difference
2797 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2798 write a Perl script that can interpret lines of an input file either as
2799 byte characters or as UTF-8, which is what "perltest" was being required to
2800 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2801 can't do is switch easily at run time between having the "use utf8;" pragma
2802 or not. In the end, I fudged it by using the RunPerlTest script to insert
2803 "use utf8;" explicitly for the UTF-8 tests.
2804
2805 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2806 the end of the subject string, contrary to the documentation and to what
2807 Perl does. This was true of both matching functions. Now it matches only at
2808 the start of the subject and immediately after *internal* newlines.
2809
2810 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2811 a pointer to an int instead of a pointer to an unsigned long int. This
2812 caused problems on 64-bit systems.
2813
2814 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2815 instance of the 'standard' template library not being so standard".
2816
2817 10. There was no check on the number of named subpatterns nor the maximum
2818 length of a subpattern name. The product of these values is used to compute
2819 the size of the memory block for a compiled pattern. By supplying a very
2820 long subpattern name and a large number of named subpatterns, the size
2821 computation could be caused to overflow. This is now prevented by limiting
2822 the length of names to 32 characters, and the number of named subpatterns
2823 to 10,000.
2824
2825 11. Subpatterns that are repeated with specific counts have to be replicated in
2826 the compiled pattern. The size of memory for this was computed from the
2827 length of the subpattern and the repeat count. The latter is limited to
2828 65535, but there was no limit on the former, meaning that integer overflow
2829 could in principle occur. The compiled length of a repeated subpattern is
2830 now limited to 30,000 bytes in order to prevent this.
2831
2832 12. Added the optional facility to have named substrings with the same name.
2833
2834 13. Added the ability to use a named substring as a condition, using the
2835 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2836 are numbers (not recommended). Forward references are permitted.
2837
2838 14. Added forward references in named backreferences (if you see what I mean).
2839
2840 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2841 pattern could run off the end of the subject. For example, the pattern
2842 "(?s)(.{1,5})"8 did this with the subject "ab".
2843
2844 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2845 PCRE_CASELESS was set when matching characters that were quantified with ?
2846 or *.
2847
2848 17. A character class other than a single negated character that had a minimum
2849 but no maximum quantifier - for example [ab]{6,} - was not handled
2850 correctly by pce_dfa_exec(). It would match only one character.
2851
2852 18. A valid (though odd) pattern that looked like a POSIX character
2853 class but used an invalid character after [ (for example [[,abc,]]) caused
2854 pcre_compile() to give the error "Failed: internal error: code overflow" or
2855 in some cases to crash with a glibc free() error. This could even happen if
2856 the pattern terminated after [[ but there just happened to be a sequence of
2857 letters, a binary zero, and a closing ] in the memory that followed.
2858
2859 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2860 over the years. Originally (before any Unicode support), just the bottom 8
2861 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2862 output from "man perlunicode" includes this:
2863
2864 The regular expression compiler produces polymorphic opcodes. That
2865 is, the pattern adapts to the data and automatically switches to
2866 the Unicode character scheme when presented with Unicode data--or
2867 instead uses a traditional byte scheme when presented with byte
2868 data.
2869
2870 Sadly, a wide octal escape does not cause a switch, and in a string with
2871 no other multibyte characters, these octal escapes are treated as before.
2872 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2873 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2874 Unicode string.
2875
2876 I have not perpetrated such confusion in PCRE. Up till now, it took just
2877 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2878 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2879 translate to the appropriate multibyte character.
2880
2881 29. Applied some refactoring to reduce the number of warnings from Microsoft
2882 and Borland compilers. This has included removing the fudge introduced
2883 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2884 a warning about an unused variable.
2885
2886 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2887 characters since release 4.0, because Perl (from release 5.004) does not.
2888 [Or at least, is documented not to: some releases seem to be in conflict
2889 with the documentation.] However, when a pattern was studied with
2890 pcre_study() and all its branches started with \s, PCRE still included VT
2891 as a possible starting character. Of course, this did no harm; it just
2892 caused an unnecessary match attempt.
2893
2894 22. Removed a now-redundant internal flag bit that recorded the fact that case
2895 dependency changed within the pattern. This was once needed for "required
2896 byte" processing, but is no longer used. This recovers a now-scarce options
2897 bit. Also moved the least significant internal flag bit to the most-
2898 significant bit of the word, which was not previously used (hangover from
2899 the days when it was an int rather than a uint) to free up another bit for
2900 the future.
2901
2902 23. Added support for CRLF line endings as well as CR and LF. As well as the
2903 default being selectable at build time, it can now be changed at runtime
2904 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2905 specify that it is scanning data with non-default line endings.
2906
2907 24. Changed the definition of CXXLINK to make it agree with the definition of
2908 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2909
2910 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2911 recursions. This makes a big different to stack usage for some patterns.
2912
2913 26. If a subpattern containing a named recursion or subroutine reference such
2914 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2915 the space required for the compiled pattern went wrong and gave too small a
2916 value. Depending on the environment, this could lead to "Failed: internal
2917 error: code overflow at offset 49" or "glibc detected double free or
2918 corruption" errors.
2919
2920 27. Applied patches from Google (a) to support the new newline modes and (b) to
2921 advance over multibyte UTF-8 characters in GlobalReplace.
2922
2923 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2924 difference for some implementation of PCRE in some Windows version.
2925
2926 29. Added some extra testing facilities to pcretest:
2927
2928 \q<number> in a data line sets the "match limit" value
2929 \Q<number> in a data line sets the "match recursion limt" value
2930 -S <number> sets the stack size, where <number> is in megabytes
2931
2932 The -S option isn't available for Windows.
2933
2934
2935 Version 6.6 06-Feb-06
2936 ---------------------
2937
2938 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2939 in pcreposix.h. I have copied the definition from pcre.h.
2940
2941 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2942 because pcre.h is no longer a built file.
2943
2944 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2945 not normally included in the compiled code.
2946
2947
2948 Version 6.5 01-Feb-06
2949 ---------------------
2950
2951 1. When using the partial match feature with pcre_dfa_exec(), it was not
2952 anchoring the second and subsequent partial matches at the new starting
2953 point. This could lead to incorrect results. For example, with the pattern
2954 /1234/, partially matching against "123" and then "a4" gave a match.
2955
2956 2. Changes to pcregrep:
2957
2958 (a) All non-match returns from pcre_exec() were being treated as failures
2959 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2960 error message is output. Some extra information is given for the
2961 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2962 probably the only errors that are likely to be caused by users (by
2963 specifying a regex that has nested indefinite repeats, for instance).
2964 If there are more than 20 of these errors, pcregrep is abandoned.
2965
2966 (b) A binary zero was treated as data while matching, but terminated the
2967 output line if it was written out. This has been fixed: binary zeroes
2968 are now no different to any other data bytes.
2969
2970 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2971 used to set a locale for matching. The --locale=xxxx long option has
2972 been added (no short equivalent) to specify a locale explicitly on the
2973 pcregrep command, overriding the environment variables.
2974
2975 (d) When -B was used with -n, some line numbers in the output were one less
2976 than they should have been.
2977
2978 (e) Added the -o (--only-matching) option.
2979
2980 (f) If -A or -C was used with -c (count only), some lines of context were
2981 accidentally printed for the final match.
2982
2983 (g) Added the -H (--with-filename) option.
2984
2985 (h) The combination of options -rh failed to suppress file names for files
2986 that were found from directory arguments.
2987
2988 (i) Added the -D (--devices) and -d (--directories) options.
2989
2990 (j) Added the -F (--fixed-strings) option.
2991
2992 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2993
2994 (l) Added the --colo(u)r option.
2995
2996 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2997 is not present by default.
2998
2999 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3000 items such as (?R) or (?1), when the recursion could match a number of
3001 alternatives. If it matched one of the alternatives, but subsequently,
3002 outside the recursion, there was a failure, the code tried to back up into
3003 the recursion. However, because of the way PCRE is implemented, this is not
3004 possible, and the result was an incorrect result from the match.
3005
3006 In order to prevent this happening, the specification of recursion has
3007 been changed so that all such subpatterns are automatically treated as
3008 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3009
3010 4. I had overlooked the fact that, in some locales, there are characters for
3011 which isalpha() is true but neither isupper() nor islower() are true. In
3012 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3013 and ordfeminine) are like this. This affected the treatment of \w and \W
3014 when they appeared in character classes, but not when they appeared outside
3015 a character class. The bit map for "word" characters is now created
3016 separately from the results of isalnum() instead of just taking it from the
3017 upper, lower, and digit maps. (Plus the underscore character, of course.)
3018
3019 5. The above bug also affected the handling of POSIX character classes such as
3020 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3021 permanent tables. Instead, the bit maps for such a class were previously
3022 created as the appropriate unions of the upper, lower, and digit bitmaps.
3023 Now they are created by subtraction from the [[:word:]] class, which has
3024 its own bitmap.
3025
3026 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3027 It is created by subtracting the vertical space characters (\x09, \x0a,
3028 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3029 subtraction was done in the overall bitmap for a character class, meaning
3030 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3031 be recognized. This bug has been fixed.
3032
3033 7. Patches from the folks at Google:
3034
3035 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3036 real life, but is still worth protecting against".
3037
3038 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3039 regular expressions".
3040
3041 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3042 have it.
3043
3044 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3045 "configure" and the latter not, in order to fix a problem somebody had
3046 with compiling the Arg class on HP-UX.
3047
3048 (e) Improve the error-handling of the C++ wrapper a little bit.
3049
3050 (f) New tests for checking recursion limiting.
