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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1190 - (show annotations)
Tue Oct 30 16:49:19 2012 UTC (6 years, 10 months ago) by ph10
File size: 239188 byte(s)
pcretest was not diagnosing characters > 0x7fffffff in 8-bit mode.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.32
5 ------------
6
7 1. Improved JIT compiler optimizations for first character search and single
8 character iterators.
9
10 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11 Patch by Daniel Richard G.
12
13 3. Single character iterator optimizations in the JIT compiler.
14
15 4. Improved JIT compiler optimizations for character ranges.
16
17 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18 Reported by Giuseppe D'Angelo.
19
20 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21 of a line, was being set incorrectly in cases where .* appeared inside
22 atomic brackets at the start of a pattern, or where there was a subsequent
23 *PRUNE or *SKIP.
24
25 7. Improved instruction cache flush for POWER/PowerPC.
26 Patch by Daniel Richard G.
27
28 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
29 grep:
30
31 (a) There is now no limit to the number of patterns to be matched.
32
33 (b) An error is given if a pattern is too long.
34
35 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
36 are now supported.
37
38 (d) --exclude-from and --include-from (multiple use) have been added.
39
40 (e) Exclusions and inclusions now apply to all files and directories, not
41 just to those obtained from scanning a directory recursively.
42
43 (f) Multiple uses of -f and --file-list are now supported.
44
45 (g) In a Windows environment, the default for -d has been changed from
46 "read" (the GNU grep default) to "skip", because otherwise the presence
47 of a directory in the file list provokes an error.
48
49 (h) The documentation has been revised and clarified in places.
50
51 9. Improve the matching speed of capturing brackets.
52
53 10. Changed the meaning of \X so that it now matches a Unicode extended
54 grapheme cluster.
55
56 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
57 out POSIX threads when JIT support is configured.
58
59 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
60
61 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
62 structure could go wrong in environments where size_t is not the same size
63 as int.
64
65 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
66 set.
67
68 15. The EBCDIC support had decayed; later updates to the code had included
69 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
70 general tidy up of EBCDIC-related issues, and the documentation was also
71 not quite right. There is now a test that can be run on ASCII systems to
72 check some of the EBCDIC-related things (but is it not a full test).
73
74 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
75 in a small tidy to the code.
76
77 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
78
79 18. If the --only-matching (-o) option in pcregrep is specified multiple
80 times, each one causes appropriate output. For example, -o1 -o2 outputs the
81 substrings matched by the 1st and 2nd capturing parentheses. A separating
82 string can be specified by --om-separator (default empty).
83
84 19. Improving the first n character searches.
85
86 20. Turn case lists for horizontal and vertical white space into macros so that
87 they are defined only once.
88
89 21. This set of changes together give more compatible Unicode case-folding
90 behaviour for characters that have more than one other case when UCP
91 support is available.
92
93 (a) The Unicode property table now has offsets into a new table of sets of
94 three or more characters that are case-equivalent. The MultiStage2.py
95 script that generates these tables (the pcre_ucd.c file) now scans
96 CaseFolding.txt instead of UnicodeData.txt for character case
97 information.
98
99 (b) The code for adding characters or ranges of characters to a character
100 class has been abstracted into a generalized function that also handles
101 case-independence. In UTF-mode with UCP support, this uses the new data
102 to handle characters with more than one other case.
103
104 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
105 whose other case is greater than 256 are now correctly matched
106 caselessly. Previously, the high codepoint matched the low one, but not
107 vice versa.
108
109 (d) The processing of \h, \H, \v, and \ in character classes now makes use
110 of the new class addition function, using character lists defined as
111 macros alongside the case definitions of 20 above.
112
113 (e) Caseless back references now work with characters that have more than
114 one other case.
115
116 (f) General caseless matching of characters with more than one other case
117 is supported.
118
119 22. Unicode character properties were updated from Unicode 6.2.0
120
121 23. Improved CMake support under Windows. Patch by Daniel Richard G.
122
123 24. Add support for 32-bit character strings, and UTF-32
124
125 25. Major JIT compiler update (code refactoring and bugfixing).
126 Experimental Sparc 32 support is added.
127
128 26. Applied a modified version of Daniel Richard G's patch to create
129 pcre.h.generic and config.h.generic by "make" instead of in the
130 PrepareRelease script.
131
132 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
133 pcre_compile.c when checking for a zero character.
134
135 28. Introducing a native interface for JIT. Through this interface, the compiled
136 machine code can be directly executed. The purpose of this interface is to
137 provide fast pattern matching, so several sanity checks are not performed.
138 However, feature tests are still performed. The new interface provides
139 1.4x speedup compared to the old one.
140
141 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
142 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
143 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
144
145 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
146 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
147 the "old" RFC 2279). Instead, it ended up passing a negative length to
148 pcre_exec().
149
150
151 Version 8.31 06-July-2012
152 -------------------------
153
154 1. Fixing a wrong JIT test case and some compiler warnings.
155
156 2. Removed a bashism from the RunTest script.
157
158 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
159 to unsigned type, result still unsigned" that was given by an MS compiler
160 on encountering the code "-sizeof(xxx)".
161
162 4. Partial matching support is added to the JIT compiler.
163
164 5. Fixed several bugs concerned with partial matching of items that consist
165 of more than one character:
166
167 (a) /^(..)\1/ did not partially match "aba" because checking references was
168 done on an "all or nothing" basis. This also applied to repeated
169 references.
170
171 (b) \R did not give a hard partial match if \r was found at the end of the
172 subject.
173
174 (c) \X did not give a hard partial match after matching one or more
175 characters at the end of the subject.
176
177 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
178 a partial match for the string "\r".
179
180 (e) When newline was set to CRLF, the metacharacter "." did not recognize
181 a partial match for a CR character at the end of the subject string.
182
183 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
184 running pcretest, the text "(JIT)" added to the output whenever JIT is
185 actually used to run the match.
186
187 7. Individual JIT compile options can be set in pcretest by following -s+[+]
188 or /S+[+] with a digit between 1 and 7.
189
190 8. OP_NOT now supports any UTF character not just single-byte ones.
191
192 9. (*MARK) control verb is now supported by the JIT compiler.
193
194 10. The command "./RunTest list" lists the available tests without actually
195 running any of them. (Because I keep forgetting what they all are.)
196
197 11. Add PCRE_INFO_MAXLOOKBEHIND.
198
199 12. Applied a (slightly modified) user-supplied patch that improves performance
200 when the heap is used for recursion (compiled with --disable-stack-for-
201 recursion). Instead of malloc and free for each heap frame each time a
202 logical recursion happens, frames are retained on a chain and re-used where
203 possible. This sometimes gives as much as 30% improvement.
204
205 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
206 call.
207
208 14. As documented, (*COMMIT) is now confined to within a positive assertion.
209
210 15. It is now possible to link pcretest with libedit as an alternative to
211 libreadline.
212
213 16. (*COMMIT) control verb is now supported by the JIT compiler.
214
215 17. The Unicode data tables have been updated to Unicode 6.1.0.
216
217 18. Added --file-list option to pcregrep.
218
219 19. Added binary file support to pcregrep, including the -a, --binary-files,
220 -I, and --text options.
221
222 20. The madvise function is renamed for posix_madvise for QNX compatibility
223 reasons. Fixed by Giuseppe D'Angelo.
224
225 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
226
227 22. Changed the option for creating symbolic links for 16-bit man pages from
228 -s to -sf so that re-installing does not cause issues.
229
230 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
231
232 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
233 matches in certain environments (the workspace was not being correctly
234 retained). Also added to pcre_dfa_exec() a simple plausibility check on
235 some of the workspace data at the beginning of a restart.
236
237 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
238 was not doing so when it should - probably a typo introduced by SVN 528
239 (change 8.10/14).
240
241 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
242 \w+ when the character tables indicated that \x{c4} was a word character.
243 There were several related cases, all because the tests for doing a table
244 lookup were testing for characters less than 127 instead of 255.
245
246 27. If a pattern contains capturing parentheses that are not used in a match,
247 their slots in the ovector are set to -1. For those that are higher than
248 any matched groups, this happens at the end of processing. In the case when
249 there were back references that the ovector was too small to contain
250 (causing temporary malloc'd memory to be used during matching), and the
251 highest capturing number was not used, memory off the end of the ovector
252 was incorrectly being set to -1. (It was using the size of the temporary
253 memory instead of the true size.)
254
255 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
256 ovector size, it uses memory at the end of the block that it has got.
257
258 29. Check for an overlong MARK name and give an error at compile time. The
259 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
260
261 30. JIT compiler update.
262
263 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
264 Rill for the patch.
265
266 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
267
268 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
269
270 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
271 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
272 was enabled.
273
274 35. Improve JIT code generation for greedy plus quantifier.
275
276 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
277 1 to "aa" instead of to an empty string. The bug affected repeated groups
278 that could potentially match an empty string.
279
280 37. Optimizing single character iterators in JIT.
281
282 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
283 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
284 codepoints that are too big for the mode are faulted, and in a UTF mode,
285 disallowed codepoints are also faulted.
286
287 39. If PCRE was compiled with UTF support, in three places in the DFA
288 matcher there was code that should only have been obeyed in UTF mode, but
289 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
290 processing when bytes with values greater than 127 were present. In 16-bit
291 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
292 both cases the values are those that cannot be the first data item in a UTF
293 character. The three items that might have provoked this were recursions,
294 possessively repeated groups, and atomic groups.
295
296 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
297 and pcregrep, because some OS require shared objects to be explicitly
298 passed to ld, causing the link step to fail if they are not.
299
300 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
301 mode, patterns that started with \h* or \R* might be incorrectly matched.
302
303
304 Version 8.30 04-February-2012
305 -----------------------------
306
307 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
308 name is defined in ctype.h.
309
310 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
311 only in quite long subpatterns.
312
313 3. Removed the function pcre_info(), which has been obsolete and deprecated
314 since it was replaced by pcre_fullinfo() in February 2000.
315
316 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
317 match a (*MARK), and the match failed at the start of the subject, a
318 reference to memory before the start of the subject could occur. This bug
319 was introduced by fix 17 of release 8.21.
320
321 5. A reference to an unset group with zero minimum repetition was giving
322 totally wrong answers (in non-JavaScript-compatibility mode). For example,
323 /(another)?(\1?)test/ matched against "hello world test". This bug was
324 introduced in release 8.13.
325
326 6. Add support for 16-bit character strings (a large amount of work involving
327 many changes and refactorings).
328
329 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
330 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
331 from a file.
332
333 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
334 rounding is not applied in this particular case).
335
336 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
337 if they appear, or are escaped, in patterns.
338
339 10. Get rid of a number of -Wunused-but-set-variable warnings.
340
341 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
342 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
343 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
344 also returns the mark "x". This bug applied to capturing parentheses,
345 non-capturing parentheses, and atomic parentheses. It also applied to some
346 assertions.
347
348 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
349 information out of configure.ac instead of relying on pcre.h.generic, which
350 is not stored in the repository.
351
352 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
353 -lreadline.
354
355 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
356
357 15. Applied Graycode's patch to put the top-level frame on the stack rather
358 than the heap when not using the stack for recursion. This gives a
359 performance improvement in many cases when recursion is not deep.
360
361 16. Experimental code added to "pcretest -C" to output the stack frame size.
362
363
364 Version 8.21 12-Dec-2011
365 ------------------------
366
367 1. Updating the JIT compiler.
368
369 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
370 are added as well.
371
372 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
373 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
374 calling _pcre_jit_exec. Some extra comments are added.
375
376 4. (*MARK) settings inside atomic groups that do not contain any capturing
377 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
378 was introduced by change 18 for 8.20.
379
380 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
381 ECMA-262 standard.
382
383 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
384 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
385 This bug was probably introduced by change 9 of 8.13.
386
387 7. While fixing 6 above, I noticed that a number of other items were being
388 incorrectly rejected as "not fixed length". This arose partly because newer
389 opcodes had not been added to the fixed-length checking code. I have (a)
390 corrected the bug and added tests for these items, and (b) arranged for an
391 error to occur if an unknown opcode is encountered while checking for fixed
392 length instead of just assuming "not fixed length". The items that were
393 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
394 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
395 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
396
397 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
398 being incorrectly compiled and would have given unpredicatble results.
399
400 9. A possessively repeated subpattern with minimum repeat count greater than
401 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
402 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
403 the first (A) could occur when it should not.
404
405 10. Add a cast and remove a redundant test from the code.
406
407 11. JIT should use pcre_malloc/pcre_free for allocation.
408
409 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
410 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
411 is anything other than /usr, -L is still shown).
412
413 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
414
415 14. Perl does not support \N without a following name in a [] class; PCRE now
416 also gives an error.
417
418 15. If a forward reference was repeated with an upper limit of around 2000,
419 it caused the error "internal error: overran compiling workspace". The
420 maximum number of forward references (including repeats) was limited by the
421 internal workspace, and dependent on the LINK_SIZE. The code has been
422 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
423 the default depends on LINK_SIZE. There is a new upper limit (for safety)
424 of around 200,000 forward references. While doing this, I also speeded up
425 the filling in of repeated forward references.