3051
3052 8. The pcre_memmove() function, which is used only if the environment does not
3053 have a standard memmove() function (and is therefore rarely compiled),
3054 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3055 returning a result (though PCRE never actually uses the result).
3056
3057 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3058 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3059 returned instead of calling malloc() with an overflowing number that would
3060 most likely cause subsequent chaos.
3061
3062 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3063
3064 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3065 with this option is matched, the nmatch and pmatch options of regexec() are
3066 ignored.
3067
3068 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3069 provided in case anyone wants to the the POSIX interface with UTF-8
3070 strings.
3071
3072 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3073 C++ linking (needed for some HP-UX environments).
3074
3075 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3076 (unused parameter) and in the pcre_printint() function (omitted "default"
3077 switch label when the default is to do nothing).
3078
3079 15. Added some code to make it possible, when PCRE is compiled as a C++
3080 library, to replace subject pointers for pcre_exec() with a smart pointer
3081 class, thus making it possible to process discontinuous strings.
3082
3083 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3084 much the same function. They were added by different people who were trying
3085 to make PCRE easy to compile on non-Unix systems. It has been suggested
3086 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3087 for compiling on Windows systems. I have therefore replaced it with
3088 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3089 defaults to "extern" for C or "extern C" for C++, which works fine on
3090 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3091 SCOPE with something explicit in config.h. In addition:
3092
3093 (a) pcreposix.h still had just "extern" instead of either of these macros;
3094 I have replaced it with PCRE_DATA_SCOPE.
3095
3096 (b) Functions such as _pcre_xclass(), which are internal to the library,
3097 but external in the C sense, all had PCRE_EXPORT in their definitions.
3098 This is apparently wrong for the Windows case, so I have removed it.
3099 (It makes no difference on Unix-like systems.)
3100
3101 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3102 of recursive calls to match(). This is different to MATCH_LIMIT because
3103 that limits the total number of calls to match(), not all of which increase
3104 the depth of recursion. Limiting the recursion depth limits the amount of
3105 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3106 when PCRE is compiled, and changed at run time. A patch from Google adds
3107 this functionality to the C++ interface.
3108
3109 18. Changes to the handling of Unicode character properties:
3110
3111 (a) Updated the table to Unicode 4.1.0.
3112
3113 (b) Recognize characters that are not in the table as "Cn" (undefined).
3114
3115 (c) I revised the way the table is implemented to a much improved format
3116 which includes recognition of ranges. It now supports the ranges that
3117 are defined in UnicodeData.txt, and it also amalgamates other
3118 characters into ranges. This has reduced the number of entries in the
3119 table from around 16,000 to around 3,000, thus reducing its size
3120 considerably. I realized I did not need to use a tree structure after
3121 all - a binary chop search is just as efficient. Having reduced the
3122 number of entries, I extended their size from 6 bytes to 8 bytes to
3123 allow for more data.
3124
3125 (d) Added support for Unicode script names via properties such as \p{Han}.
3126
3127 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3128 matching that character.
3129
3130 20. When matching a repeated Unicode property with a minimum greater than zero,
3131 (for example \pL{2,}), PCRE could look past the end of the subject if it
3132 reached it while seeking the minimum number of characters. This could
3133 happen only if some of the characters were more than one byte long, because
3134 there is a check for at least the minimum number of bytes.
3135
3136 21. Refactored the implementation of \p and \P so as to be more general, to
3137 allow for more different types of property in future. This has changed the
3138 compiled form incompatibly. Anybody with saved compiled patterns that use
3139 \p or \P will have to recompile them.
3140
3141 22. Added "Any" and "L&" to the supported property types.
3142
3143 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3144 but give a compile time error if the value is greater than 0xff.
3145
3146 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3147 accidentally not being installed or uninstalled.
3148
3149 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3150 made were to insert the current release number. This seemed silly, because
3151 it made things harder for people building PCRE on systems that don't run
3152 "configure". I have turned pcre.h into a distributed file, no longer built
3153 by "configure", with the version identification directly included. There is
3154 no longer a pcre.h.in file.
3155
3156 However, this change necessitated a change to the pcre-config script as
3157 well. It is built from pcre-config.in, and one of the substitutions was the
3158 release number. I have updated configure.ac so that ./configure now finds
3159 the release number by grepping pcre.h.
3160
3161 26. Added the ability to run the tests under valgrind.
3162
3163
3164 Version 6.4 05-Sep-05
3165 ---------------------
3166
3167 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3168 "--" to be printed when multiple files were scanned, even when none of the
3169 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3170 consider it to be a bug, and have restored the previous behaviour.
3171
3172 2. A couple of code tidies to get rid of compiler warnings.
3173
3174 3. The pcretest program used to cheat by referring to symbols in the library
3175 whose names begin with _pcre_. These are internal symbols that are not
3176 really supposed to be visible externally, and in some environments it is
3177 possible to suppress them. The cheating is now confined to including
3178 certain files from the library's source, which is a bit cleaner.
3179
3180 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3181 file's purpose clearer.
3182
3183 5. Reorganized pcre_ucp_findchar().
3184
3185
3186 Version 6.3 15-Aug-05
3187 ---------------------
3188
3189 1. The file libpcre.pc.in did not have general read permission in the tarball.
3190
3191 2. There were some problems when building without C++ support:
3192
3193 (a) If C++ support was not built, "make install" and "make test" still
3194 tried to test it.
3195
3196 (b) There were problems when the value of CXX was explicitly set. Some
3197 changes have been made to try to fix these, and ...
3198
3199 (c) --disable-cpp can now be used to explicitly disable C++ support.
3200
3201 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3202 backslash in a target when C++ was disabled. This confuses some
3203 versions of "make", apparently. Using an intermediate variable solves
3204 this. (Same for CPP_LOBJ.)
3205
3206 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3207 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3208 necessary on certain architectures.
3209
3210 4. Added a setting of -export-symbols-regex to the link command to remove
3211 those symbols that are exported in the C sense, but actually are local
3212 within the library, and not documented. Their names all begin with
3213 "_pcre_". This is not a perfect job, because (a) we have to except some
3214 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3215 available (and never for static libraries). I have made a note to try to
3216 find a way round (a) in the future.
3217
3218
3219 Version 6.2 01-Aug-05
3220 ---------------------
3221
3222 1. There was no test for integer overflow of quantifier values. A construction
3223 such as {1111111111111111} would give undefined results. What is worse, if
3224 a minimum quantifier for a parenthesized subpattern overflowed and became
3225 negative, the calculation of the memory size went wrong. This could have
3226 led to memory overwriting.
3227
3228 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3229
3230 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3231 operating environments where this matters.
3232
3233 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3234 PCRE options from within the C++ wrapper.
3235
3236 5. Named capturing subpatterns were not being correctly counted when a pattern
3237 was compiled. This caused two problems: (a) If there were more than 100
3238 such subpatterns, the calculation of the memory needed for the whole
3239 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3240 back references of the form \12, where the number was greater than 9, were
3241 not recognized as back references, even though there were sufficient
3242 previous subpatterns.
3243
3244 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3245 versions of gcc, e.g. 2.95.4.
3246
3247
3248 Version 6.1 21-Jun-05
3249 ---------------------
3250
3251 1. There was one reference to the variable "posix" in pcretest.c that was not
3252 surrounded by "#if !defined NOPOSIX".
3253
3254 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3255 the cross-check on the old pcre_info() function, for the benefit of the
3256 cut-down version of PCRE that is currently imported into Exim.
3257
3258 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3259 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3260 patterns that start (?i) but I don't think that matters. The use of (?i) is
3261 just an example; this all applies to the other options as well.
3262
3263 4. Since libtool seems to echo the compile commands it is issuing, the output
3264 from "make" can be reduced a bit by putting "@" in front of each libtool
3265 compile command.
3266
3267 5. Patch from the folks at Google for configure.in to be a bit more thorough
3268 in checking for a suitable C++ installation before trying to compile the
3269 C++ stuff. This should fix a reported problem when a compiler was present,
3270 but no suitable headers.
3271
3272 6. The man pages all had just "PCRE" as their title. I have changed them to
3273 be the relevant file name. I have also arranged that these names are
3274 retained in the file doc/pcre.txt, which is a concatenation in text format
3275 of all the man pages except the little individual ones for each function.
3276
3277 7. The NON-UNIX-USE file had not been updated for the different set of source
3278 files that come with release 6. I also added a few comments about the C++
3279 wrapper.
3280
3281
3282 Version 6.0 07-Jun-05
3283 ---------------------
3284
3285 1. Some minor internal re-organization to help with my DFA experiments.
3286
3287 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3288 didn't matter for the library itself when fully configured, but did matter
3289 when compiling without UCP support, or within Exim, where the ucp files are
3290 not imported.
3291
3292 3. Refactoring of the library code to split up the various functions into
3293 different source modules. The addition of the new DFA matching code (see
3294 below) to a single monolithic source would have made it really too
3295 unwieldy, quite apart from causing all the code to be include in a
3296 statically linked application, when only some functions are used. This is
3297 relevant even without the DFA addition now that patterns can be compiled in
3298 one application and matched in another.
3299
3300 The downside of splitting up is that there have to be some external
3301 functions and data tables that are used internally in different modules of
3302 the library but which are not part of the API. These have all had their
3303 names changed to start with "_pcre_" so that they are unlikely to clash
3304 with other external names.
3305
3306 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3307 a different (DFA) algorithm. Although it is slower than the original
3308 function, it does have some advantages for certain types of matching
3309 problem.
3310
3311 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3312 including restarting after a partial match.
3313
3314 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3315 defined when compiling for Windows was sent to me. I have put it into the
3316 code, though I have no means of testing or verifying it.
3317
3318 7. Added the pcre_refcount() auxiliary function.
3319
3320 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3321 match before or at the first newline in the subject string. In pcretest,
3322 the /f option on a pattern can be used to set this.
3323
3324 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3325 would behave wrongly. This has been present in PCRE since release 4.0.
3326
3327 10. A number of changes to the pcregrep command:
3328
3329 (a) Refactored how -x works; insert ^(...)$ instead of setting
3330 PCRE_ANCHORED and checking the length, in preparation for adding
3331 something similar for -w.
3332
3333 (b) Added the -w (match as a word) option.
3334
3335 (c) Refactored the way lines are read and buffered so as to have more
3336 than one at a time available.
3337
3338 (d) Implemented a pcregrep test script.
3339
3340 (e) Added the -M (multiline match) option. This allows patterns to match
3341 over several lines of the subject. The buffering ensures that at least
3342 8K, or the rest of the document (whichever is the shorter) is available
3343 for matching (and similarly the previous 8K for lookbehind assertions).
3344
3345 (f) Changed the --help output so that it now says
3346
3347 -w, --word-regex(p)
3348
3349 instead of two lines, one with "regex" and the other with "regexp"
3350 because that confused at least one person since the short forms are the
3351 same. (This required a bit of code, as the output is generated
3352 automatically from a table. It wasn't just a text change.)
3353
3354 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3355 option but starts with a hyphen. Could be a pattern or a path name
3356 starting with a hyphen, for instance.
3357
3358 (h) "-" can be given as a file name to represent stdin.
3359
3360 (i) When file names are being printed, "(standard input)" is used for
3361 the standard input, for compatibility with GNU grep. Previously
3362 "<stdin>" was used.
3363
3364 (j) The option --label=xxx can be used to supply a name to be used for
3365 stdin when file names are being printed. There is no short form.
3366
3367 (k) Re-factored the options decoding logic because we are going to add
3368 two more options that take data. Such options can now be given in four
3369 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3370
3371 (l) Added the -A, -B, and -C options for requesting that lines of context
3372 around matches be printed.
3373
3374 (m) Added the -L option to print the names of files that do not contain
3375 any matching lines, that is, the complement of -l.
3376
3377 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3378 continue to scan other files.
3379
3380 (o) The -s option was incorrectly implemented. For compatibility with other
3381 greps, it now suppresses the error message for a non-existent or non-
3382 accessible file (but not the return code). There is a new option called
3383 -q that suppresses the output of matching lines, which was what -s was
3384 previously doing.
3385
3386 (p) Added --include and --exclude options to specify files for inclusion
3387 and exclusion when recursing.
3388
3389 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3390 Hopefully, it now does.
3391
3392 12. Missing cast in pcre_study().
3393
3394 13. Added an "uninstall" target to the makefile.
3395
3396 14. Replaced "extern" in the function prototypes in Makefile.in with
3397 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3398 world, but is set differently for Windows.
3399
3400 15. Added a second compiling function called pcre_compile2(). The only
3401 difference is that it has an extra argument, which is a pointer to an
3402 integer error code. When there is a compile-time failure, this is set
3403 non-zero, in addition to the error test pointer being set to point to an
3404 error message. The new argument may be NULL if no error number is required
3405 (but then you may as well call pcre_compile(), which is now just a
3406 wrapper). This facility is provided because some applications need a
3407 numeric error indication, but it has also enabled me to tidy up the way
3408 compile-time errors are handled in the POSIX wrapper.
3409
3410 16. Added VPATH=.libs to the makefile; this should help when building with one
3411 prefix path and installing with another. (Or so I'm told by someone who
3412 knows more about this stuff than I do.)
3413
3414 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3415 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3416 match everything, including newlines. This is not POSIX-compatible, but
3417 somebody wanted the feature. From pcretest it can be activated by using
3418 both the P and the s flags.
3419
3420 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3421
3422 19. libpcre.pc was being incorrectly installed as executable.
3423
3424 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3425 it now also looks for '\r' so that it will work unmodified on Windows.
3426
3427 21. Added Google's contributed C++ wrapper to the distribution.
3428
3429 22. Added some untidy missing memory free() calls in pcretest, to keep
3430 Electric Fence happy when testing.
3431
3432
3433
3434 Version 5.0 13-Sep-04
3435 ---------------------
3436
3437 1. Internal change: literal characters are no longer packed up into items
3438 containing multiple characters in a single byte-string. Each character
3439 is now matched using a separate opcode. However, there may be more than one
3440 byte in the character in UTF-8 mode.
3441
3442 2. The pcre_callout_block structure has two new fields: pattern_position and
3443 next_item_length. These contain the offset in the pattern to the next match
3444 item, and its length, respectively.
3445
3446 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3447 insertion of callouts before each pattern item. Added the /C option to
3448 pcretest to make use of this.
3449
3450 4. On the advice of a Windows user, the lines
3451
3452 #if defined(_WIN32) || defined(WIN32)
3453 _setmode( _fileno( stdout ), 0x8000 );
3454 #endif /* defined(_WIN32) || defined(WIN32) */
3455
3456 have been added to the source of pcretest. This apparently does useful
3457 magic in relation to line terminators.
3458
3459 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3460 for the benefit of those environments where the "b" makes a difference.
3461
3462 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3463 to know about it. I have put a hack into configure.in that adds in code
3464 to set GCC=yes if CC=icc. This seems to end up at a point in the
3465 generated configure script that is early enough to affect the setting of
3466 compiler options, which is what is needed, but I have no means of testing
3467 whether it really works. (The user who reported this had patched the
3468 generated configure script, which of course I cannot do.)
3469
3470 LATER: After change 22 below (new libtool files), the configure script
3471 seems to know about icc (and also ecc). Therefore, I have commented out
3472 this hack in configure.in.
3473
3474 7. Added support for pkg-config (2 patches were sent in).
3475
3476 8. Negated POSIX character classes that used a combination of internal tables
3477 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3478 [[:^ascii]]. Typically, they would match almost any characters. The other
3479 POSIX classes were not broken in this way.
3480
3481 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3482 to find the match, as PCRE was deluded into thinking that the match had to
3483 start at the start point or following a newline. The same bug applied to
3484 patterns with negative forward assertions or any backward assertions
3485 preceding ".*" at the start, unless the pattern required a fixed first
3486 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3487
3488 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3489 starting at the last subject character, bytes beyond the end of the subject
3490 string were read.
3491
3492 11. Renamed the variable "class" as "classbits" to make life easier for C++
3493 users. (Previously there was a macro definition, but it apparently wasn't
3494 enough.)
3495
3496 12. Added the new field "tables" to the extra data so that tables can be passed
3497 in at exec time, or the internal tables can be re-selected. This allows
3498 a compiled regex to be saved and re-used at a later time by a different
3499 program that might have everything at different addresses.
3500
3501 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3502 -R library as well as a -L library.
3503
3504 14. The debugging options of pcretest (-d on the command line or D on a
3505 pattern) showed incorrect output for anything following an extended class
3506 that contained multibyte characters and which was followed by a quantifier.
3507
3508 15. Added optional support for general category Unicode character properties
3509 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3510 support. It adds about 90K to the size of the library. The meanings of the
3511 inbuilt class escapes such as \d and \s have NOT been changed.
3512
3513 16. Updated pcredemo.c to include calls to free() to release the memory for the
3514 compiled pattern.
3515
3516 17. The generated file chartables.c was being created in the source directory
3517 instead of in the building directory. This caused the build to fail if the
3518 source directory was different from the building directory, and was
3519 read-only.
3520
3521 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3522 file. No doubt somebody will tell me if they don't make sense... Also added
3523 Dan Mooney's comments about building on OpenVMS.
3524
3525 19. Added support for partial matching via the PCRE_PARTIAL option for
3526 pcre_exec() and the \P data escape in pcretest.
3527
3528 20. Extended pcretest with 3 new pattern features:
3529
3530 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3531 write the compiled pattern to the file whose name is "rest-of-line".
3532 This is a straight binary dump of the data, with the saved pointer to
3533 the character tables forced to be NULL. The study data, if any, is
3534 written too. After writing, pcretest reads a new pattern.
3535
3536 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3537 compiled pattern from the given file. There must not be any
3538 occurrences of "<" in the file name (pretty unlikely); if there are,
3539 pcretest will instead treat the initial "<" as a pattern delimiter.
3540 After reading in the pattern, pcretest goes on to read data lines as
3541 usual.
3542
3543 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3544 and 16-bit fields in a compiled pattern, to simulate a pattern that
3545 was compiled on a host of opposite endianness.
3546
3547 21. The pcre-exec() function can now cope with patterns that were compiled on
3548 hosts of opposite endianness, with this restriction:
3549
3550 As for any compiled expression that is saved and used later, the tables
3551 pointer field cannot be preserved; the extra_data field in the arguments
3552 to pcre_exec() should be used to pass in a tables address if a value
3553 other than the default internal tables were used at compile time.
3554
3555 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3556 now diagnosed as an error. Previously, most of the time, a negative number
3557 would have been treated as zero, but if in addition "ovector" was passed as
3558 NULL, a crash could occur.
3559
3560 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3561 new versions from the libtool 1.5 distribution (the last one is a copy of
3562 a file called libtool.m4). This seems to have fixed the need to patch
3563 "configure" to support Darwin 1.3 (which I used to do). However, I still
3564 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3565 workstation).
3566
3567 24. Changed the PCRE licence to be the more standard "BSD" licence.