426
427 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
428 incorrectly expecting the subject to contain another "a" after the start.
429
430 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
431 in the match, the SKIP should be ignored. This was not happening; instead
432 the SKIP was being treated as NOMATCH. For patterns such as
433 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
434 tested.
435
436 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
437 now much more compatible with Perl, in particular in cases where the result
438 is a non-match for a non-anchored pattern. For example, if
439 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
440 "m", where previously it did not return a name. A side effect of this
441 change is that for partial matches, the last encountered mark name is
442 returned, as for non matches. A number of tests that were previously not
443 Perl-compatible have been moved into the Perl-compatible test files. The
444 refactoring has had the pleasing side effect of removing one argument from
445 the match() function, thus reducing its stack requirements.
446
447 19. If the /S+ option was used in pcretest to study a pattern using JIT,
448 subsequent uses of /S (without +) incorrectly behaved like /S+.
449
450 21. Retrieve executable code size support for the JIT compiler and fixing
451 some warnings.
452
453 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
454 not work when the shorter character appeared right at the end of the
455 subject string.
456
457 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
458 systems.
459
460 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
461 output it when the /M option is used in pcretest.
462
463 25. The CheckMan script was not being included in the distribution. Also, added
464 an explicit "perl" to run Perl scripts from the PrepareRelease script
465 because this is reportedly needed in Windows.
466
467 26. If study data was being save in a file and studying had not found a set of
468 "starts with" bytes for the pattern, the data written to the file (though
469 never used) was taken from uninitialized memory and so caused valgrind to
470 complain.
471
472 27. Updated RunTest.bat as provided by Sheri Pierce.
473
474 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
475
476 29. Computation of memory usage for the table of capturing group names was
477 giving an unnecessarily large value.
478
479
480 Version 8.20 21-Oct-2011
481 ------------------------
482
483 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
484 a POSIX class. After further experiments with Perl, which convinced me that
485 Perl has bugs and confusions, a closing square bracket is no longer allowed
486 in a POSIX name. This bug also affected patterns with classes that started
487 with full stops.
488
489 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
490 captured substring, but while checking the failing first alternative,
491 substring 1 is temporarily captured. If the output vector supplied to
492 pcre_exec() was not big enough for this capture, the yield of the function
493 was still zero ("insufficient space for captured substrings"). This cannot
494 be totally fixed without adding another stack variable, which seems a lot
495 of expense for a edge case. However, I have improved the situation in cases
496 such as /(a)(b)x|abc/ matched against "abc", where the return code
497 indicates that fewer than the maximum number of slots in the ovector have
498 been set.
499
500 3. Related to (2) above: when there are more back references in a pattern than
501 slots in the output vector, pcre_exec() uses temporary memory during
502 matching, and copies in the captures as far as possible afterwards. It was
503 using the entire output vector, but this conflicts with the specification
504 that only 2/3 is used for passing back captured substrings. Now it uses
505 only the first 2/3, for compatibility. This is, of course, another edge
506 case.
507
508 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
509 main code base, and can be used by building with --enable-jit. When this is
510 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
511 runtime --no-jit option is given.
512
513 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
514 ovector, the return from the function was zero, implying that there were
515 other matches that did not fit. The correct "exactly full" value is now
516 returned.
517
518 6. If a subpattern that was called recursively or as a subroutine contained
519 (*PRUNE) or any other control that caused it to give a non-standard return,
520 invalid errors such as "Error -26 (nested recursion at the same subject
521 position)" or even infinite loops could occur.
522
523 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
524 computing the minimum length on reaching *ACCEPT, and so ended up with the
525 wrong value of 1 rather than 0. Further investigation indicates that
526 computing a minimum subject length in the presence of *ACCEPT is difficult
527 (think back references, subroutine calls), and so I have changed the code
528 so that no minimum is registered for a pattern that contains *ACCEPT.
529
530 8. If (*THEN) was present in the first (true) branch of a conditional group,
531 it was not handled as intended. [But see 16 below.]
532
533 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
534 Sheri Pierce.
535
536 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
537 the first byte in a match must be "a".
538
539 11. Change 17 for 8.13 increased the recursion depth for patterns like
540 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
541 pattern contains any instances of (*THEN). If it does not, the old
542 optimizations are restored. It would be nice to do this on a per-group
543 basis, but at the moment that is not feasible.
544
545 12. In some environments, the output of pcretest -C is CRLF terminated. This
546 broke RunTest's code that checks for the link size. A single white space
547 character after the value is now allowed for.
548
549 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
550 For "fr", it uses the Windows-specific input and output files.
551
552 14. If (*THEN) appeared in a group that was called recursively or as a
553 subroutine, it did not work as intended. [But see next item.]
554
555 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
556 pattern fragments (but not containing any | characters). If A and B are
557 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
558 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
559 D. In other words, Perl considers parentheses that do not contain any |
560 characters to be part of a surrounding alternative, whereas PCRE was
561 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
562 differently. PCRE now behaves in the same way as Perl, except in the case
563 of subroutine/recursion calls such as (?1) which have in any case always
564 been different (but PCRE had them first :-).
565
566 16. Related to 15 above: Perl does not treat the | in a conditional group as
567 creating alternatives. Such a group is treated in the same way as an
568 ordinary group without any | characters when processing (*THEN). PCRE has
569 been changed to match Perl's behaviour.
570
571 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
572 RunGrepTest script failed.
573
574 18. Change 22 for version 13 caused atomic groups to use more stack. This is
575 inevitable for groups that contain captures, but it can lead to a lot of
576 stack use in large patterns. The old behaviour has been restored for atomic
577 groups that do not contain any capturing parentheses.
578
579 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
580 suppress the check for a minimum subject length at run time. (If it was
581 given to pcre_exec() or pcre_dfa_exec() it did work.)
582
583 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
584 fail to work when decoding hex characters in data strings in EBCDIC
585 environments.
586
587 21. It appears that in at least one Mac OS environment, the isxdigit() function
588 is implemented as a macro that evaluates to its argument more than once,
589 contravening the C 90 Standard (I haven't checked a later standard). There
590 was an instance in pcretest which caused it to go wrong when processing
591 \x{...} escapes in subject strings. The has been rewritten to avoid using
592 things like p++ in the argument of isxdigit().
593
594
595 Version 8.13 16-Aug-2011
596 ------------------------
597
598 1. The Unicode data tables have been updated to Unicode 6.0.0.
599
600 2. Two minor typos in pcre_internal.h have been fixed.
601
602 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
603 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
604 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
605
606 4. There were a number of related bugs in the code for matching backrefences
607 caselessly in UTF-8 mode when codes for the characters concerned were
608 different numbers of bytes. For example, U+023A and U+2C65 are an upper
609 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
610 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
611 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
612 2-byte code at the end of the subject (it thought there wasn't enough data
613 left).
614
615 5. Comprehensive information about what went wrong is now returned by
616 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
617 as the output vector has at least 2 elements. The offset of the start of
618 the failing character and a reason code are placed in the vector.
619
620 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
621 now returned is for the first byte of the failing character, instead of the
622 last byte inspected. This is an incompatible change, but I hope it is small
623 enough not to be a problem. It makes the returned offset consistent with
624 pcre_exec() and pcre_dfa_exec().
625
626 7. pcretest now gives a text phrase as well as the error number when
627 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
628 failure, the offset and reason code are output.
629
630 8. When \R was used with a maximizing quantifier it failed to skip backwards
631 over a \r\n pair if the subsequent match failed. Instead, it just skipped
632 back over a single character (\n). This seems wrong (because it treated the
633 two characters as a single entity when going forwards), conflicts with the
634 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
635 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
636 has been changed.
637
638 9. Some internal refactoring has changed the processing so that the handling
639 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
640 time (the PCRE_DOTALL option was changed this way some time ago: version
641 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
642 which was always a bit of a fudge. It also means that there is one less
643 argument for the match() function, which reduces its stack requirements
644 slightly. This change also fixes an incompatibility with Perl: the pattern
645 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
646
647 10. More internal refactoring has drastically reduced the number of recursive
648 calls to match() for possessively repeated groups such as (abc)++ when
649 using pcre_exec().
650
651 11. While implementing 10, a number of bugs in the handling of groups were
652 discovered and fixed:
653
654 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
655 (a|)*(?1) gave a compile-time internal error.
656 ((a|)+)+ did not notice that the outer group could match an empty string.
657 (^a|^)+ was not marked as anchored.
658 (.*a|.*)+ was not marked as matching at start or after a newline.
659
660 12. Yet more internal refactoring has removed another argument from the match()
661 function. Special calls to this function are now indicated by setting a
662 value in a variable in the "match data" data block.
663
664 13. Be more explicit in pcre_study() instead of relying on "default" for
665 opcodes that mean there is no starting character; this means that when new
666 ones are added and accidentally left out of pcre_study(), testing should
667 pick them up.
668
669 14. The -s option of pcretest has been documented for ages as being an old
670 synonym of -m (show memory usage). I have changed it to mean "force study
671 for every regex", that is, assume /S for every regex. This is similar to -i
672 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
673 using it. It makes it easier to run collections of tests with and without
674 study enabled, and thereby test pcre_study() more easily. All the standard
675 tests are now run with and without -s (but some patterns can be marked as
676 "never study" - see 20 below).
677
678 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
679 restoration of the capturing data to the outer values was not happening
680 correctly.
681
682 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
683 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
684 pattern had matched an empty string, and so incorrectly returned a no
685 match.
686
687 17. There was optimizing code for the last branch of non-capturing parentheses,
688 and also for the obeyed branch of a conditional subexpression, which used
689 tail recursion to cut down on stack usage. Unfortunately, now that there is
690 the possibility of (*THEN) occurring in these branches, tail recursion is
691 no longer possible because the return has to be checked for (*THEN). These
692 two optimizations have therefore been removed. [But see 8.20/11 above.]
693
694 18. If a pattern containing \R was studied, it was assumed that \R always
695 matched two bytes, thus causing the minimum subject length to be
696 incorrectly computed because \R can also match just one byte.
697
698 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
699 was incorrectly computed.
700
701 20. If /S is present twice on a test pattern in pcretest input, it now
702 *disables* studying, thereby overriding the use of -s on the command line
703 (see 14 above). This is necessary for one or two tests to keep the output
704 identical in both cases.
705
706 21. When (*ACCEPT) was used in an assertion that matched an empty string and
707 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
708
709 22. When an atomic group that contained a capturing parenthesis was
710 successfully matched, but the branch in which it appeared failed, the
711 capturing was not being forgotten if a higher numbered group was later
712 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
713 group 1 to "a", when in fact it should be unset. This applied to multi-
714 branched capturing and non-capturing groups, repeated or not, and also to
715 positive assertions (capturing in negative assertions does not happen
716 in PCRE) and also to nested atomic groups.
717
718 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
719 subject after a captured substring, to make it easier to tell which of a
720 number of identical substrings has been captured.
721
722 24. The way atomic groups are processed by pcre_exec() has been changed so that
723 if they are repeated, backtracking one repetition now resets captured
724 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
725 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
726 "aaa". Previously, it would have been "a". As part of this code
727 refactoring, the way recursive calls are handled has also been changed.
728
729 25. If an assertion condition captured any substrings, they were not passed
730 back unless some other capturing happened later. For example, if
731 (?(?=(a))a) was matched against "a", no capturing was returned.
732
733 26. When studying a pattern that contained subroutine calls or assertions,
734 the code for finding the minimum length of a possible match was handling
735 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
736 group 1 called group 2 while simultaneously a separate group 2 called group
737 1). A stack overflow occurred in this case. I have fixed this by limiting
738 the recursion depth to 10.
739
740 27. Updated RunTest.bat in the distribution to the version supplied by Tom
741 Fortmann. This supports explicit test numbers on the command line, and has
742 argument validation and error reporting.
743
744 28. An instance of \X with an unlimited repeat could fail if at any point the
745 first character it looked at was a mark character.
746
747 29. Some minor code refactoring concerning Unicode properties and scripts
748 should reduce the stack requirement of match() slightly.
749
750 30. Added the '=' option to pcretest to check the setting of unused capturing
751 slots at the end of the pattern, which are documented as being -1, but are
752 not included in the return count.
753
754 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
755 compiled something random. Now it gives a compile-time error (as does
756 Perl).
757
758 32. A *MARK encountered during the processing of a positive assertion is now
759 recorded and passed back (compatible with Perl).
760
761 33. If --only-matching or --colour was set on a pcregrep call whose pattern
762 had alternative anchored branches, the search for a second match in a line
763 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
764 matched the line "0102" twice. The same bug affected patterns that started
765 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
766 twice.
767
768 34. Previously, PCRE did not allow quantification of assertions. However, Perl
769 does, and because of capturing effects, quantifying parenthesized
770 assertions may at times be useful. Quantifiers are now allowed for
771 parenthesized assertions.
772
773 35. A minor code tidy in pcre_compile() when checking options for \R usage.