3568
3569
3570 Version 4.5 01-Dec-03
3571 ---------------------
3572
3573 1. There has been some re-arrangement of the code for the match() function so
3574 that it can be compiled in a version that does not call itself recursively.
3575 Instead, it keeps those local variables that need separate instances for
3576 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3577 needs to "recurse". Keeping track of where control must go is done by means
3578 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3579 hide most of the details from the main code, and operates only if
3580 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3581 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3582 operating.
3583
3584 To make it easier for callers to provide specially tailored get/free
3585 functions for this usage, two new functions, pcre_stack_malloc, and
3586 pcre_stack_free, are used. They are always called in strict stacking order,
3587 and the size of block requested is always the same.
3588
3589 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3590 PCRE has been compiled to use the stack or the heap for recursion. The
3591 -C option of pcretest uses this to show which version is compiled.
3592
3593 A new data escape \S, is added to pcretest; it causes the amounts of store
3594 obtained and freed by both kinds of malloc/free at match time to be added
3595 to the output.
3596
3597 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3598 what's available on my current Linux desktop machine.
3599
3600 3. When matching a UTF-8 string, the test for a valid string at the start has
3601 been extended. If start_offset is not zero, PCRE now checks that it points
3602 to a byte that is the start of a UTF-8 character. If not, it returns
3603 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3604 this is necessary because there may be backward assertions in the pattern.
3605 When matching the same subject several times, it may save resources to use
3606 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3607
3608 4. The code for checking the validity of UTF-8 strings has been tightened so
3609 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3610 containing "overlong sequences".
3611
3612 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3613 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3614 should have been "&", but it just so happened that all the cases this let
3615 through by mistake were picked up later in the function.
3616
3617 6. I had used a variable called "isblank" - this is a C99 function, causing
3618 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3619
3620 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3621 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3622 so the version is shown; (c) stop "make test" echoing "./RunTest".
3623
3624 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3625
3626 9. The prototype for memmove() for systems that don't have it was using
3627 size_t, but the inclusion of the header that defines size_t was later. I've
3628 moved the #includes for the C headers earlier to avoid this.
3629
3630 10. Added some adjustments to the code to make it easier to compiler on certain
3631 special systems:
3632
3633 (a) Some "const" qualifiers were missing.
3634 (b) Added the macro EXPORT before all exported functions; by default this
3635 is defined to be empty.
3636 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3637 that it reads its output file name as an argument instead of writing
3638 to the standard output and assuming this can be redirected.
3639
3640 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3641 class containing characters with values greater than 255, PCRE compilation
3642 went into a loop.
3643
3644 12. A recursive reference to a subpattern that was within another subpattern
3645 that had a minimum quantifier of zero caused PCRE to crash. For example,
3646 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3647 recursion. If the recursively-called subpattern itself had a zero repeat,
3648 that was OK.
3649
3650 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3651 buffer into which it was copied (for escape processing) was still set at
3652 1024, so long lines caused crashes.
3653
3654 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3655 "internal error: code overflow...". This applied to any character class
3656 that was followed by a possessive quantifier.
3657
3658 15. Modified the Makefile to add libpcre.la as a prerequisite for
3659 libpcreposix.la because I was told this is needed for a parallel build to
3660 work.
3661
3662 16. If a pattern that contained .* following optional items at the start was
3663 studied, the wrong optimizing data was generated, leading to matching
3664 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3665 matching string must start with a or b or c. The correct conclusion for
3666 this pattern is that a match can start with any character.
3667
3668
3669 Version 4.4 13-Aug-03
3670 ---------------------
3671
3672 1. In UTF-8 mode, a character class containing characters with values between
3673 127 and 255 was not handled correctly if the compiled pattern was studied.
3674 In fixing this, I have also improved the studying algorithm for such
3675 classes (slightly).
3676
3677 2. Three internal functions had redundant arguments passed to them. Removal
3678 might give a very teeny performance improvement.
3679
3680 3. Documentation bug: the value of the capture_top field in a callout is *one
3681 more than* the number of the hightest numbered captured substring.
3682
3683 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3684 in incorrectly linking with a previously installed version. They now link
3685 explicitly with libpcre.la.
3686
3687 5. configure.in no longer needs to recognize Cygwin specially.
3688
3689 6. A problem in pcre.in for Windows platforms is fixed.
3690
3691 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3692 pcretest, it used to include the size of the study block as part of its
3693 output. Unfortunately, the structure contains a field that has a different
3694 size on different hardware architectures. This meant that the tests that
3695 showed this size failed. As the block is currently always of a fixed size,
3696 this information isn't actually particularly useful in pcretest output, so
3697 I have just removed it.
3698
3699 8. Three pre-processor statements accidentally did not start in column 1.
3700 Sadly, there are *still* compilers around that complain, even though
3701 standard C has not required this for well over a decade. Sigh.
3702
3703 9. In pcretest, the code for checking callouts passed small integers in the
3704 callout_data field, which is a void * field. However, some picky compilers
3705 complained about the casts involved for this on 64-bit systems. Now
3706 pcretest passes the address of the small integer instead, which should get
3707 rid of the warnings.
3708
3709 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3710 both compile and run time, and gives an error if an invalid UTF-8 sequence
3711 is found. There is a option for disabling this check in cases where the
3712 string is known to be correct and/or the maximum performance is wanted.
3713
3714 11. In response to a bug report, I changed one line in Makefile.in from
3715
3716 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3717 to
3718 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3719
3720 to look similar to other lines, but I have no way of telling whether this
3721 is the right thing to do, as I do not use Windows. No doubt I'll get told
3722 if it's wrong...
3723
3724
3725 Version 4.3 21-May-03
3726 ---------------------
3727
3728 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3729 Makefile.
3730
3731 2. Some refactoring to improve the quality of the code:
3732
3733 (i) The utf8_table... variables are now declared "const".
3734
3735 (ii) The code for \cx, which used the "case flipping" table to upper case
3736 lower case letters, now just substracts 32. This is ASCII-specific,
3737 but the whole concept of \cx is ASCII-specific, so it seems
3738 reasonable.
3739
3740 (iii) PCRE was using its character types table to recognize decimal and
3741 hexadecimal digits in the pattern. This is silly, because it handles
3742 only 0-9, a-f, and A-F, but the character types table is locale-
3743 specific, which means strange things might happen. A private
3744 table is now used for this - though it costs 256 bytes, a table is
3745 much faster than multiple explicit tests. Of course, the standard
3746 character types table is still used for matching digits in subject
3747 strings against \d.
3748
3749 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3750 ending in _t are). So I've renamed it as ESC_tee.
3751
3752 3. The first argument for regexec() in the POSIX wrapper should have been
3753 defined as "const".
3754
3755 4. Changed pcretest to use malloc() for its buffers so that they can be
3756 Electric Fenced for debugging.
3757
3758 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3759 to read one or more bytes before the start of the subject string. Often this
3760 had no effect on PCRE's behaviour, but in some circumstances it could
3761 provoke a segmentation fault.
3762
3763 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3764 to try to read one or more bytes before the start of the subject string.
3765
3766 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3767 UTF-8 support could misbehave in various ways if the subject string
3768 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3769 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3770 back over UTF-8 characters.)
3771
3772
3773 Version 4.2 14-Apr-03
3774 ---------------------
3775
3776 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3777
3778 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3779 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3780 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3781 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3782 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3783 and BUILD_EXEEXT
3784 Note: automatic setting of the BUILD variables is not yet working
3785 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3786 compile-time but not at link-time
3787 [LINK]: use for linking executables only
3788 make different versions for Windows and non-Windows
3789 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3790 libraries
3791 [LINK_FOR_BUILD]: new variable
3792 [OBJEXT]: use throughout
3793 [EXEEXT]: use throughout
3794 <winshared>: new target
3795 <wininstall>: new target
3796 <dftables.o>: use native compiler
3797 <dftables>: use native linker
3798 <install>: handle Windows platform correctly
3799 <clean>: ditto
3800 <check>: ditto
3801 copy DLL to top builddir before testing
3802
3803 As part of these changes, -no-undefined was removed again. This was reported
3804 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3805 in any case.
3806
3807 3. Some tidies to get rid of compiler warnings:
3808
3809 . In the match_data structure, match_limit was an unsigned long int, whereas
3810 match_call_count was an int. I've made them both unsigned long ints.
3811
3812 . In pcretest the fact that a const uschar * doesn't automatically cast to
3813 a void * provoked a warning.
3814
3815 . Turning on some more compiler warnings threw up some "shadow" variables
3816 and a few more missing casts.
3817
3818 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3819 option, a class that contained a single character with a value between 128
3820 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3821
3822 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3823 option, a class that contained several characters, but with at least one
3824 whose value was between 128 and 255 caused PCRE to crash.
3825
3826
3827 Version 4.1 12-Mar-03
3828 ---------------------
3829
3830 1. Compiling with gcc -pedantic found a couple of places where casts were
3831 needed, and a string in dftables.c that was longer than standard compilers are
3832 required to support.
3833
3834 2. Compiling with Sun's compiler found a few more places where the code could
3835 be tidied up in order to avoid warnings.
3836
3837 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3838 first of these names is deprecated in the latest Autoconf in favour of the name
3839 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3840 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3841 analogy I have changed it to CFLAGS_FOR_BUILD.
3842
3843 4. Added -no-undefined to the linking command in the Makefile, because this is
3844 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3845 linking step for the pcreposix library.
3846
3847 5. PCRE was failing to diagnose the case of two named groups with the same
3848 name.
3849
3850 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3851 literal character that is needed in the subject for a match, and scans along to
3852 ensure that it is present before embarking on the full matching process. This
3853 saves time in cases of nested unlimited repeats that are never going to match.
3854 Problem: the scan can take a lot of time if the subject is very long (e.g.
3855 megabytes), thus penalizing straightforward matches. It is now done only if the
3856 amount of subject to be scanned is less than 1000 bytes.
3857
3858 7. A lesser problem with the same optimization is that it was recording the
3859 first character of an anchored pattern as "needed", thus provoking a search
3860 right along the subject, even when the first match of the pattern was going to
3861 fail. The "needed" character is now not set for anchored patterns, unless it
3862 follows something in the pattern that is of non-fixed length. Thus, it still
3863 fulfils its original purpose of finding quick non-matches in cases of nested
3864 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3865
3866
3867 Version 4.0 17-Feb-03
3868 ---------------------
3869
3870 1. If a comment in an extended regex that started immediately after a meta-item
3871 extended to the end of string, PCRE compiled incorrect data. This could lead to
3872 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3873
3874 2. Moved to autoconf 2.53 and libtool 1.4.2.
3875
3876 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3877 the special perltest8 script is no longer needed - all the tests can be run
3878 from a single perltest script.
3879
3880 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3881 by \s. It has now been removed in PCRE. This means it isn't recognized as
3882 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3883 class [:space:] *does* include VT, thereby creating a mess.
3884
3885 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3886 space and tab.
3887
3888 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3889 its new features into the main test script, reducing the number of scripts.
3890
3891 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3892 were backward compatible, and made the (?i) apply to the whole pattern, as if
3893 /i were given. Now it behaves more logically, and applies the option setting
3894 only to what follows. PCRE has been changed to follow suit. However, if it
3895 finds options settings right at the start of the pattern, it extracts them into
3896 the global options, as before. Thus, they show up in the info data.
3897
3898 8. Added support for the \Q...\E escape sequence. Characters in between are
3899 treated as literals. This is slightly different from Perl in that $ and @ are
3900 also handled as literals inside the quotes. In Perl, they will cause variable
3901 interpolation. Note the following examples:
3902
3903 Pattern PCRE matches Perl matches
3904
3905 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3906 \Qabc\$xyz\E abc\$xyz abc\$xyz
3907 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3908
3909 For compatibility with Perl, \Q...\E sequences are recognized inside character
3910 classes as well as outside them.
3911
3912 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3913 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3914 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3915 signed/unsigned warnings.
3916
3917 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3918 option for pcretest, so I've replaced it by a simple function that does just
3919 that job.
3920
3921 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3922 "pcregrep -".
3923
3924 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3925 Java package. This provides some syntactic sugar for simple cases of what my
3926 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3927 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3928 item, you can use this simplified notation. Note that only makes sense with
3929 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3930 greediness, whatever the setting of the PCRE_UNGREEDY option.
3931
3932 13. A change of greediness default within a pattern was not taking effect at
3933 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3934 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3935 was abstracted outside.
3936
3937 14. PCRE now supports the \G assertion. It is true when the current matching
3938 position is at the start point of the match. This differs from \A when the
3939 starting offset is non-zero. Used with the /g option of pcretest (or similar
3940 code), it works in the same way as it does for Perl's /g option. If all
3941 alternatives of a regex begin with \G, the expression is anchored to the start
3942 match position, and the "anchored" flag is set in the compiled expression.
3943
3944 15. Some bugs concerning the handling of certain option changes within patterns
3945 have been fixed. These applied to options other than (?ims). For example,
3946 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3947 been the other way round. Some of this was related to change 7 above.
3948
3949 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3950 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3951 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3952 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3953
3954 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3955 mode. Unlike ".", it always matches newline, whatever the setting of
3956 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3957 assertions. Perl allows it, but it doesn't (in general) work because it can't
3958 calculate the length of the lookbehind. At least, that's the case for Perl
3959 5.8.0 - I've been told they are going to document that it doesn't work in
3960 future.
3961
3962 18. Added an error diagnosis for escapes that PCRE does not support: these are
3963 \L, \l, \N, \P, \p, \U, \u, and \X.
3964
3965 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3966 reading past the end of the pattern in cases such as /[abcd/.
3967
3968 20. PCRE was getting more memory than necessary for patterns with classes that
3969 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3970
3971 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3972 compiling PCRE for use with Virtual Pascal.
3973
3974 22. Small fix to the Makefile to make it work properly if the build is done
3975 outside the source tree.
3976
3977 23. Added a new extension: a condition to go with recursion. If a conditional
3978 subpattern starts with (?(R) the "true" branch is used if recursion has
3979 happened, whereas the "false" branch is used only at the top level.
3980
3981 24. When there was a very long string of literal characters (over 255 bytes
3982 without UTF support, over 250 bytes with UTF support), the computation of how
3983 much memory was required could be incorrect, leading to segfaults or other
3984 strange effects.
3985
3986 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3987 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3988 there was a subsequent back reference to those brackets. This meant that, for
3989 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3990 possible to check for precisely this case. All we can do is abandon the
3991 optimization if .* occurs inside capturing brackets when there are any back
3992 references whatsoever. (See below for a better fix that came later.)
3993
3994 26. The handling of the optimization for finding the first character of a
3995 non-anchored pattern, and for finding a character that is required later in the
3996 match were failing in some cases. This didn't break the matching; it just
3997 failed to optimize when it could. The way this is done has been re-implemented.
3998
3999 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4000
4001 28. Added a new feature that provides some of the functionality that Perl
4002 provides with (?{...}). The facility is termed a "callout". The way it is done
4003 in PCRE is for the caller to provide an optional function, by setting
4004 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4005 global variable. By default it is unset, which disables all calling out. To get
4006 the function called, the regex must include (?C) at appropriate points. This
4007 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4008 This provides a means of identifying different callout points. When PCRE
4009 reaches such a point in the regex, if pcre_callout has been set, the external
4010 function is called. It is provided with data in a structure called
4011 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4012 matching continues; if it returns a non-zero value, the match at the current
4013 point fails. However, backtracking will occur if possible. [This was changed
4014 later and other features added - see item 49 below.]
4015
4016 29. pcretest is upgraded to test the callout functionality. It provides a
4017 callout function that displays information. By default, it shows the start of
4018 the match and the current position in the text. There are some new data escapes
4019 to vary what happens:
4020
4021 \C+ in addition, show current contents of captured substrings
4022 \C- do not supply a callout function
4023 \C!n return 1 when callout number n is reached
4024 \C!n!m return 1 when callout number n is reached for the mth time
4025
4026 30. If pcregrep was called with the -l option and just a single file name, it
4027 output "<stdin>" if a match was found, instead of the file name.
4028
4029 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4030 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4031 pcre_exec(). This saves a malloc/free per call. The default value of
4032 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4033 when configuring.
4034
4035 32. The default maximum size of a compiled pattern is 64K. There have been a
4036 few cases of people hitting this limit. The code now uses macros to handle the
4037 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4038 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4039 configuring. Tests 2 and 5 work only with 2-byte links because they output
4040 debugging information about compiled patterns.
4041
4042 33. Internal code re-arrangements:
4043
4044 (a) Moved the debugging function for printing out a compiled regex into
4045 its own source file (printint.c) and used #include to pull it into
4046 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4047 separate copies.
4048
4049 (b) Defined the list of op-code names for debugging as a macro in
4050 internal.h so that it is next to the definition of the opcodes.
4051
4052 (c) Defined a table of op-code lengths for simpler skipping along compiled
4053 code. This is again a macro in internal.h so that it is next to the
4054 definition of the opcodes.
4055
4056 34. Added support for recursive calls to individual subpatterns, along the
4057 lines of Robin Houston's patch (but implemented somewhat differently).
4058
4059 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4060 allow it to read and process whole directories in Win32. This code was
4061 contributed by Lionel Fourquaux; it has not been tested by me.
4062
4063 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4064 used to name a group. Names consist of alphanumerics and underscores, and must
4065 be unique. Back references use the syntax (?P=name) and recursive calls use
4066 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4067 numbers. The function pcre_fullinfo() can be used after compilation to extract
4068 a name/number map. There are three relevant calls:
4069
4070 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4071 PCRE_INFO_NAMECOUNT yields the number of entries
4072 PCRE_INFO_NAMETABLE yields a pointer to the map.
4073
4074 The map is a vector of fixed-size entries. The size of each entry depends on
4075 the length of the longest name used. The first two bytes of each entry are the
4076 group number, most significant byte first. There follows the corresponding
4077 name, zero terminated. The names are in alphabetical order.
4078
4079 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4080 case instead of 255. Making it the same both with and without UTF-8 support
4081 means that the same test output works with both.
4082
4083 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4084 calling malloc() with a zero argument.
4085
4086 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4087 optimization. I've improved things by keeping a bitmap of backreferences with
4088 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4089 fact referenced, the optimization can be applied. It is unlikely that a
4090 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4091 the match to follow \n) will appear inside brackets with a number greater than
4092 31, but if it does, any back reference > 31 suppresses the optimization.
4093
4094 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4095 of disabling numbered capturing parentheses. Any opening parenthesis that is
4096 not followed by ? behaves as if it were followed by ?: but named parentheses
4097 can still be used for capturing (and they will acquire numbers in the usual
4098 way).