774
775 36. \g was being checked for fancy things in a character class, when it should
776 just be a literal "g".
777
778 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
779 appearance of a nested POSIX class supersedes an apparent external class.
780 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
781 unescaped square brackets may also appear as part of class names. For
782 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
783 more like Perl. (But see 8.20/1 above.)
784
785 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
786 was because it thought it was \N{name}, which is not supported).
787
788 39. Add minix to OS list not supporting the -S option in pcretest.
789
790 40. PCRE tries to detect cases of infinite recursion at compile time, but it
791 cannot analyze patterns in sufficient detail to catch mutual recursions
792 such as ((?1))((?2)). There is now a runtime test that gives an error if a
793 subgroup is called recursively as a subpattern for a second time at the
794 same position in the subject string. In previous releases this might have
795 been caught by the recursion limit, or it might have run out of stack.
796
797 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
798 happen only once. PCRE was, however incorrectly giving a compile time error
799 "recursive call could loop indefinitely" because it cannot analyze the
800 pattern in sufficient detail. The compile time test no longer happens when
801 PCRE is compiling a conditional subpattern, but actual runaway loops are
802 now caught at runtime (see 40 above).
803
804 42. It seems that Perl allows any characters other than a closing parenthesis
805 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
806 has been changed to be the same.
807
808 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
809 as not to get warnings when autogen.sh is called. Also changed
810 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
811
812 44. To help people who use pcregrep to scan files containing exceedingly long
813 lines, the following changes have been made:
814
815 (a) The default value of the buffer size parameter has been increased from
816 8K to 20K. (The actual buffer used is three times this size.)
817
818 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
819 PCRE is built.
820
821 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
822 to be set at run time.
823
824 (d) Numerical values in pcregrep options can be followed by K or M, for
825 example --buffer-size=50K.
826
827 (e) If a line being scanned overflows pcregrep's buffer, an error is now
828 given and the return code is set to 2.
829
830 45. Add a pointer to the latest mark to the callout data block.
831
832 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
833 partial match of an empty string instead of no match. This was specific to
834 the use of ".".
835
836 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
837 complete match instead of a partial match. This bug was dependent on both
838 the PCRE_UTF8 and PCRE_DOTALL options being set.
839
840 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
841 starting byte set, because \b was not being ignored.
842
843
844 Version 8.12 15-Jan-2011
845 ------------------------
846
847 1. Fixed some typos in the markup of the man pages, and wrote a script that
848 checks for such things as part of the documentation building process.
849
850 2. On a big-endian 64-bit system, pcregrep did not correctly process the
851 --match-limit and --recursion-limit options (added for 8.11). In
852 particular, this made one of the standard tests fail. (The integer value
853 went into the wrong half of a long int.)
854
855 3. If the --colour option was given to pcregrep with -v (invert match), it
856 did strange things, either producing crazy output, or crashing. It should,
857 of course, ignore a request for colour when reporting lines that do not
858 match.
859
860 4. Another pcregrep bug caused similar problems if --colour was specified with
861 -M (multiline) and the pattern match finished with a line ending.
862
863 5. In pcregrep, when a pattern that ended with a literal newline sequence was
864 matched in multiline mode, the following line was shown as part of the
865 match. This seems wrong, so I have changed it.
866
867 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
868 the check for further matches in the same line (so they could be coloured)
869 to overrun the end of the current line. If another match was found, it was
870 incorrectly shown (and then shown again when found in the next line).
871
872 7. If pcregrep was compiled under Windows, there was a reference to the
873 function pcregrep_exit() before it was defined. I am assuming this was
874 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
875 reported by a user. I've moved the definition above the reference.
876
877
878 Version 8.11 10-Dec-2010
879 ------------------------
880
881 1. (*THEN) was not working properly if there were untried alternatives prior
882 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
883 backtracked to try for "b" instead of moving to the next alternative branch
884 at the same level (in this case, to look for "c"). The Perl documentation
885 is clear that when (*THEN) is backtracked onto, it goes to the "next
886 alternative in the innermost enclosing group".
887
888 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
889 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
890 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
891 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
892 (*THEN).
893
894 3. If \s appeared in a character class, it removed the VT character from
895 the class, even if it had been included by some previous item, for example
896 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
897 of \s, but is part of the POSIX "space" class.)
898
899 4. A partial match never returns an empty string (because you can always
900 match an empty string at the end of the subject); however the checking for
901 an empty string was starting at the "start of match" point. This has been
902 changed to the "earliest inspected character" point, because the returned
903 data for a partial match starts at this character. This means that, for
904 example, /(?<=abc)def/ gives a partial match for the subject "abc"
905 (previously it gave "no match").
906
907 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
908 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
909 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
910 has an implication that the given string is incomplete (because a partial
911 match is preferred over a full match). For this reason, these items now
912 give a partial match in this situation. [Aside: previously, the one case
913 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
914 match rather than a full match, which was wrong by the old rules, but is
915 now correct.]
916
917 6. There was a bug in the handling of #-introduced comments, recognized when
918 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
919 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
920 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
921 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
922 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
923 places in pcre_compile().
924
925 7. Related to (6) above, when pcre_compile() was skipping #-introduced
926 comments when looking ahead for named forward references to subpatterns,
927 the only newline sequence it recognized was NL. It now handles newlines
928 according to the set newline convention.
929
930 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
931 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
932 cater for a lack of strerror(). These oversights have been fixed.
933
934 9. Added --match-limit and --recursion-limit to pcregrep.
935
936 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
937
938 11. When the -o option was used, pcregrep was setting a return code of 1, even
939 when matches were found, and --line-buffered was not being honoured.
940
941 12. Added an optional parentheses number to the -o and --only-matching options
942 of pcregrep.
943
944 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
945 can match an empty string. The code to do it in pcretest and pcredemo
946 needed fixing:
947
948 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
949 only one byte after an empty string match just before CRLF (this case
950 just got forgotten; "any" and "anycrlf" were OK).
951
952 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
953 mode when an empty string match preceded an ASCII character followed by
954 a non-ASCII character. (The code for advancing by one character rather
955 than one byte was nonsense.)
956
957 (c) The pcredemo.c sample program did not have any code at all to handle
958 the cases when CRLF is a valid newline sequence.
959
960 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
961 as a starting offset was within the subject string. There is now a new
962 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
963 negative or greater than the length of the string. In order to test this,
964 pcretest is extended to allow the setting of negative starting offsets.
965
966 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
967 starting offset points to the beginning of a UTF-8 character was
968 unnecessarily clumsy. I tidied it up.
969
970 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
971 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
972
973 17. Nobody had reported that the --include_dir option, which was added in
974 release 7.7 should have been called --include-dir (hyphen, not underscore)
975 for compatibility with GNU grep. I have changed it to --include-dir, but
976 left --include_dir as an undocumented synonym, and the same for
977 --exclude-dir, though that is not available in GNU grep, at least as of
978 release 2.5.4.
979
980 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
981 characters from a string of bytes) have been redefined so as not to use
982 loops, in order to improve performance in some environments. At the same
983 time, I abstracted some of the common code into auxiliary macros to save
984 repetition (this should not affect the compiled code).
985
986 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
987 compile-time error is now given if \c is not followed by an ASCII
988 character, that is, a byte less than 128. (In EBCDIC mode, the code is
989 different, and any byte value is allowed.)
990
991 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
992 START_OPTIMIZE option, which is now allowed at compile time - but just
993 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
994 to pcregrep and other applications that have no direct access to PCRE
995 options. The new /Y option in pcretest sets this option when calling
996 pcre_compile().
997
998 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
999 back references. Groups containing recursive back references were forced to
1000 be atomic by that change, but in the case of named groups, the amount of
1001 memory required was incorrectly computed, leading to "Failed: internal
1002 error: code overflow". This has been fixed.
1003
1004 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1005 pcretest.c, to avoid build problems in some Borland environments.
1006
1007
1008 Version 8.10 25-Jun-2010
1009 ------------------------
1010
1011 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1012 THEN.
1013
1014 2. (*ACCEPT) was not working when inside an atomic group.
1015
1016 3. Inside a character class, \B is treated as a literal by default, but
1017 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1018 causes the error). The code is unchanged, but I tidied the documentation.
1019
1020 4. Inside a character class, PCRE always treated \R and \X as literals,
1021 whereas Perl faults them if its -w option is set. I have changed PCRE so
1022 that it faults them when PCRE_EXTRA is set.
1023
1024 5. Added support for \N, which always matches any character other than
1025 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1026
1027 6. When compiling pcregrep with newer versions of gcc which may have
1028 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1029 declared with attribute warn_unused_result" were given. Just casting the
1030 result to (void) does not stop the warnings; a more elaborate fudge is
1031 needed. I've used a macro to implement this.
1032
1033 7. Minor change to pcretest.c to avoid a compiler warning.
1034
1035 8. Added four artifical Unicode properties to help with an option to make
1036 \s etc use properties (see next item). The new properties are: Xan
1037 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1038
1039 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1040 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1041 this option. Modified pcretest to add /W to test this facility. Added
1042 REG_UCP to make it available via the POSIX interface.
1043
1044 10. Added --line-buffered to pcregrep.
1045
1046 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1047 studied, and the match started with a letter with a code point greater than
1048 127 whose first byte was different to the first byte of the other case of
1049 the letter, the other case of this starting letter was not recognized
1050 (#976).
1051
1052 12. If a pattern that was studied started with a repeated Unicode property
1053 test, for example, \p{Nd}+, there was the theoretical possibility of
1054 setting up an incorrect bitmap of starting bytes, but fortunately it could
1055 not have actually happened in practice until change 8 above was made (it
1056 added property types that matched character-matching opcodes).
1057
1058 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1059 possible starting bytes for non-anchored patterns.
1060
1061 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1062 \R, and also a number of cases that involve Unicode properties, both
1063 explicit and implicit when PCRE_UCP is set.
1064
1065 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1066 input, it could crash or give wrong results if characters with values
1067 greater than 0xc0 were present in the subject string. (Detail: it assumed
1068 UTF-8 input when processing these items.)
1069
1070 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1071 size_t is 64-bit (#991).
1072
1073 17. Added a check for running out of memory when PCRE is compiled with
1074 --disable-stack-for-recursion (#990).
1075
1076 18. If the last data line in a file for pcretest does not have a newline on
1077 the end, a newline was missing in the output.
1078
1079 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1080 less than 128) in its various bitmaps. However, there is a facility for
1081 generating tables according to the current locale when PCRE is compiled. It
1082 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1083 characters, are recognized by isspace() and therefore were getting set in
1084 these tables, and indeed these tables seem to approximate to ISO 8859. This
1085 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1086 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1087 which of course cannot start UTF-8 characters. I have changed the code so
1088 that only real ASCII characters (less than 128) and the correct starting
1089 bytes for UTF-8 encodings are set for characters greater than 127 when in
1090 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1091 altogether.)
1092
1093 20. Added the /T option to pcretest so as to be able to run tests with non-
1094 standard character tables, thus making it possible to include the tests
1095 used for 19 above in the standard set of tests.
1096
1097 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1098 reference to a subpattern the other side of a comment that contains an
1099 opening parenthesis caused either an internal compiling error, or a
1100 reference to the wrong subpattern.
1101
1102
1103 Version 8.02 19-Mar-2010
1104 ------------------------
1105
1106 1. The Unicode data tables have been updated to Unicode 5.2.0.
1107
1108 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1109 configured.
1110
1111 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1112 original author of that file, following a query about its status.
1113
1114 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1115 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1116
1117 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1118 quantifier applied to a forward-referencing subroutine call, could compile
1119 incorrect code or give the error "internal error: previously-checked
1120 referenced subpattern not found".
1121
1122 6. Both MS Visual Studio and Symbian OS have problems with initializing
1123 variables to point to external functions. For these systems, therefore,
1124 pcre_malloc etc. are now initialized to local functions that call the
1125 relevant global functions.
1126
1127 7. There were two entries missing in the vectors called coptable and poptable
1128 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1129 I've fixed the data, and added a kludgy way of testing at compile time that
1130 the lengths are correct (equal to the number of opcodes).
1131
1132 8. Following on from 7, I added a similar kludge to check the length of the
1133 eint vector in pcreposix.c.
1134
1135 9. Error texts for pcre_compile() are held as one long string to avoid too
1136 much relocation at load time. To find a text, the string is searched,
1137 counting zeros. There was no check for running off the end of the string,
1138 which could happen if a new error number was added without updating the
1139 string.
1140
1141 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1142
1143 11. \K was not working if it appeared in an atomic group or in a group that
1144 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1145 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1146 the assertion is positive, but not if it is negative.
1147
1148 12. Change 11 fortuitously reduced the size of the stack frame used in the
1149 "match()" function of pcre_exec.c by one pointer. Forthcoming
1150 implementation of support for (*MARK) will need an extra pointer on the
1151 stack; I have reserved it now, so that the stack frame size does not
1152 decrease.