4099
4100 41. Redesigned the return codes from the match() function into yes/no/error so
4101 that errors can be passed back from deep inside the nested calls. A malloc
4102 failure while inside a recursive subpattern call now causes the
4103 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4104
4105 42. It is now possible to set a limit on the number of times the match()
4106 function is called in a call to pcre_exec(). This facility makes it possible to
4107 limit the amount of recursion and backtracking, though not in a directly
4108 obvious way, because the match() function is used in a number of different
4109 circumstances. The count starts from zero for each position in the subject
4110 string (for non-anchored patterns). The default limit is, for compatibility, a
4111 large number, namely 10 000 000. You can change this in two ways:
4112
4113 (a) When configuring PCRE before making, you can use --with-match-limit=n
4114 to set a default value for the compiled library.
4115
4116 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4117 a different value is set. See 45 below.
4118
4119 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4120
4121 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4122 of things that can be changed at compile time. The first argument specifies
4123 what is wanted and the second points to where the information is to be placed.
4124 The current list of available information is:
4125
4126 PCRE_CONFIG_UTF8
4127
4128 The output is an integer that is set to one if UTF-8 support is available;
4129 otherwise it is set to zero.
4130
4131 PCRE_CONFIG_NEWLINE
4132
4133 The output is an integer that it set to the value of the code that is used for
4134 newline. It is either LF (10) or CR (13).
4135
4136 PCRE_CONFIG_LINK_SIZE
4137
4138 The output is an integer that contains the number of bytes used for internal
4139 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4140
4141 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4142
4143 The output is an integer that contains the threshold above which the POSIX
4144 interface uses malloc() for output vectors. See item 31 above.
4145
4146 PCRE_CONFIG_MATCH_LIMIT
4147
4148 The output is an unsigned integer that contains the default limit of the number
4149 of match() calls in a pcre_exec() execution. See 42 above.
4150
4151 44. pcretest has been upgraded by the addition of the -C option. This causes it
4152 to extract all the available output from the new pcre_config() function, and to
4153 output it. The program then exits immediately.
4154
4155 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4156 order to support additional features. One way would have been to define
4157 pcre_exec2() (for example) with extra arguments, but this would not have been
4158 extensible, and would also have required all calls to the original function to
4159 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4160 is used for passing in "extra" data from pcre_study().
4161
4162 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4163 contains the following fields:
4164
4165 flags a bitmap indicating which of the following fields are set
4166 study_data opaque data from pcre_study()
4167 match_limit a way of specifying a limit on match() calls for a specific
4168 call to pcre_exec()
4169 callout_data data for callouts (see 49 below)
4170
4171 The flag bits are also defined in pcre.h, and are
4172
4173 PCRE_EXTRA_STUDY_DATA
4174 PCRE_EXTRA_MATCH_LIMIT
4175 PCRE_EXTRA_CALLOUT_DATA
4176
4177 The pcre_study() function now returns one of these new pcre_extra blocks, with
4178 the actual study data pointed to by the study_data field, and the
4179 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4180 before. That is, this change is entirely upwards-compatible and requires no
4181 change to existing code.
4182
4183 If you want to pass in additional data to pcre_exec(), you can either place it
4184 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4185 block.
4186
4187 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4188 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4189 times with different match limits, until it finds the minimum value needed for
4190 pcre_exec() to complete. The value is then output. This can be instructive; for
4191 most simple matches the number is quite small, but for pathological cases it
4192 gets very large very quickly.
4193
4194 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4195 returns the size of the data block pointed to by the study_data field in a
4196 pcre_extra block, that is, the value that was passed as the argument to
4197 pcre_malloc() when PCRE was getting memory in which to place the information
4198 created by pcre_study(). The fourth argument should point to a size_t variable.
4199 pcretest has been extended so that this information is shown after a successful
4200 pcre_study() call when information about the compiled regex is being displayed.
4201
4202 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4203 because what follows is always an absolute path. (Later: it turns out that this
4204 is more than cosmetic for MinGW, because it doesn't like empty path
4205 components.)
4206
4207 49. Some changes have been made to the callout feature (see 28 above):
4208
4209 (i) A callout function now has three choices for what it returns:
4210
4211 0 => success, carry on matching
4212 > 0 => failure at this point, but backtrack if possible
4213 < 0 => serious error, return this value from pcre_exec()
4214
4215 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4216 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4217 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4218 use by callout functions. It will never be used by PCRE itself.
4219
4220 (ii) The pcre_extra structure (see 45 above) has a void * field called
4221 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4222 pcre_callout_block structure has a field of the same name. The contents of
4223 the field passed in the pcre_extra structure are passed to the callout
4224 function in the corresponding field in the callout block. This makes it
4225 easier to use the same callout-containing regex from multiple threads. For
4226 testing, the pcretest program has a new data escape
4227
4228 \C*n pass the number n (may be negative) as callout_data
4229
4230 If the callout function in pcretest receives a non-zero value as
4231 callout_data, it returns that value.
4232
4233 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4234 there were some redundant $(CFLAGS) in commands that are now specified as
4235 $(LINK), which already includes $(CFLAGS).
4236
4237 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4238 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4239 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4240 one-byte characters throughout. Note that case-insensitive matching applies
4241 only to characters whose values are less than 256. PCRE doesn't support the
4242 notion of cases for higher-valued characters.
4243
4244 (i) A character class whose characters are all within 0-255 is handled as
4245 a bit map, and the map is inverted for negative classes. Previously, a
4246 character > 255 always failed to match such a class; however it should
4247 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4248
4249 (ii) A negated character class with a single character < 255 is coded as
4250 "not this character" (OP_NOT). This wasn't working properly when the test
4251 character was multibyte, either singly or repeated.
4252
4253 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4254 mode, for example: \x{100}{2,3}.
4255
4256 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4257 singly or repeated) now correctly test multibyte characters. However,
4258 PCRE doesn't recognize any characters with values greater than 255 as
4259 digits, spaces, or word characters. Such characters always match \D, \S,
4260 and \W, and never match \d, \s, or \w.
4261
4262 (v) Classes may now contain characters and character ranges with values
4263 greater than 255. For example: [ab\x{100}-\x{400}].
4264
4265 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4266 PCRE in UTF-8 mode.
4267
4268 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4269 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4270 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4271 value.)
4272
4273 53. The single man page has become too large. I have therefore split it up into
4274 a number of separate man pages. These also give rise to individual HTML pages;
4275 these are now put in a separate directory, and there is an index.html page that
4276 lists them all. Some hyperlinking between the pages has been installed.
4277
4278 54. Added convenience functions for handling named capturing parentheses.
4279
4280 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4281 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4282 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4283 are faulted.
4284
4285 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4286 calling configure. These values are used when compiling the dftables.c program
4287 which is run to generate the source of the default character tables. They
4288 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4289 you will need to set these values.
4290
4291 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4292
4293
4294 Version 3.9 02-Jan-02
4295 ---------------------
4296
4297 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4298
4299 2. If --disable-static was given, the building process failed when trying to
4300 build pcretest and pcregrep. (For some reason it was using libtool to compile
4301 them, which is not right, as they aren't part of the library.)
4302
4303
4304 Version 3.8 18-Dec-01
4305 ---------------------
4306
4307 1. The experimental UTF-8 code was completely screwed up. It was packing the
4308 bytes in the wrong order. How dumb can you get?
4309
4310
4311 Version 3.7 29-Oct-01
4312 ---------------------
4313
4314 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4315 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4316 this didn't happen under Solaris 8, where I normally test things.
4317
4318 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4319 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4320 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4321 this fix an hour or so after the initial 3.7 release.)
4322
4323
4324 Version 3.6 23-Oct-01
4325 ---------------------
4326
4327 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4328 offsets passed as NULL with zero offset count.
4329
4330 2. The config.guess and config.sub files had not been updated when I moved to
4331 the latest autoconf.
4332
4333
4334 Version 3.5 15-Aug-01
4335 ---------------------
4336
4337 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4338 had been forgotten.
4339
4340 2. By using declared but undefined structures, we can avoid using "void"
4341 definitions in pcre.h while keeping the internal definitions of the structures
4342 private.
4343
4344 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4345 user point of view, this means that both static and shared libraries are built
4346 by default, but this can be individually controlled. More of the work of
4347 handling this static/shared cases is now inside libtool instead of PCRE's make
4348 file.
4349
4350 4. The pcretest utility is now installed along with pcregrep because it is
4351 useful for users (to test regexs) and by doing this, it automatically gets
4352 relinked by libtool. The documentation has been turned into a man page, so
4353 there are now .1, .txt, and .html versions in /doc.
4354
4355 5. Upgrades to pcregrep:
4356 (i) Added long-form option names like gnu grep.
4357 (ii) Added --help to list all options with an explanatory phrase.
4358 (iii) Added -r, --recursive to recurse into sub-directories.
4359 (iv) Added -f, --file to read patterns from a file.
4360
4361 6. pcre_exec() was referring to its "code" argument before testing that
4362 argument for NULL (and giving an error if it was NULL).
4363
4364 7. Upgraded Makefile.in to allow for compiling in a different directory from
4365 the source directory.
4366
4367 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4368 options bits, the pointer it was passed was to an int instead of to an unsigned
4369 long int. This mattered only on 64-bit systems.
4370
4371 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4372 generated) instead of pcre.in, which it its source. Also made the same change
4373 in several of the .c files.
4374
4375 10. A new release of gcc defines printf() as a macro, which broke pcretest
4376 because it had an ifdef in the middle of a string argument for printf(). Fixed
4377 by using separate calls to printf().
4378
4379 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4380 script, to force use of CR or LF instead of \n in the source. On non-Unix
4381 systems, the value can be set in config.h.
4382
4383 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4384 absolute limit. Changed the text of the error message to make this clear, and
4385 likewise updated the man page.
4386
4387 13. The limit of 99 on the number of capturing subpatterns has been removed.