1153
1154 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1155 item in branch that calls a recursion is a subroutine call - as in the
1156 second branch in the above example - was incorrectly given the compile-
1157 time error "recursive call could loop indefinitely" because pcre_compile()
1158 was not correctly checking the subroutine for matching a non-empty string.
1159
1160 14. The checks for overrunning compiling workspace could trigger after an
1161 overrun had occurred. This is a "should never occur" error, but it can be
1162 triggered by pathological patterns such as hundreds of nested parentheses.
1163 The checks now trigger 100 bytes before the end of the workspace.
1164
1165 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1166
1167
1168 Version 8.01 19-Jan-2010
1169 ------------------------
1170
1171 1. If a pattern contained a conditional subpattern with only one branch (in
1172 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1173 computed the wrong minimum data length (which is of course zero for such
1174 subpatterns). This could cause incorrect "no match" results.
1175
1176 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1177 the pattern is reset in the first branch, pcre_compile() failed with
1178 "internal error: code overflow at offset...". This happened only when
1179 the reset was to the original external option setting. (An optimization
1180 abstracts leading options settings into an external setting, which was the
1181 cause of this.)
1182
1183 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1184 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1185 assertion pattern did not match (meaning that the assertion was true), it
1186 was incorrectly treated as false if the SKIP had been reached during the
1187 matching. This also applied to assertions used as conditions.
1188
1189 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1190 assertion subpattern, including such a pattern used as a condition,
1191 unpredictable results occurred, instead of the error return
1192 PCRE_ERROR_DFA_UITEM.
1193
1194 5. The C++ GlobalReplace function was not working like Perl for the special
1195 situation when an empty string is matched. It now does the fancy magic
1196 stuff that is necessary.
1197
1198 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1199 removed. (These were left over from very, very early versions of PCRE.)
1200
1201 7. Some cosmetic changes to the code to make life easier when compiling it
1202 as part of something else:
1203
1204 (a) Change DEBUG to PCRE_DEBUG.
1205
1206 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1207 called "current" as "current_branch", to prevent a collision with the
1208 Linux macro when compiled as a kernel module.
1209
1210 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1211 prevent a collision with the Linux macro when compiled as a kernel
1212 module.
1213
1214 8. In pcre_compile() there are some checks for integer overflows that used to
1215 cast potentially large values to (double). This has been changed to that
1216 when building, a check for int64_t is made, and if it is found, it is used
1217 instead, thus avoiding the use of floating point arithmetic. (There is no
1218 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1219 double.
1220
1221 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1222 2005 (difference between two addresses compared to an unsigned value).
1223
1224 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1225 custom one, because of the following reported problem in Windows:
1226
1227 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1228 under Win32.
1229 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1230 therefore missing the function definition.
1231 - The compiler thus generates a "C" signature for the test function.
1232 - The linker fails to find the "C" function.
1233 - PCRE fails to configure if asked to do so against libbz2.
1234
1235 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1236 messages were output:
1237
1238 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1239 rerunning libtoolize, to keep the correct libtool macros in-tree.
1240 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1241
1242 I have done both of these things.
1243
1244 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1245 most of the time, it *can* run out if it is given a pattern that contains a
1246 runaway infinite recursion. I updated the discussion in the pcrestack man
1247 page.
1248
1249 13. Now that we have gone to the x.xx style of version numbers, the minor
1250 version may start with zero. Using 08 or 09 is a bad idea because users
1251 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1252 interpreted as invalid octal numbers. I've updated the previous comment in
1253 configure.ac, and also added a check that gives an error if 08 or 09 are
1254 used.
1255
1256 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1257 causing partial matching to fail when the end of the subject matched \W
1258 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1259
1260 15. There were some discrepancies between the declarations in pcre_internal.h
1261 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1262 their definitions. The declarations used "const uschar *" and the
1263 definitions used USPTR. Even though USPTR is normally defined as "const
1264 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1265 reported that: "This difference in casting confuses some C++ compilers, for
1266 example, SunCC recognizes above declarations as different functions and
1267 generates broken code for hbpcre." I have changed the declarations to use
1268 USPTR.
1269
1270 16. GNU libtool is named differently on some systems. The autogen.sh script now
1271 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1272 (FreeBSD).
1273
1274 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1275 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1276 comment: "Figure out how to create a longlong from a string: strtoll and
1277 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1278 instance, but it only takes 2 args instead of 3!"
1279
1280 18. A subtle bug concerned with back references has been fixed by a change of
1281 specification, with a corresponding code fix. A pattern such as
1282 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1283 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1284 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1285 same bug. Such groups have to be quantified to be useful, or contained
1286 inside another quantified group. (If there's no repetition, the reference
1287 can never match.) The problem arises because, having left the group and
1288 moved on to the rest of the pattern, a later failure that backtracks into
1289 the group uses the captured value from the final iteration of the group
1290 rather than the correct earlier one. I have fixed this in PCRE by forcing
1291 any group that contains a reference to itself to be an atomic group; that
1292 is, there cannot be any backtracking into it once it has completed. This is
1293 similar to recursive and subroutine calls.
1294
1295
1296 Version 8.00 19-Oct-09
1297 ----------------------
1298
1299 1. The table for translating pcre_compile() error codes into POSIX error codes
1300 was out-of-date, and there was no check on the pcre_compile() error code
1301 being within the table. This could lead to an OK return being given in
1302 error.
1303
1304 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1305 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1306 in a Windows environment.
1307
1308 3. The pcregrep --count option prints the count for each file even when it is
1309 zero, as does GNU grep. However, pcregrep was also printing all files when
1310 --files-with-matches was added. Now, when both options are given, it prints
1311 counts only for those files that have at least one match. (GNU grep just
1312 prints the file name in this circumstance, but including the count seems
1313 more useful - otherwise, why use --count?) Also ensured that the
1314 combination -clh just lists non-zero counts, with no names.
1315
1316 4. The long form of the pcregrep -F option was incorrectly implemented as
1317 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1318 but it seems right to fix it, and I didn't think it was worth preserving
1319 the old behaviour.
1320
1321 5. The command line items --regex=pattern and --regexp=pattern were not
1322 recognized by pcregrep, which required --regex pattern or --regexp pattern
1323 (with a space rather than an '='). The man page documented the '=' forms,
1324 which are compatible with GNU grep; these now work.
1325
1326 6. No libpcreposix.pc file was created for pkg-config; there was just
1327 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1328
1329 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1330 when UCP support is not needed, by modifying the Python script that
1331 generates it from Unicode data files. This should not matter if the module
1332 is correctly used as a library, but I received one complaint about 50K of
1333 unwanted data. My guess is that the person linked everything into his
1334 program rather than using a library. Anyway, it does no harm.
1335
1336 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1337 was a minimum greater than 1 for a wide character in a possessive
1338 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1339 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1340 character. Chaos in the form of incorrect output or a compiling loop could
1341 result.
1342
1343 9. The restrictions on what a pattern can contain when partial matching is
1344 requested for pcre_exec() have been removed. All patterns can now be
1345 partially matched by this function. In addition, if there are at least two
1346 slots in the offset vector, the offset of the earliest inspected character
1347 for the match and the offset of the end of the subject are set in them when
1348 PCRE_ERROR_PARTIAL is returned.
1349
1350 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1351 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1352 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1353 and may be more useful for multi-segment matching.
1354
1355 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1356 used to be given if ever the end of the subject was reached; now it is
1357 given only if matching could not proceed because another character was
1358 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1359 string "Z", which now yields "no match" instead of "partial match". In the
1360 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1361 final character ended with (*FAIL).
1362
1363 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1364 if the pattern had a "must contain" character that was already found in the
1365 earlier partial match, unless partial matching was again requested. For
1366 example, with the pattern /dog.(body)?/, the "must contain" character is
1367 "g". If the first part-match was for the string "dog", restarting with
1368 "sbody" failed. This bug has been fixed.
1369
1370 13. The string returned by pcre_dfa_exec() after a partial match has been
1371 changed so that it starts at the first inspected character rather than the
1372 first character of the match. This makes a difference only if the pattern
1373 starts with a lookbehind assertion or \b or \B (\K is not supported by
1374 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1375 matching functions compatible, and I think it's the right thing to do.
1376
1377 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1378 so that the demonstration program is easily available in environments where
1379 PCRE has not been installed from source.
1380
1381 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1382 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1383 library.
1384
1385 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1386 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1387 is not the first non-POSIX option to be added. Clearly some people find
1388 these options useful.
1389
1390 17. If a caller to the POSIX matching function regexec() passes a non-zero
1391 value for nmatch with a NULL value for pmatch, the value of
1392 nmatch is forced to zero.
1393
1394 18. RunGrepTest did not have a test for the availability of the -u option of
1395 the diff command, as RunTest does. It now checks in the same way as
1396 RunTest, and also checks for the -b option.
1397
1398 19. If an odd number of negated classes containing just a single character
1399 interposed, within parentheses, between a forward reference to a named
1400 subpattern and the definition of the subpattern, compilation crashed with
1401 an internal error, complaining that it could not find the referenced
1402 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1403 [The bug was that it was starting one character too far in when skipping
1404 over the character class, thus treating the ] as data rather than
1405 terminating the class. This meant it could skip too much.]
1406
1407 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1408 /g option in pcretest when the pattern contains \K, which makes it possible
1409 to have an empty string match not at the start, even when the pattern is
1410 anchored. Updated pcretest and pcredemo to use this option.
1411
1412 21. If the maximum number of capturing subpatterns in a recursion was greater
1413 than the maximum at the outer level, the higher number was returned, but
1414 with unset values at the outer level. The correct (outer level) value is
1415 now given.
1416
1417 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1418 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1419 make it do so. The string so far is captured, making this feature
1420 compatible with Perl.
1421
1422 23. The tests have been re-organized, adding tests 11 and 12, to make it
1423 possible to check the Perl 5.10 features against Perl 5.10.
1424
1425 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1426 pattern matches a fixed length string. PCRE did not allow this; now it
1427 does. Neither allows recursion.
1428
1429 25. I finally figured out how to implement a request to provide the minimum
1430 length of subject string that was needed in order to match a given pattern.
1431 (It was back references and recursion that I had previously got hung up
1432 on.) This code has now been added to pcre_study(); it finds a lower bound
1433 to the length of subject needed. It is not necessarily the greatest lower
1434 bound, but using it to avoid searching strings that are too short does give
1435 some useful speed-ups. The value is available to calling programs via
1436 pcre_fullinfo().
1437
1438 26. While implementing 25, I discovered to my embarrassment that pcretest had
1439 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1440 study optimizations had never been tested with that matching function.
1441 Oops. What is worse, even when it was passed study data, there was a bug in
1442 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1443 were also very few tests of studied patterns with pcre_dfa_exec().
1444
1445 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1446 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1447 on the other side of the coin, they are no longer allowed to have different
1448 names, because these cannot be distinguished in PCRE, and this has caused
1449 confusion. (This is a difference from Perl.)
1450
1451 28. When duplicate subpattern names are present (necessarily with different
1452 numbers, as required by 27 above), and a test is made by name in a
1453 conditional pattern, either for a subpattern having been matched, or for
1454 recursion in such a pattern, all the associated numbered subpatterns are
1455 tested, and the overall condition is true if the condition is true for any
1456 one of them. This is the way Perl works, and is also more like the way
1457 testing by number works.
1458
1459
1460 Version 7.9 11-Apr-09
1461 ---------------------
1462
1463 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1464 (pcretest), all targets were linked against these libraries. This included
1465 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1466 libraries. This caused unwanted dependencies to be created. This problem
1467 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1468 pcretest is linked with readline.
1469
1470 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1471 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1472 moved outside it again, because FALSE and TRUE are already defined in AIX,
1473 but BOOL is not.
1474
1475 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1476 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1477
1478 4. The pcregrep documentation said spaces were inserted as well as colons (or
1479 hyphens) following file names and line numbers when outputting matching
1480 lines. This is not true; no spaces are inserted. I have also clarified the
1481 wording for the --colour (or --color) option.
1482
1483 5. In pcregrep, when --colour was used with -o, the list of matching strings
1484 was not coloured; this is different to GNU grep, so I have changed it to be
1485 the same.
1486
1487 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1488 each matching line was coloured. Now it goes on to look for further matches
1489 of any of the test patterns, which is the same behaviour as GNU grep.
1490
1491 7. A pattern that could match an empty string could cause pcregrep to loop; it
1492 doesn't make sense to accept an empty string match in pcregrep, so I have
1493 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1494 seems to be how GNU grep behaves.
1495
1496 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1497 start or after a newline", because the conditional assertion was not being
1498 correctly handled. The rule now is that both the assertion and what follows
1499 in the first alternative must satisfy the test.
1500
1501 9. If auto-callout was enabled in a pattern with a conditional group whose
1502 condition was an assertion, PCRE could crash during matching, both with
1503 pcre_exec() and pcre_dfa_exec().
1504
1505 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1506 used for matching.
1507
1508 11. Unicode property support in character classes was not working for
1509 characters (bytes) greater than 127 when not in UTF-8 mode.