4388 The new limit is 65535, which I hope will not be a "real" limit.
4389
4390
4391 Version 3.4 22-Aug-00
4392 ---------------------
4393
4394 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4395
4396 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4397
4398
4399 Version 3.3 01-Aug-00
4400 ---------------------
4401
4402 1. If an octal character was given, but the value was greater than \377, it
4403 was not getting masked to the least significant bits, as documented. This could
4404 lead to crashes in some systems.
4405
4406 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4407 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4408
4409 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4410 These just pass their arguments on to (pcre_free)(), but they are provided
4411 because some uses of PCRE bind it to non-C systems that can call its functions,
4412 but cannot call free() or pcre_free() directly.
4413
4414 4. Add "make test" as a synonym for "make check". Corrected some comments in
4415 the Makefile.
4416
4417 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4418 Makefile.
4419
4420 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4421 command called pgrep for grepping around the active processes.
4422
4423 7. Added the beginnings of support for UTF-8 character strings.
4424
4425 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4426 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4427 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4428 out for the ar command.)
4429
4430
4431 Version 3.2 12-May-00
4432 ---------------------
4433
4434 This is purely a bug fixing release.
4435
4436 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4437 of ZA. This was just one example of several cases that could provoke this bug,
4438 which was introduced by change 9 of version 2.00. The code for breaking
4439 infinite loops after an iteration that matches an empty string was't working
4440 correctly.
4441
4442 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4443 when matched against abbab (for example). After matching an empty string, it
4444 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4445 caused it to match further down the string than it should.
4446
4447 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4448 was there because it doesn't seem to be needed, and it causes trouble on some
4449 systems, as it is not a Standard C header. It has been removed.
4450
4451 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4452 were reported on the Macintosh. The changes were from
4453
4454 while ((c = *(++ptr)) != 0 && c != '\n');
4455 to
4456 while ((c = *(++ptr)) != 0 && c != '\n') ;
4457
4458 Totally extraordinary, but if that's what it takes...
4459
4460 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4461 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4462 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4463 assumes the way PCRE uses memmove() (always moving upwards).
4464
4465 6. PCRE is being used in one environment where strchr() is not available. There
4466 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4467 faster code anyway.
4468
4469
4470 Version 3.1 09-Feb-00
4471 ---------------------
4472
4473 The only change in this release is the fixing of some bugs in Makefile.in for
4474 the "install" target:
4475
4476 (1) It was failing to install pcreposix.h.
4477
4478 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4479
4480
4481 Version 3.0 01-Feb-00
4482 ---------------------
4483
4484 1. Add support for the /+ modifier to perltest (to output $` like it does in
4485 pcretest).
4486
4487 2. Add support for the /g modifier to perltest.
4488
4489 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4490 matches null strings.
4491
4492 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4493 pattern. Perl treats empty patterns specially - it reuses the most recent
4494 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4495 effect.
4496
4497 5. The POSIX interface was broken in that it was just handing over the POSIX
4498 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4499 required a bigger vector, with some working space on the end. This means that
4500 the POSIX wrapper now has to get and free some memory, and copy the results.
4501
4502 6. Added some simple autoconf support, placing the test data and the
4503 documentation in separate directories, re-organizing some of the
4504 information files, and making it build pcre-config (a GNU standard). Also added
4505 libtool support for building PCRE as a shared library, which is now the
4506 default.
4507
4508 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4509 09 are not valid octal constants. Single digits will be used for minor values
4510 less than 10.
4511
4512 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4513 existing programs that set these in the POSIX interface can use PCRE without
4514 modification.
4515
4516 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4517 return all that pcre_info() returns, plus additional data. The pcre_info()
4518 function is retained for compatibility, but is considered to be obsolete.
4519
4520 10. Added experimental recursion feature (?R) to handle one common case that
4521 Perl 5.6 will be able to do with (?p{...}).
4522
4523 11. Added support for POSIX character classes like [:alpha:], which Perl is
4524 adopting.
4525
4526
4527 Version 2.08 31-Aug-99
4528 ----------------------
4529
4530 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4531 trying to match at the startoffset position, but instead was moving forward to
4532 the next newline as if a previous match had failed.
4533
4534 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4535 and could get into a loop if a null string was matched other than at the start
4536 of the subject.
4537
4538 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4539 be distinguished at compile time, and for completeness also added PCRE_DATE.
4540
4541 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4542 in GnuWin32 environments.
4543
4544
4545 Version 2.07 29-Jul-99
4546 ----------------------
4547
4548 1. The documentation is now supplied in plain text form and HTML as well as in
4549 the form of man page sources.
4550
4551 2. C++ compilers don't like assigning (void *) values to other pointer types.
4552 In particular this affects malloc(). Although there is no problem in Standard
4553 C, I've put in casts to keep C++ compilers happy.
4554
4555 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4556 should be (const char *).
4557
4558 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4559 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4560 However, I haven't made this a standard facility. The documentation doesn't
4561 mention it, and the Makefile doesn't support it.
4562
4563 5. The Makefile now contains an "install" target, with editable destinations at
4564 the top of the file. The pcretest program is not installed.
4565
4566 6. pgrep -V now gives the PCRE version number and date.
4567
4568 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4569 causing the entire string to be ignored, instead of just the last character.
4570
4571 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4572 non-matching string, it can take a very, very long time, even for strings of
4573 quite modest length, because of the nested recursion. PCRE now does better in
4574 some of these cases. It does this by remembering the last required literal
4575 character in the pattern, and pre-searching the subject to ensure it is present
4576 before running the real match. In other words, it applies a heuristic to detect
4577 some types of certain failure quickly, and in the above example, if presented
4578 with a string that has no trailing " it gives "no match" very quickly.
4579
4580 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4581 other alternatives are tried instead.
4582
4583
4584 Version 2.06 09-Jun-99
4585 ----------------------
4586
4587 1. Change pcretest's output for amount of store used to show just the code
4588 space, because the remainder (the data block) varies in size between 32-bit and
4589 64-bit systems.
4590
4591 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4592 start matching at. This allows lookbehinds to work when searching for multiple
4593 occurrences in a string.
4594
4595 3. Added additional options to pcretest for testing multiple occurrences:
4596
4597 /+ outputs the rest of the string that follows a match
4598 /g loops for multiple occurrences, using the new startoffset argument
4599 /G loops for multiple occurrences by passing an incremented pointer
4600
4601 4. PCRE wasn't doing the "first character" optimization for patterns starting
4602 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4603 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4604 the letter 'x'. On long subject strings, this gives a significant speed-up.
4605
4606
4607 Version 2.05 21-Apr-99
4608 ----------------------
4609
4610 1. Changed the type of magic_number from int to long int so that it works
4611 properly on 16-bit systems.
4612
4613 2. Fixed a bug which caused patterns starting with .* not to work correctly
4614 when the subject string contained newline characters. PCRE was assuming
4615 anchoring for such patterns in all cases, which is not correct because .* will
4616 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4617 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4618 must be retried after every newline in the subject.
4619
4620
4621 Version 2.04 18-Feb-99
4622 ----------------------
4623
4624 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4625 computation of the store needed to hold the pattern was incorrect (too large).
4626 If such patterns were nested a few deep, this could multiply and become a real
4627 problem.
4628
4629 2. Added /M option to pcretest to show the memory requirement of a specific
4630 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4631
4632 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4633 compiled in such a way that the backtracking after subsequent failure was
4634 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4635 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4636
4637
4638 Version 2.03 02-Feb-99
4639 ----------------------
4640
4641 1. Fixed typo and small mistake in man page.
4642
4643 2. Added 4th condition (GPL supersedes if conflict) and created separate
4644 LICENCE file containing the conditions.
4645
4646 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4647 Perl, that is the internal \ allows the delimiter to be included in the
4648 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4649 the final delimiter, add \ to the end of the pattern (to test the error).
4650
4651 4. Added the convenience functions for extracting substrings after a successful
4652 match. Updated pcretest to make it able to test these functions.
4653
4654
4655 Version 2.02 14-Jan-99
4656 ----------------------
4657
4658 1. Initialized the working variables associated with each extraction so that
4659 their saving and restoring doesn't refer to uninitialized store.
4660
4661 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4662 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4663 fix the problem.
4664
4665 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4666 calls, and wasn't printing the correct value for compiling calls. Increased the
4667 default value of LOOPREPEAT, and the number of significant figures in the
4668 times.
4669
4670 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4671
4672 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4673 a building problem on Windows NT with a FAT file system.
4674
4675
4676 Version 2.01 21-Oct-98
4677 ----------------------
4678
4679 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4680 to character tables built by pcre_maketables() in the current locale. If NULL
4681 is passed, the default tables are used.
4682
4683
4684 Version 2.00 24-Sep-98
4685 ----------------------
4686
4687 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4688 it any more.
4689
4690 2. Allow quantification of (?>) groups, and make it work correctly.
4691
4692 3. The first character computation wasn't working for (?>) groups.
4693
4694 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4695 end of the subject) and add 5.005's \z, which really does match only at the
4696 very end of the subject.
4697
4698 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4699
4700 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4701 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4702 localized options. All options to pcre_study() were also removed.
4703
4704 7. Add other new features from 5.005:
4705
4706 $(?<= positive lookbehind
4707 $(?<! negative lookbehind
4708 (?imsx-imsx) added the unsetting capability
4709 such a setting is global if at outer level; local otherwise
4710 (?imsx-imsx:) non-capturing groups with option setting
4711 (?(cond)re|re) conditional pattern matching
4712
4713 A backreference to itself in a repeated group matches the previous
4714 captured string.