1510
1511 12. Added the -M command line option to pcretest.
1512
1513 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1514
1515 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1516
1517 16. Added comments and documentation about mis-use of no_arg in the C++
1518 wrapper.
1519
1520 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1521 from Martin Jerabek that uses macro names for all relevant character and
1522 string constants.
1523
1524 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1525 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1526 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1527 these, but not everybody uses configure.
1528
1529 19. A conditional group that had only one branch was not being correctly
1530 recognized as an item that could match an empty string. This meant that an
1531 enclosing group might also not be so recognized, causing infinite looping
1532 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1533 with the subject "ab", where knowledge that the repeated group can match
1534 nothing is needed in order to break the loop.
1535
1536 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1537 exec(), but without supplying a callout function, matching went wrong.
1538
1539 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1540 leak if the size of the offset vector was greater than 30. When the vector
1541 is smaller, the saved offsets during recursion go onto a local stack
1542 vector, but for larger vectors malloc() is used. It was failing to free
1543 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1544 error, in fact).
1545
1546 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1547 heapframe that is used only when UTF-8 support is enabled. This caused no
1548 problem, but was untidy.
1549
1550 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1551 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1552 included within another project.
1553
1554 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1555 slightly modified by me:
1556
1557 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1558 not building pcregrep.
1559
1560 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1561 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1562
1563 25. Forward references, both numeric and by name, in patterns that made use of
1564 duplicate group numbers, could behave incorrectly or give incorrect errors,
1565 because when scanning forward to find the reference group, PCRE was not
1566 taking into account the duplicate group numbers. A pattern such as
1567 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1568
1569 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1570 the feature of a custom pointer more persuasive (as requested by a user).
1571
1572 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1573 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1574 pre-defined.
1575
1576 28. Added support for (*UTF8) at the start of a pattern.
1577
1578 29. Arrange for flags added by the "release type" setting in CMake to be shown
1579 in the configuration summary.
1580
1581
1582 Version 7.8 05-Sep-08
1583 ---------------------
1584
1585 1. Replaced UCP searching code with optimized version as implemented for Ad
1586 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1587 stage table and inline lookup instead of a function, giving speed ups of 2
1588 to 5 times on some simple patterns that I tested. Permission was given to
1589 distribute the MultiStage2.py script that generates the tables (it's not in
1590 the tarball, but is in the Subversion repository).
1591
1592 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1593 scripts.
1594
1595 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1596 a group with a zero qualifier. The result of the study could be incorrect,
1597 or the function might crash, depending on the pattern.
1598
1599 4. Caseless matching was not working for non-ASCII characters in back
1600 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1601 It now works when Unicode Property Support is available.
1602
1603 5. In pcretest, an escape such as \x{de} in the data was always generating
1604 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1605 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1606 truncation.
1607
1608 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1609
1610 7. Added two (int) casts to pcregrep when printing the difference of two
1611 pointers, in case they are 64-bit values.
1612
1613 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1614 test 2 if it fails.
1615
1616 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1617 and a #define of that name to empty if it is not externally set. This is to
1618 allow users of MSVC to set it if necessary.
1619
1620 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1621 the convenience functions in the pcre_get.c source file.
1622
1623 11. An option change at the start of a pattern that had top-level alternatives
1624 could cause overwriting and/or a crash. This command provoked a crash in
1625 some environments:
1626
1627 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1628
1629 This potential security problem was recorded as CVE-2008-2371.
1630
1631 12. For a pattern where the match had to start at the beginning or immediately
1632 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1633 pcre_dfa_exec() could read past the end of the passed subject if there was
1634 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1635 pcretest so that it places the subject at the end of its malloc-ed buffer.
1636
1637 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1638 exec() might read past the end of the data buffer in UTF-8 mode.
1639
1640 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1641 the data contained the byte 0x85 as part of a UTF-8 character within its
1642 first line. This applied both to normal and DFA matching.
1643
1644 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1645 /^[^d]*?$/8 failed to match "abc".
1646
1647 16. Added a missing copyright notice to pcrecpp_internal.h.
1648
1649 17. Make it more clear in the documentation that values returned from
1650 pcre_exec() in ovector are byte offsets, not character counts.
1651
1652 18. Tidied a few places to stop certain compilers from issuing warnings.
1653
1654 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1655 supplied by Stefan Weber. I made a further small update for 7.8 because
1656 there is a change of source arrangements: the pcre_searchfuncs.c module is
1657 replaced by pcre_ucd.c.
1658
1659
1660 Version 7.7 07-May-08
1661 ---------------------
1662
1663 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1664 a string to a long long, pretend we don't even have a long long." This is
1665 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1666
1667 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1668 pre-7.6 versions, which defined a global no_arg variable instead of putting
1669 it in the RE class. (See also #8 below.)
1670
1671 3. Remove a line of dead code, identified by coverity and reported by Nuno
1672 Lopes.
1673
1674 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1675
1676 (1) The include/exclude patterns were being applied to the whole pathnames
1677 of files, instead of just to the final components.
1678
1679 (2) If there was more than one level of directory, the subdirectories were
1680 skipped unless they satisfied the include/exclude conditions. This is
1681 inconsistent with GNU grep (and could even be seen as contrary to the
1682 pcregrep specification - which I improved to make it absolutely clear).
1683 The action now is always to scan all levels of directory, and just
1684 apply the include/exclude patterns to regular files.
1685
1686 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1687 --exclude_dir in the tests to avoid scanning .svn directories.
1688
1689 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1690 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1691 doesn't support NULs in patterns.
1692
1693 7. Added some missing "const"s to declarations of static tables in
1694 pcre_compile.c and pcre_dfa_exec.c.
1695
1696 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1697 caused by fix #2 above. (Subsequently also a second patch to fix the
1698 first patch. And a third patch - this was a messy problem.)
1699
1700 9. Applied Craig's patch to remove the use of push_back().
1701
1702 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1703 matching function regexec().
1704
1705 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1706 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1707 references. PCRE supports relative numbers with this syntax (I don't think
1708 Oniguruma does).
1709
1710 12. Previously, a group with a zero repeat such as (...){0} was completely
1711 omitted from the compiled regex. However, this means that if the group
1712 was called as a subroutine from elsewhere in the pattern, things went wrong
1713 (an internal error was given). Such groups are now left in the compiled
1714 pattern, with a new opcode that causes them to be skipped at execution
1715 time.
1716
1717 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1718 to the way PCRE behaves:
1719
1720 (a) A lone ] character is dis-allowed (Perl treats it as data).
1721
1722 (b) A back reference to an unmatched subpattern matches an empty string
1723 (Perl fails the current match path).
1724
1725 (c) A data ] in a character class must be notated as \] because if the
1726 first data character in a class is ], it defines an empty class. (In
1727 Perl it is not possible to have an empty class.) The empty class []
1728 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1729 The negative empty class [^] matches any one character, independently
1730 of the DOTALL setting.
1731
1732 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1733 non-existent subpattern following a character class starting with ']' and
1734 containing () gave an internal compiling error instead of "reference to
1735 non-existent subpattern". Fortunately, when the pattern did exist, the
1736 compiled code was correct. (When scanning forwards to check for the
1737 existencd of the subpattern, it was treating the data ']' as terminating
1738 the class, so got the count wrong. When actually compiling, the reference
1739 was subsequently set up correctly.)
1740
1741 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1742 it was being rejected as not supported by pcre_dfa_exec(), even though
1743 other assertions are supported. I have made pcre_dfa_exec() support
1744 (*FAIL).
1745
1746 16. The implementation of 13c above involved the invention of a new opcode,
1747 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1748 cannot be changed at match time, I realized I could make a small
1749 improvement to matching performance by compiling OP_ALLANY instead of
1750 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1751 on the OP_ANY path.
1752
1753 17. Compiling pcretest on Windows with readline support failed without the
1754 following two fixes: (1) Make the unistd.h include conditional on
1755 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1756
1757 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1758 ncurses library to be included for pcretest when ReadLine support is
1759 requested, but also to allow for it to be overridden. This patch came from
1760 Daniel Bergström.
1761
1762 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1763 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1764 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1765 spotting this.
1766
1767
1768 Version 7.6 28-Jan-08
1769 ---------------------
1770
1771 1. A character class containing a very large number of characters with
1772 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1773 overflow.
1774
1775 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1776 HAVE_LONG_LONG is not defined.
1777
1778 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1779 bring them up to date and include new features. This patch includes:
1780
1781 - Fixed PH's badly added libz and libbz2 support.
1782 - Fixed a problem with static linking.
1783 - Added pcredemo. [But later removed - see 7 below.]
1784 - Fixed dftables problem and added an option.
1785 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1786 HAVE_LONG_LONG.
1787 - Added readline support for pcretest.
1788 - Added an listing of the option settings after cmake has run.
1789
1790 4. A user submitted a patch to Makefile that makes it easy to create
1791 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1792 Makefile.am that cause it to include this special target, without
1793 affecting anything else. Note that the same mingw target plus all
1794 the other distribution libraries and programs are now supported
1795 when configuring with CMake (see 6 below) instead of with
1796 Configure/Make.
1797
1798 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1799 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1800 exported in the Windows port". It has not yet been confirmed that the patch
1801 solves the problem, but it does no harm.
1802
1803 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1804 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1805 with CMake, and also correct the comment about stack recursion.
1806
1807 7. Remove the automatic building of pcredemo from the ./configure system and
1808 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1809 of a program that users should build themselves after PCRE is installed, so
1810 building it automatically is not really right. What is more, it gave
1811 trouble in some build environments.
1812
1813 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1814
1815
1816 Version 7.5 10-Jan-08
1817 ---------------------
1818
1819 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1820 values in parens when parsing an RE using the C++ wrapper."
1821
1822 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1823 Characters greater than 255 were excluded from the class instead of being
1824 included.
1825
1826 3. The same bug as (2) above applied to negated POSIX classes such as
1827 [:^space:].
1828
1829 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1830 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1831 I have changed it.
1832
1833 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1834 first named subpattern) and a construct such as (?&a) would reference the
1835 first named subpattern whose name started with "a" (in other words, the
1836 length check was missing). Both these problems are fixed. "Subpattern name
1837 expected" is now given for (?&) (a zero-length name), and this patch also
1838 makes it give the same error for \k'' (previously it complained that that
1839 was a reference to a non-existent subpattern).
1840
1841 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1842 this is right because (?- can be followed by option settings as well as by
1843 digits. I have, however, made the messages clearer.
1844
1845 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1846 than the number used in the conditional) now cause a compile-time error.
1847 This is actually not compatible with Perl, which accepts such patterns, but
1848 treats the conditional as always being FALSE (as PCRE used to), but it
1849 seems to me that giving a diagnostic is better.
1850
1851 8. Change "alphameric" to the more common word "alphanumeric" in comments
1852 and messages.
1853
1854 9. Fix two occurrences of "backslash" in comments that should have been
1855 "backspace".
1856
1857 10. Remove two redundant lines of code that can never be obeyed (their function
1858 was moved elsewhere).
1859
1860 11. The program that makes PCRE's Unicode character property table had a bug
1861 which caused it to generate incorrect table entries for sequences of
1862 characters that have the same character type, but are in different scripts.
1863 It amalgamated them into a single range, with the script of the first of
1864 them. In other words, some characters were in the wrong script. There were
1865 thirteen such cases, affecting characters in the following ranges:
1866
1867 U+002b0 - U+002c1
1868 U+0060c - U+0060d
1869 U+0061e - U+00612
1870 U+0064b - U+0065e
1871 U+0074d - U+0076d
1872 U+01800 - U+01805
1873 U+01d00 - U+01d77
1874 U+01d9b - U+01dbf
1875 U+0200b - U+0200f
1876 U+030fc - U+030fe
1877 U+03260 - U+0327f
1878 U+0fb46 - U+0fbb1
1879 U+10450 - U+1049d
1880
1881 12. The -o option (show only the matching part of a line) for pcregrep was not
1882 compatible with GNU grep in that, if there was more than one match in a
1883 line, it showed only the first of them. It now behaves in the same way as
1884 GNU grep.
1885
1886 13. If the -o and -v options were combined for pcregrep, it printed a blank
1887 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1888 does the same. The return code can be used to tell if there were any
1889 non-matching lines.
1890
1891 14. Added --file-offsets and --line-offsets to pcregrep.
1892
1893 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1894 infinitely looping recursion. The bug was that positive lookaheads were not
1895 being skipped when checking for a possible empty match (negative lookaheads
1896 and both kinds of lookbehind were skipped).
1897
1898 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1899 inclusion of <windows.h> to before rather than after the definition of
1900 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1901
1902 17. Specifying a possessive quantifier with a specific limit for a Unicode
1903 character property caused pcre_compile() to compile bad code, which led at
1904 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1905 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1906 caused the error; without that there was no problem.
1907
1908 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1909
1910 19. Added --enable-pcretest-libreadline.
1911
1912 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1913 RE::GlobalReplace(). As a result, the number of replacements returned was
1914 double what it should be. I removed one of the increments, but Craig sent a
1915 later patch that removed the other one (the right fix) and added unit tests
1916 that check the return values (which was not done before).
1917
1918 21. Several CMake things:
1919
1920 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1921 the names libpcre and libpcreposix, not just pcre and pcreposix.
1922
1923 (2) The above change means that pcretest and pcregrep are now correctly
1924 linked with the newly-built libraries, not previously installed ones.
1925
1926 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1927
1928 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1929 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1930 UTF-8 newline character). The key issue is that the pattern starts .*;
1931 this means that the match must be either at the beginning, or after a
1932 newline. The bug was in the code for advancing after a failed match and
1933 checking that the new position followed a newline. It was not taking
1934 account of UTF-8 characters correctly.
1935
1936 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1937 character classes. PCRE was not treating the sequence [:...:] as a
1938 character class unless the ... were all letters. Perl, however, seems to
1939 allow any characters between [: and :], though of course it rejects as
1940 unknown any "names" that contain non-letters, because all the known class
1941 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1942 for example, whereas PCRE did not - it did not recognize a POSIX character
1943 class. This seemed a bit dangerous, so the code has been changed to be
1944 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1945 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1946 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1947 Perl does, and where it didn't before.
1948
1949 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1950 Windows environments %n is disabled by default.
1951
1952
1953 Version 7.4 21-Sep-07
1954 ---------------------
1955
1956 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1957 means that a class such as [\s] counted as "explicit reference to CR or
1958 LF". That isn't really right - the whole point of the change was to try to
1959 help when there was an actual mention of one of the two characters. So now
1960 the change happens only if \r or \n (or a literal CR or LF) character is
1961 encountered.
1962
1963 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1964 of both had grown to the point where there were only 3 bits left.
1965 Fortunately, there was spare space in the data structure, and so I have
1966 moved the internal flags into a new 16-bit field to free up more option
1967 bits.
1968
1969 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1970 but did not set the internal JCHANGED flag - either of these is enough to
1971 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1972 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1973 start sets both bits.
1974
1975 4. Added options (at build time, compile time, exec time) to change \R from
1976 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1977
1978 5. doc/pcresyntax.html was missing from the distribution.
1979
1980 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1981 compatibility, even though it is no longer used.
1982
1983 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1984 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1985 windows.h file is present (where different names are used). [This was
1986 reversed later after testing - see 16 below.]
1987
1988 8. Changed all #include <config.h> to #include "config.h". There were also
1989 some further <pcre.h> cases that I changed to "pcre.h".
1990
1991 9. When pcregrep was used with the --colour option, it missed the line ending
1992 sequence off the lines that it output.
1993
1994 10. It was pointed out to me that arrays of string pointers cause lots of
1995 relocations when a shared library is dynamically loaded. A technique of
1996 using a single long string with a table of offsets can drastically reduce
1997 these. I have refactored PCRE in four places to do this. The result is
1998 dramatic:
1999
2000 Originally: 290
2001 After changing UCP table: 187
2002 After changing error message table: 43
2003 After changing table of "verbs" 36
2004 After changing table of Posix names 22
2005
2006 Thanks to the folks working on Gregex for glib for this insight.
2007
2008 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2009 unicode-properties was also set.
2010
2011 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2012
2013 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2014 checked only for CRLF.
2015
2016 14. Added casts to pcretest.c to avoid compiler warnings.
2017
2018 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2019
2020 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2021 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2022 entirely. This removes changes made in 7 above.
2023
2024 17. The CMake files have been updated, and there is now more information about
2025 building with CMake in the NON-UNIX-USE document.
2026
2027
2028 Version 7.3 28-Aug-07
2029 ---------------------
2030
2031 1. In the rejigging of the build system that eventually resulted in 7.1, the
2032 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2033 brackets there is not right, since it causes compilers to look for an
2034 installed pcre.h, not the version that is in the source that is being
2035 compiled (which of course may be different). I have changed it back to:
2036
2037 #include "pcre.h"
2038
2039 I have a vague recollection that the change was concerned with compiling in
2040 different directories, but in the new build system, that is taken care of
2041 by the VPATH setting the Makefile.
2042
2043 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2044 when the subject happened to end in the byte 0x85 (e.g. if the last
2045 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2046 characters but of course it shouldn't be taken as a newline when it is part
2047 of another character. The bug was that, for an unlimited repeat of . in
2048 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2049 characters when looking for a newline.
2050
2051 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2052
2053 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2054 in debug output.
2055
2056 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2057 long printing in the pcrecpp unittest when running under MinGW.
2058
2059 6. ESC_K was left out of the EBCDIC table.
2060
2061 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2062 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2063 limit also applies to "virtual nesting" when a pattern is recursive, and in
2064 this case 1000 isn't so big. I have been able to remove this limit at the
2065 expense of backing off one optimization in certain circumstances. Normally,
2066 when pcre_exec() would call its internal match() function recursively and
2067 immediately return the result unconditionally, it uses a "tail recursion"
2068 feature to save stack. However, when a subpattern that can match an empty
2069 string has an unlimited repetition quantifier, it no longer makes this
2070 optimization. That gives it a stack frame in which to save the data for
2071 checking that an empty string has been matched. Previously this was taken
2072 from the 1000-entry workspace that had been reserved. So now there is no
2073 explicit limit, but more stack is used.
2074
2075 8. Applied Daniel's patches to solve problems with the import/export magic
2076 syntax that is required for Windows, and which was going wrong for the
2077 pcreposix and pcrecpp parts of the library. These were overlooked when this
2078 problem was solved for the main library.
2079
2080 9. There were some crude static tests to avoid integer overflow when computing
2081 the size of patterns that contain repeated groups with explicit upper
2082 limits. As the maximum quantifier is 65535, the maximum group length was
2083 set at 30,000 so that the product of these two numbers did not overflow a
2084 32-bit integer. However, it turns out that people want to use groups that
2085 are longer than 30,000 bytes (though not repeat them that many times).
2086 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2087 made it possible to implement the integer overflow checks in a much more
2088 dynamic way, which I have now done. The artificial limitation on group
2089 length has been removed - we now have only the limit on the total length of
2090 the compiled pattern, which depends on the LINK_SIZE setting.
2091
2092 10. Fixed a bug in the documentation for get/copy named substring when
2093 duplicate names are permitted. If none of the named substrings are set, the
2094 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2095 empty string.
2096
2097 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2098 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2099 because the ] is interpreted as the first data character and the
2100 terminating ] is not found. PCRE has been made compatible with Perl in this
2101 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2102 cause memory overwriting.
2103
2104 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2105 string has been matched (to stop an infinite loop). It was not recognizing
2106 a conditional subpattern that could match an empty string if that
2107 subpattern was within another subpattern. For example, it looped when
2108 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2109 condition was not nested. This bug has been fixed.
2110
2111 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2112 past the start of the subject in the presence of bytes with the top bit
2113 set, for example "\x8aBCD".
2114
2115 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2116 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2117
2118 14. Optimized (?!) to (*FAIL).
2119
2120 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2121 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2122 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2123 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2124 does: it's just the validity check that is more restrictive.
2125
2126 16. Inserted checks for integer overflows during escape sequence (backslash)
2127 processing, and also fixed erroneous offset values for syntax errors during
2128 backslash processing.
2129
2130 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2131 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2132
2133 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2134 caused an overrun.
2135
2136 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2137 something other than just ASCII characters) inside a group that had an
2138 unlimited repeat caused a loop at compile time (while checking to see
2139 whether the group could match an empty string).
2140
2141 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2142 [\P{Any}] did so. (Error in the code for printing property names.)
2143
2144 21. An orphan \E inside a character class could cause a crash.
2145
2146 22. A repeated capturing bracket such as (A)? could cause a wild memory
2147 reference during compilation.
2148
2149 23. There are several functions in pcre_compile() that scan along a compiled
2150 expression for various reasons (e.g. to see if it's fixed length for look
2151 behind). There were bugs in these functions when a repeated \p or \P was
2152 present in the pattern. These operators have additional parameters compared
2153 with \d, etc, and these were not being taken into account when moving along
2154 the compiled data. Specifically:
2155
2156 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2157 length.
2158
2159 (b) An item such as \pL+ within a repeated group could cause crashes or
2160 loops.
2161
2162 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2163 "reference to non-existent subpattern" error.
2164
2165 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2166
2167 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2168 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2169
2170 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2171
2172 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2173 character were causing crashes (broken optimization).
2174
2175 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2176 \p or \P) caused a compile-time loop.
2177
2178 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2179 break. For example, the unstudied pattern [\r\n]A does not match the string
2180 "\r\nA" because change 7.0/46 below moves the current point on by two
2181 characters after failing to match at the start. However, the pattern \nA
2182 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2183 the same is true. There doesn't seem any very clean way out of this, but
2184 what I have chosen to do makes the common cases work: PCRE now takes note
2185 of whether there can be an explicit match for \r or \n anywhere in the
2186 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2187 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2188 pattern has explicit CR or LF references.
2189
2190 29. Added (*CR) etc for changing newline setting at start of pattern.
2191
2192
2193 Version 7.2 19-Jun-07
2194 ---------------------
2195
2196 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2197 which is apparently normally available under Windows.
2198
2199 2. Re-jig the pcregrep tests with different newline settings in an attempt
2200 to make them independent of the local environment's newline setting.
2201
2202 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2203
2204 4. Some of the "internals" tests were previously cut out when the link size
2205 was not 2, because the output contained actual offsets. The recent new
2206 "Z" feature of pcretest means that these can be cut out, making the tests
2207 usable with all link sizes.
2208
2209 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2210 stack recursion. This gives a massive performance boost under BSD, but just
2211 a small improvement under Linux. However, it saves one field in the frame
2212 in all cases.
2213
2214 6. Added more features from the forthcoming Perl 5.10:
2215
2216 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2217 recursion call. It refers to the nth most recently opened parentheses.
2218
2219 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2220 to be opened parentheses.
2221
2222 (c) Conditions that refer to capturing parentheses can be specified
2223 relatively, for example, (?(-2)... or (?(+3)...
2224
2225 (d) \K resets the start of the current match so that everything before
2226 is not part of it.
2227
2228 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2229
2230 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2231 reference syntax.
2232
2233 (g) (?| introduces a group in which the numbering of parentheses in each
2234 alternative starts with the same number.
2235
2236 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2237
2238 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2239 PCRE_INFO_JCHANGED.
2240
2241 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2242 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2243 for detecting groups that can match an empty string.
2244
2245 9. A pattern with a very large number of alternatives (more than several
2246 hundred) was running out of internal workspace during the pre-compile
2247 phase, where pcre_compile() figures out how much memory will be needed. A
2248 bit of new cunning has reduced the workspace needed for groups with
2249 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2250 workspace instead of running out of the 4096 that are available.
2251
2252 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2253
2254 11. Applied patch from Google to remove an optimization that didn't quite work.
2255 The report of the bug said:
2256
2257 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2258 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2259 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2260
2261 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2262 it matched the wrong number of bytes.
2263
2264
2265 Version 7.1 24-Apr-07
2266 ---------------------
2267
2268 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2269 that is more "standard", making use of automake and other Autotools. There
2270 is some re-arrangement of the files and adjustment of comments consequent
2271 on this.
2272
2273 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2274 for recursive directory scanning broke on some systems because the files
2275 are not scanned in any specific order and on different systems the order
2276 was different. A call to "sort" has been inserted into RunGrepTest for the
2277 approprate test as a short-term fix. In the longer term there may be an
2278 alternative.
2279
2280 3. I had an email from Eric Raymond about problems translating some of PCRE's
2281 man pages to HTML (despite the fact that I distribute HTML pages, some
2282 people do their own conversions for various reasons). The problems
2283 concerned the use of low-level troff macros .br and .in. I have therefore
2284 removed all such uses from the man pages (some were redundant, some could
2285 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2286 HTML has been updated to handle .nf/.fi and to complain if it encounters
2287 .br or .in.
2288
2289 4. Updated comments in configure.ac that get placed in config.h.in and also
2290 arranged for config.h to be included in the distribution, with the name
2291 config.h.generic, for the benefit of those who have to compile without
2292 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2293
2294 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2295 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2296 makevp.bat for the current PCRE, using the additional files
2297 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2298
2299 6. A Windows user reported a minor discrepancy with test 2, which turned out
2300 to be caused by a trailing space on an input line that had got lost in his
2301 copy. The trailing space was an accident, so I've just removed it.
2302
2303 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2304 that is needed.
2305
2306 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2307 as "const" (a) because they are and (b) because it helps the PHP
2308 maintainers who have recently made a script to detect big data structures
2309 in the php code that should be moved to the .rodata section. I remembered
2310 to update Builducptable as well, so it won't revert if ucptable.h is ever
2311 re-created.
2312
2313 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2314 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2315 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2316 support is not required. This saves 1.5-2K of code, which is important in
2317 some applications.
2318
2319 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2320 so as not to refer to the tables, even though these functions will never be
2321 called when UTF-8 support is disabled. Otherwise there are problems with a
2322 shared library.