4715
4716 8. General tidying up of studying (both automatic and via "study")
4717 consequential on the addition of new assertions.
4718
4719 9. As in 5.005, unlimited repeated groups that could match an empty substring
4720 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4721 runtime if any iteration does actually match an empty substring.
4722
4723 10. Include the RunTest script in the distribution.
4724
4725 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4726 discrepancies, some of which were old and were also with respect to 5.004. They
4727 have now been fixed.
4728
4729
4730 Version 1.09 28-Apr-98
4731 ----------------------
4732
4733 1. A negated single character class followed by a quantifier with a minimum
4734 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4735 program crashes, or just wrong answers. This did not apply to negated classes
4736 containing more than one character, or to minima other than one.
4737
4738
4739 Version 1.08 27-Mar-98
4740 ----------------------
4741
4742 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4743
4744 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4745 latter must appear before anything that relies on it in the pattern.
4746
4747
4748 Version 1.07 16-Feb-98
4749 ----------------------
4750
4751 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4752 repeat of a potentially empty string).
4753
4754
4755 Version 1.06 23-Jan-98
4756 ----------------------
4757
4758 1. Added Markus Oberhumer's little patches for C++.
4759
4760 2. Literal strings longer than 255 characters were broken.
4761
4762
4763 Version 1.05 23-Dec-97
4764 ----------------------
4765
4766 1. Negated character classes containing more than one character were failing if
4767 PCRE_CASELESS was set at run time.
4768
4769
4770 Version 1.04 19-Dec-97
4771 ----------------------
4772
4773 1. Corrected the man page, where some "const" qualifiers had been omitted.
4774
4775 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4776 input syntax.
4777
4778 3. Fixed memory leak which occurred when a regex with back references was
4779 matched with an offsets vector that wasn't big enough. The temporary memory
4780 that is used in this case wasn't being freed if the match failed.
4781
4782 4. Tidied pcretest to ensure it frees memory that it gets.
4783
4784 5. Temporary memory was being obtained in the case where the passed offsets
4785 vector was exactly big enough.
4786
4787 6. Corrected definition of offsetof() from change 5 below.
4788
4789 7. I had screwed up change 6 below and broken the rules for the use of
4790 setjmp(). Now fixed.
4791
4792
4793 Version 1.03 18-Dec-97
4794 ----------------------
4795
4796 1. A erroneous regex with a missing opening parenthesis was correctly
4797 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4798 on some systems.
4799
4800 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4801 it was reported that one broken compiler failed on the former because "code" is
4802 also an independent variable.
4803
4804 3. The erroneous regex a[]b caused an array overrun reference.
4805
4806 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4807 fail on data ending with that character. (It was going on too far, and checking
4808 the next character, typically a binary zero.) This was specific to the
4809 optimized code for single-character negative classes.
4810
4811 5. Added a contributed patch from the TIN world which does the following:
4812
4813 + Add an undef for memmove, in case the the system defines a macro for it.
4814
4815 + Add a definition of offsetof(), in case there isn't one. (I don't know
4816 the reason behind this - offsetof() is part of the ANSI standard - but
4817 it does no harm).
4818
4819 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4820 most of the places where whitespace preceded '#'. I have given up and
4821 allowed the remaining 2 cases to be at the margin.
4822
4823 + Rename some variables in pcre to eliminate shadowing. This seems very
4824 pedantic, but does no harm, of course.
4825
4826 6. Moved the call to setjmp() into its own function, to get rid of warnings
4827 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4828
4829 7. Constructs such as \d{8,} were compiling into the equivalent of
4830 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4831 outcome, but in this particular case used more store than had been allocated,
4832 which caused the bug to be discovered because it threw up an internal error.
4833
4834 8. The debugging code in both pcre and pcretest for outputting the compiled
4835 form of a regex was going wrong in the case of back references followed by
4836 curly-bracketed repeats.
4837
4838
4839 Version 1.02 12-Dec-97
4840 ----------------------
4841
4842 1. Typos in pcre.3 and comments in the source fixed.
4843
4844 2. Applied a contributed patch to get rid of places where it used to remove
4845 'const' from variables, and fixed some signed/unsigned and uninitialized
4846 variable warnings.
4847
4848 3. Added the "runtest" target to Makefile.
4849
4850 4. Set default compiler flag to -O2 rather than just -O.
4851
4852
4853 Version 1.01 19-Nov-97
4854 ----------------------
4855
4856 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4857 like /([ab]*)*/, that is, for classes with more than one character in them.
4858
4859 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4860 as /((?>a*))*/ (a PCRE_EXTRA facility).
4861
4862
4863 Version 1.00 18-Nov-97
4864 ----------------------
4865
4866 1. Added compile-time macros to support systems such as SunOS4 which don't have
4867 memmove() or strerror() but have other things that can be used instead.
4868
4869 2. Arranged that "make clean" removes the executables.
4870
4871
4872 Version 0.99 27-Oct-97
4873 ----------------------
4874
4875 1. Fixed bug in code for optimizing classes with only one character. It was
4876 initializing a 32-byte map regardless, which could cause it to run off the end
4877 of the memory it had got.
4878
4879 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4880
4881
4882 Version 0.98 22-Oct-97
4883 ----------------------
4884
4885 1. Fixed bug in code for handling temporary memory usage when there are more
4886 back references than supplied space in the ovector. This could cause segfaults.
4887
4888
4889 Version 0.97 21-Oct-97
4890 ----------------------
4891
4892 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4893
4894 2. Optimized negated single characters not to use a bit map.
4895
4896 3. Brought error texts together as macro definitions; clarified some of them;
4897 fixed one that was wrong - it said "range out of order" when it meant "invalid
4898 escape sequence".
4899
4900 4. Changed some char * arguments to const char *.
4901
4902 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4903
4904 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4905 pcretest.
4906
4907
4908 Version 0.96 16-Oct-97
4909 ----------------------
4910
4911 1. Added a simple "pgrep" utility to the distribution.
4912
4913 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4914 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4915 where "ddd" means "one or more decimal digits".
4916
4917 3. Fixed serious bug. If a pattern had a back reference, but the call to
4918 pcre_exec() didn't supply a large enough ovector to record the related
4919 identifying subpattern, the match always failed. PCRE now remembers the number
4920 of the largest back reference, and gets some temporary memory in which to save
4921 the offsets during matching if necessary, in order to ensure that
4922 backreferences always work.
4923
4924 4. Increased the compatibility with Perl in a number of ways:
4925
4926 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4927 to request this handling. The option can be set at compile or exec time.
4928
4929 (b) $ matches before a terminating newline by default; an option
4930 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4931 mode). The option can be set at compile or exec time.
4932
4933 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4934 the same as Perl's. If the decimal number it represents is less than 10
4935 or there aren't that many previous left capturing parentheses, an octal
4936 escape is read. Inside a character class, it's always an octal escape,
4937 even if it is a single digit.
4938
4939 (d) An escaped but undefined alphabetic character is taken as a literal,
4940 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4941 escapes.
4942
4943 (e) {0} is now permitted. (The previous item is removed from the compiled
4944 pattern).
4945
4946 5. Changed all the names of code files so that the basic parts are no longer
4947 than 10 characters, and abolished the teeny "globals.c" file.
4948
4949 6. Changed the handling of character classes; they are now done with a 32-byte
4950 bit map always.
4951
4952 7. Added the -d and /D options to pcretest to make it possible to look at the
4953 internals of compilation without having to recompile pcre.
4954
4955
4956 Version 0.95 23-Sep-97
4957 ----------------------
4958
4959 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4960 \x20 at the start of a run of normal characters. These were being treated as
4961 real characters, instead of the source characters being re-checked.
4962
4963
4964 Version 0.94 18-Sep-97
4965 ----------------------
4966
4967 1. The functions are now thread-safe, with the caveat that the global variables
4968 containing pointers to malloc() and free() or alternative functions are the
4969 same for all threads.
4970
4971 2. Get pcre_study() to generate a bitmap of initial characters for non-
4972 anchored patterns when this is possible, and use it if passed to pcre_exec().
4973
4974
4975 Version 0.93 15-Sep-97
4976 ----------------------
4977
4978 1. /(b)|(:+)/ was computing an incorrect first character.
4979
4980 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4981 but not actually doing anything yet.
4982
4983 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4984 as Perl does (e.g. [-az] or [az-]).
4985
4986 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4987 all possible positions.
4988
4989 5. Split up into different modules to avoid including unneeded functions in a
4990 compiled binary. However, compile and exec are still in one module. The "study"
4991 function is split off.
4992
4993 6. The character tables are now in a separate module whose source is generated
4994 by an auxiliary program - but can then be edited by hand if required. There are
4995 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4996 toupper() in the code.
4997
4998 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4999 make them global. Abolish the function for setting them, as the caller can now
5000 set them directly.
5001
5002
5003 Version 0.92 11-Sep-97
5004 ----------------------
5005
5006 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5007 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5008
5009 2. Caseless matching was not working in character classes if the characters in
5010 the pattern were in upper case.
5011
5012 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5013
5014 4. Make PCRE_ANCHORED public and accept as a compile option.
5015
5016 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5017 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5018 pass them.
5019
5020 6. Give an error if bad option bits passed at compile or run time.
5021
5022 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5023 pcretest to cause it to pass that flag.
5024
5025 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5026 options, and the first character, if set.
5027
5028 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5029
5030
5031 Version 0.91 10-Sep-97
5032 ----------------------
5033
5034 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5035 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5036
5037 2. PCRE was looping on encountering an indefinitely repeated back reference to
5038 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5039 Perl does - treats the match as successful.
5040
5041 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5