2323
2324 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2325
2326 (a) It was defining its arguments as char * instead of void *.
2327
2328 (b) It was assuming that all moves were upwards in memory; this was true
2329 a long time ago when I wrote it, but is no longer the case.
2330
2331 The emulated memove() is provided for those environments that have neither
2332 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2333 is clearly not the case, as these two bugs were recently reported.
2334
2335 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2336 and Detrail to create the HTML documentation, the .txt form of the man
2337 pages, and it removes trailing spaces from listed files. It also creates
2338 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2339 case, it wraps all the #defines with #ifndefs. This script should be run
2340 before "make dist".
2341
2342 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2343 with Unicode property support.
2344
2345 (a) For a maximizing quantifier, if the two different cases of the
2346 character were of different lengths in their UTF-8 codings (there are
2347 some cases like this - I found 11), and the matching function had to
2348 back up over a mixture of the two cases, it incorrectly assumed they
2349 were both the same length.
2350
2351 (b) When PCRE was configured to use the heap rather than the stack for
2352 recursion during matching, it was not correctly preserving the data for
2353 the other case of a UTF-8 character when checking ahead for a match
2354 while processing a minimizing repeat. If the check also involved
2355 matching a wide character, but failed, corruption could cause an
2356 erroneous result when trying to check for a repeat of the original
2357 character.
2358
2359 13. Some tidying changes to the testing mechanism:
2360
2361 (a) The RunTest script now detects the internal link size and whether there
2362 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2363 values substituted by "configure". (The RunGrepTest script already did
2364 this for UTF-8.) The configure.ac script no longer substitutes the
2365 relevant variables.
2366
2367 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2368 with length and offset values. This means that the output is different
2369 for different internal link sizes. Test 2 is skipped for link sizes
2370 other than 2 because of this, bypassing the problem. Unfortunately,
2371 there was also a test in test 3 (the locale tests) that used /B and
2372 failed for link sizes other than 2. Rather than cut the whole test out,
2373 I have added a new /Z option to pcretest that replaces the length and
2374 offset values with spaces. This is now used to make test 3 independent
2375 of link size. (Test 2 will be tidied up later.)
2376
2377 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2378 segmentation fault instead of returning the appropriate error message.
2379
2380 15. In multiline mode when the newline sequence was set to "any", the pattern
2381 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2382 This doesn't seem right; it now treats the CRLF combination as the line
2383 ending, and so does not match in that case. It's only a pattern such as ^$
2384 that would hit this one: something like ^ABC$ would have failed after \r
2385 and then tried again after \r\n.
2386
2387 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2388 in an attempt to make files that differ only in their line terminators
2389 compare equal. This works on Linux.
2390
2391 17. Under certain error circumstances pcregrep might try to free random memory
2392 as it exited. This is now fixed, thanks to valgrind.
2393
2394 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2395 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2396 was because its rules for how to advance for /g after matching an empty
2397 string at the end of a line did not allow for this case. They now check for
2398 it specially.
2399
2400 20. pcretest is supposed to handle patterns and data of any length, by
2401 extending its buffers when necessary. It was getting this wrong when the
2402 buffer for a data line had to be extended.
2403
2404 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2405 CRLF as a newline sequence.
2406
2407 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2408 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2409 I have nevertheless tidied it up.
2410
2411 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2412
2413 24. Added a man page for pcre-config.
2414
2415
2416 Version 7.0 19-Dec-06
2417 ---------------------
2418
2419 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2420 moving to gcc 4.1.1.
2421
2422 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2423 sys/time.h, which is documented as needed for this function. It doesn't
2424 seem to matter on Linux, but it showed up on some releases of OS X.
2425
2426 3. It seems that there are systems where bytes whose values are greater than
2427 127 match isprint() in the "C" locale. The "C" locale should be the
2428 default when a C program starts up. In most systems, only ASCII printing
2429 characters match isprint(). This difference caused the output from pcretest
2430 to vary, making some of the tests fail. I have changed pcretest so that:
2431
2432 (a) When it is outputting text in the compiled version of a pattern, bytes
2433 other than 32-126 are always shown as hex escapes.
2434
2435 (b) When it is outputting text that is a matched part of a subject string,
2436 it does the same, unless a different locale has been set for the match
2437 (using the /L modifier). In this case, it uses isprint() to decide.
2438
2439 4. Fixed a major bug that caused incorrect computation of the amount of memory
2440 required for a compiled pattern when options that changed within the
2441 pattern affected the logic of the preliminary scan that determines the
2442 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2443 that the computed length was too small. The symptoms of this bug were
2444 either the PCRE error "internal error: code overflow" from pcre_compile(),
2445 or a glibc crash with a message such as "pcretest: free(): invalid next
2446 size (fast)". Examples of patterns that provoked this bug (shown in
2447 pcretest format) are:
2448
2449 /(?-x: )/x
2450 /(?x)(?-x: \s*#\s*)/
2451 /((?i)[\x{c0}])/8
2452 /(?i:[\x{c0}])/8
2453
2454 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2455 is now done differently.
2456
2457 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2458 wrapper classes; (b) implement a new function in the C++ scanner that is
2459 more efficient than the old way of doing things because it avoids levels of
2460 recursion in the regex matching; (c) add a paragraph to the documentation
2461 for the FullMatch() function.
2462
2463 6. The escape sequence \n was being treated as whatever was defined as
2464 "newline". Not only was this contrary to the documentation, which states
2465 that \n is character 10 (hex 0A), but it also went horribly wrong when
2466 "newline" was defined as CRLF. This has been fixed.
2467
2468 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2469 was being set to -1 for the "end of line" case (supposedly a value that no
2470 character can have). Though this value is never used (the check for end of
2471 line is "zero bytes in current character"), it caused compiler complaints.
2472 I've changed it to 0xffffffff.
2473
2474 8. In pcre_version.c, the version string was being built by a sequence of
2475 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2476 string (as it is for production releases) called a macro with an empty
2477 argument. The C standard says the result of this is undefined. The gcc
2478 compiler treats it as an empty string (which was what was wanted) but it is
2479 reported that Visual C gives an error. The source has been hacked around to
2480 avoid this problem.
2481
2482 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2483 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2484 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2485 of them did).
2486
2487 10. Originally, pcretest opened its input and output without "b"; then I was
2488 told that "b" was needed in some environments, so it was added for release
2489 5.0 to both the input and output. (It makes no difference on Unix-like
2490 systems.) Later I was told that it is wrong for the input on Windows. I've
2491 now abstracted the modes into two macros, to make it easier to fiddle with
2492 them, and removed "b" from the input mode under Windows.
2493
2494 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2495
2496 12. Added -help and --help to pcretest as an official way of being reminded
2497 of the options.
2498
2499 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2500 and pcrecpp.cc because they annoy compilers at high warning levels.
2501
2502 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2503
2504 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2505 scripts are not C programs :-) and which was not noticed because it works
2506 on Linux.
2507
2508 16. pcretest is supposed to handle any length of pattern and data line (as one
2509 line or as a continued sequence of lines) by extending its input buffer if
2510 necessary. This feature was broken for very long pattern lines, leading to
2511 a string of junk being passed to pcre_compile() if the pattern was longer
2512 than about 50K.
2513
2514 17. I have done a major re-factoring of the way pcre_compile() computes the
2515 amount of memory needed for a compiled pattern. Previously, there was code
2516 that made a preliminary scan of the pattern in order to do this. That was
2517 OK when PCRE was new, but as the facilities have expanded, it has become
2518 harder and harder to keep it in step with the real compile phase, and there
2519 have been a number of bugs (see for example, 4 above). I have now found a
2520 cunning way of running the real compile function in a "fake" mode that
2521 enables it to compute how much memory it would need, while actually only
2522 ever using a few hundred bytes of working memory and without too many
2523 tests of the mode. This should make future maintenance and development
2524 easier. A side effect of this work is that the limit of 200 on the nesting
2525 depth of parentheses has been removed (though this was never a serious
2526 limitation, I suspect). However, there is a downside: pcre_compile() now
2527 runs more slowly than before (30% or more, depending on the pattern). I
2528 hope this isn't a big issue. There is no effect on runtime performance.
2529
2530 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2531 newline (only possible for the last line of a file) and it was a
2532 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2533
2534 19. Added additional timing features to pcretest. (1) The -tm option now times
2535 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2536 separate command line item, by a number that specifies the number of
2537 repeats to use when timing. The default is 50000; this gives better
2538 precision, but takes uncomfortably long for very large patterns.
2539
2540 20. Extended pcre_study() to be more clever in cases where a branch of a
2541 subpattern has no definite first character. For example, (a*|b*)[cd] would
2542 previously give no result from pcre_study(). Now it recognizes that the
2543 first character must be a, b, c, or d.
2544
2545 21. There was an incorrect error "recursive call could loop indefinitely" if
2546 a subpattern (or the entire pattern) that was being tested for matching an
2547 empty string contained only one non-empty item after a nested subpattern.
2548 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2549 incorrectly, because the \d was being skipped in the check.
2550
2551 22. The pcretest program now has a new pattern option /B and a command line
2552 option -b, which is equivalent to adding /B to every pattern. This causes
2553 it to show the compiled bytecode, without the additional information that
2554 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2555 is the same as /B/I).
2556
2557 23. A new optimization is now able automatically to treat some sequences such
2558 as a*b as a*+b. More specifically, if something simple (such as a character
2559 or a simple class like \d) has an unlimited quantifier, and is followed by
2560 something that cannot possibly match the quantified thing, the quantifier
2561 is automatically "possessified".
2562
2563 24. A recursive reference to a subpattern whose number was greater than 39
2564 went wrong under certain circumstances in UTF-8 mode. This bug could also
2565 have affected the operation of pcre_study().
2566
2567 25. Realized that a little bit of performance could be had by replacing
2568 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2569
2570 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2571
2572 27. Possessive quantifiers such as a++ were previously implemented by turning
2573 them into atomic groups such as ($>a+). Now they have their own opcodes,
2574 which improves performance. This includes the automatically created ones
2575 from 23 above.
2576
2577 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2578 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2579 the first matched character to be a colon. This applied both to named and
2580 numbered groups.
2581
2582 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2583
2584 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2585 building PCRE on Windows easier, so I have included it in the distribution.
2586
2587 31. There is now a check in pcretest against a ridiculously large number being
2588 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2589 loop, the loop is abandoned.
2590
2591 32. Forward references to subpatterns in conditions such as (?(2)...) where
2592 subpattern 2 is defined later cause pcre_compile() to search forwards in
2593 the pattern for the relevant set of parentheses. This search went wrong
2594 when there were unescaped parentheses in a character class, parentheses
2595 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2596
2597 33. "Subroutine" calls and backreferences were previously restricted to
2598 referencing subpatterns earlier in the regex. This restriction has now
2599 been removed.
2600
2601 34. Added a number of extra features that are going to be in Perl 5.10. On the
2602 whole, these are just syntactic alternatives for features that PCRE had
2603 previously implemented using the Python syntax or my own invention. The
2604 other formats are all retained for compatibility.
2605
2606 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2607 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2608 also .NET compatible.
2609
2610 (b) A recursion or subroutine call to a named group can now be defined as
2611 (?&name) as well as (?P>name).
2612
2613 (c) A backreference to a named group can now be defined as \k<name> or
2614 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2615 5.10, are also .NET compatible.
2616
2617 (d) A conditional reference to a named group can now use the syntax
2618 (?(<name>) or (?('name') as well as (?(name).
2619
2620 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2621 groups (named and numbered) that are never evaluated inline, but can be
2622 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2623 is always false. There may be only one alternative in such a group.
2624
2625 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2626 as the simple (?(R). The condition is true only if the most recent
2627 recursion is that of the given number or name. It does not search out
2628 through the entire recursion stack.
2629
2630 (g) The escape \gN or \g{N} has been added, where N is a positive or
2631 negative number, specifying an absolute or relative reference.
2632
2633 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2634 some "unreachable code" warnings.
2635
2636 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2637 things, this adds five new scripts.
2638
2639 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2640 There were also incompatibilities regarding the handling of \Q..\E inside
2641 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2642 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2643
2644 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2645 matches an empty string, and forcibly breaks the loop. There were bugs in
2646 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2647 against aaaa the result was just "a" rather than "aaaa", for example. Two
2648 separate and independent bugs (that affected different cases) have been
2649 fixed.
2650
2651 39. Refactored the code to abolish the use of different opcodes for small
2652 capturing bracket numbers. This is a tidy that I avoided doing when I
2653 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2654 The new approach is not only tidier, it makes it possible to reduce the
2655 memory needed to fix the previous bug (38).
2656
2657 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2658 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2659 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2660 mode.
2661
2662 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2663 report.
2664
2665 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2666 copy construction and assignment in the C++ wrapper.
2667
2668 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2669 couple of bugs that could have given wrong results in the "--newline=crlf"
2670 case.
2671
2672 44. Added a number of casts and did some reorganization of signed/unsigned int
2673 variables following suggestions from Dair Grant. Also renamed the variable
2674 "this" as "item" because it is a C++ keyword.
2675
2676 45. Arranged for dftables to add
2677
2678 #include "pcre_internal.h"
2679
2680 to pcre_chartables.c because without it, gcc 4.x may remove the array
2681 definition from the final binary if PCRE is built into a static library and
2682 dead code stripping is activated.
2683
2684 46. For an unanchored pattern, if a match attempt fails at the start of a
2685 newline sequence, and the newline setting is CRLF or ANY, and the next two
2686 characters are CRLF, advance by two characters instead of one.
2687
2688
2689 Version 6.7 04-Jul-06
2690 ---------------------
2691
2692 1. In order to handle tests when input lines are enormously long, pcretest has
2693 been re-factored so that it automatically extends its buffers when
2694 necessary. The code is crude, but this _is_ just a test program. The
2695 default size has been increased from 32K to 50K.
2696
2697 2. The code in pcre_study() was using the value of the re argument before
2698 testing it for NULL. (Of course, in any sensible call of the function, it
2699 won't be NULL.)
2700
2701 3. The memmove() emulation function in pcre_internal.h, which is used on
2702 systems that lack both memmove() and bcopy() - that is, hardly ever -
2703 was missing a "static" storage class specifier.
2704
2705 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2706 containing an extended class (one that cannot be represented by a bitmap
2707 because it contains high-valued characters or Unicode property items, e.g.
2708 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2709 pattern, but PCRE should not loop if you do not (it no longer does).
2710 [Detail: two cases were found: (a) a repeated subpattern containing an
2711 extended class; (b) a recursive reference to a subpattern that followed a
2712 previous extended class. It wasn't skipping over the extended class
2713 correctly when UTF-8 mode was not set.]
2714
2715 5. A negated single-character class was not being recognized as fixed-length
2716 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2717 compile error "lookbehind assertion is not fixed length".
2718
2719 6. The RunPerlTest auxiliary script was showing an unexpected difference
2720 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2721 write a Perl script that can interpret lines of an input file either as
2722 byte characters or as UTF-8, which is what "perltest" was being required to
2723 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2724 can't do is switch easily at run time between having the "use utf8;" pragma
2725 or not. In the end, I fudged it by using the RunPerlTest script to insert
2726 "use utf8;" explicitly for the UTF-8 tests.
2727
2728 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2729 the end of the subject string, contrary to the documentation and to what
2730 Perl does. This was true of both matching functions. Now it matches only at
2731 the start of the subject and immediately after *internal* newlines.
2732
2733 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2734 a pointer to an int instead of a pointer to an unsigned long int. This
2735 caused problems on 64-bit systems.
2736
2737 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2738 instance of the 'standard' template library not being so standard".
2739
2740 10. There was no check on the number of named subpatterns nor the maximum
2741 length of a subpattern name. The product of these values is used to compute
2742 the size of the memory block for a compiled pattern. By supplying a very
2743 long subpattern name and a large number of named subpatterns, the size
2744 computation could be caused to overflow. This is now prevented by limiting
2745 the length of names to 32 characters, and the number of named subpatterns
2746 to 10,000.
2747
2748 11. Subpatterns that are repeated with specific counts have to be replicated in
2749 the compiled pattern. The size of memory for this was computed from the
2750 length of the subpattern and the repeat count. The latter is limited to
2751 65535, but there was no limit on the former, meaning that integer overflow
2752 could in principle occur. The compiled length of a repeated subpattern is
2753 now limited to 30,000 bytes in order to prevent this.
2754
2755 12. Added the optional facility to have named substrings with the same name.
2756
2757 13. Added the ability to use a named substring as a condition, using the
2758 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2759 are numbers (not recommended). Forward references are permitted.
2760
2761 14. Added forward references in named backreferences (if you see what I mean).
2762
2763 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2764 pattern could run off the end of the subject. For example, the pattern
2765 "(?s)(.{1,5})"8 did this with the subject "ab".
2766
2767 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2768 PCRE_CASELESS was set when matching characters that were quantified with ?
2769 or *.
2770
2771 17. A character class other than a single negated character that had a minimum
2772 but no maximum quantifier - for example [ab]{6,} - was not handled
2773 correctly by pce_dfa_exec(). It would match only one character.
2774
2775 18. A valid (though odd) pattern that looked like a POSIX character
2776 class but used an invalid character after [ (for example [[,abc,]]) caused
2777 pcre_compile() to give the error "Failed: internal error: code overflow" or
2778 in some cases to crash with a glibc free() error. This could even happen if
2779 the pattern terminated after [[ but there just happened to be a sequence of
2780 letters, a binary zero, and a closing ] in the memory that followed.
2781
2782 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2783 over the years. Originally (before any Unicode support), just the bottom 8
2784 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2785 output from "man perlunicode" includes this:
2786
2787 The regular expression compiler produces polymorphic opcodes. That
2788 is, the pattern adapts to the data and automatically switches to
2789 the Unicode character scheme when presented with Unicode data--or
2790 instead uses a traditional byte scheme when presented with byte
2791 data.
2792
2793 Sadly, a wide octal escape does not cause a switch, and in a string with
2794 no other multibyte characters, these octal escapes are treated as before.
2795 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2796 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2797 Unicode string.
2798
2799 I have not perpetrated such confusion in PCRE. Up till now, it took just
2800 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2801 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2802 translate to the appropriate multibyte character.
2803
2804 29. Applied some refactoring to reduce the number of warnings from Microsoft
2805 and Borland compilers. This has included removing the fudge introduced
2806 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2807 a warning about an unused variable.
2808
2809 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2810 characters since release 4.0, because Perl (from release 5.004) does not.
2811 [Or at least, is documented not to: some releases seem to be in conflict
2812 with the documentation.] However, when a pattern was studied with
2813 pcre_study() and all its branches started with \s, PCRE still included VT
2814 as a possible starting character. Of course, this did no harm; it just
2815 caused an unnecessary match attempt.
2816
2817 22. Removed a now-redundant internal flag bit that recorded the fact that case
2818 dependency changed within the pattern. This was once needed for "required
2819 byte" processing, but is no longer used. This recovers a now-scarce options
2820 bit. Also moved the least significant internal flag bit to the most-
2821 significant bit of the word, which was not previously used (hangover from
2822 the days when it was an int rather than a uint) to free up another bit for
2823 the future.
2824
2825 23. Added support for CRLF line endings as well as CR and LF. As well as the
2826 default being selectable at build time, it can now be changed at runtime
2827 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2828 specify that it is scanning data with non-default line endings.
2829
2830 24. Changed the definition of CXXLINK to make it agree with the definition of
2831 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2832
2833 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2834 recursions. This makes a big different to stack usage for some patterns.
2835
2836 26. If a subpattern containing a named recursion or subroutine reference such
2837 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2838 the space required for the compiled pattern went wrong and gave too small a
2839 value. Depending on the environment, this could lead to "Failed: internal
2840 error: code overflow at offset 49" or "glibc detected double free or
2841 corruption" errors.
2842
2843 27. Applied patches from Google (a) to support the new newline modes and (b) to
2844 advance over multibyte UTF-8 characters in GlobalReplace.
2845
2846 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2847 difference for some implementation of PCRE in some Windows version.
2848
2849 29. Added some extra testing facilities to pcretest:
2850
2851 \q<number> in a data line sets the "match limit" value
2852 \Q<number> in a data line sets the "match recursion limt" value
2853 -S <number> sets the stack size, where <number> is in megabytes
2854
2855 The -S option isn't available for Windows.
2856
2857
2858 Version 6.6 06-Feb-06
2859 ---------------------
2860
2861 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2862 in pcreposix.h. I have copied the definition from pcre.h.
2863
2864 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2865 because pcre.h is no longer a built file.
2866
2867 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2868 not normally included in the compiled code.
2869
2870
2871 Version 6.5 01-Feb-06
2872 ---------------------
2873
2874 1. When using the partial match feature with pcre_dfa_exec(), it was not
2875 anchoring the second and subsequent partial matches at the new starting
2876 point. This could lead to incorrect results. For example, with the pattern
2877 /1234/, partially matching against "123" and then "a4" gave a match.
2878
2879 2. Changes to pcregrep:
2880
2881 (a) All non-match returns from pcre_exec() were being treated as failures
2882 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2883 error message is output. Some extra information is given for the
2884 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2885 probably the only errors that are likely to be caused by users (by
2886 specifying a regex that has nested indefinite repeats, for instance).
2887 If there are more than 20 of these errors, pcregrep is abandoned.
2888
2889 (b) A binary zero was treated as data while matching, but terminated the
2890 output line if it was written out. This has been fixed: binary zeroes
2891 are now no different to any other data bytes.
2892
2893 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2894 used to set a locale for matching. The --locale=xxxx long option has
2895 been added (no short equivalent) to specify a locale explicitly on the
2896 pcregrep command, overriding the environment variables.
2897
2898 (d) When -B was used with -n, some line numbers in the output were one less
2899 than they should have been.
2900
2901 (e) Added the -o (--only-matching) option.
2902
2903 (f) If -A or -C was used with -c (count only), some lines of context were
2904 accidentally printed for the final match.
2905
2906 (g) Added the -H (--with-filename) option.
2907
2908 (h) The combination of options -rh failed to suppress file names for files
2909 that were found from directory arguments.
2910
2911 (i) Added the -D (--devices) and -d (--directories) options.
2912
2913 (j) Added the -F (--fixed-strings) option.
2914
2915 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2916
2917 (l) Added the --colo(u)r option.
2918
2919 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2920 is not present by default.
2921
2922 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2923 items such as (?R) or (?1), when the recursion could match a number of
2924 alternatives. If it matched one of the alternatives, but subsequently,
2925 outside the recursion, there was a failure, the code tried to back up into
2926 the recursion. However, because of the way PCRE is implemented, this is not
2927 possible, and the result was an incorrect result from the match.
2928
2929 In order to prevent this happening, the specification of recursion has
2930 been changed so that all such subpatterns are automatically treated as
2931 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2932
2933 4. I had overlooked the fact that, in some locales, there are characters for
2934 which isalpha() is true but neither isupper() nor islower() are true. In
2935 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2936 and ordfeminine) are like this. This affected the treatment of \w and \W
2937 when they appeared in character classes, but not when they appeared outside
2938 a character class. The bit map for "word" characters is now created
2939 separately from the results of isalnum() instead of just taking it from the
2940 upper, lower, and digit maps. (Plus the underscore character, of course.)
2941
2942 5. The above bug also affected the handling of POSIX character classes such as
2943 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2944 permanent tables. Instead, the bit maps for such a class were previously
2945 created as the appropriate unions of the upper, lower, and digit bitmaps.
2946 Now they are created by subtraction from the [[:word:]] class, which has
2947 its own bitmap.
2948
2949 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2950 It is created by subtracting the vertical space characters (\x09, \x0a,
2951 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2952 subtraction was done in the overall bitmap for a character class, meaning
2953 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2954 be recognized. This bug has been fixed.
2955
2956 7. Patches from the folks at Google:
2957
2958 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2959 real life, but is still worth protecting against".
2960
2961 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2962 regular expressions".
2963
2964 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2965 have it.
2966
2967 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2968 "configure" and the latter not, in order to fix a problem somebody had
2969 with compiling the Arg class on HP-UX.
2970
2971 (e) Improve the error-handling of the C++ wrapper a little bit.
2972
2973 (f) New tests for checking recursion limiting.
2974
2975 8. The pcre_memmove() function, which is used only if the environment does not
2976 have a standard memmove() function (and is therefore rarely compiled),
2977 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2978 returning a result (though PCRE never actually uses the result).
2979
2980 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2981 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2982 returned instead of calling malloc() with an overflowing number that would
2983 most likely cause subsequent chaos.
2984
2985 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2986
2987 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2988 with this option is matched, the nmatch and pmatch options of regexec() are
2989 ignored.
2990
2991 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2992 provided in case anyone wants to the the POSIX interface with UTF-8
2993 strings.
2994
2995 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2996 C++ linking (needed for some HP-UX environments).
2997
2998 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2999 (unused parameter) and in the pcre_printint() function (omitted "default"
3000 switch label when the default is to do nothing).
3001
3002 15. Added some code to make it possible, when PCRE is compiled as a C++
3003 library, to replace subject pointers for pcre_exec() with a smart pointer
3004 class, thus making it possible to process discontinuous strings.
3005
3006 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3007 much the same function. They were added by different people who were trying
3008 to make PCRE easy to compile on non-Unix systems. It has been suggested
3009 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3010 for compiling on Windows systems. I have therefore replaced it with
3011 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3012 defaults to "extern" for C or "extern C" for C++, which works fine on
3013 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3014 SCOPE with something explicit in config.h. In addition:
3015
3016 (a) pcreposix.h still had just "extern" instead of either of these macros;
3017 I have replaced it with PCRE_DATA_SCOPE.
3018
3019 (b) Functions such as _pcre_xclass(), which are internal to the library,
3020 but external in the C sense, all had PCRE_EXPORT in their definitions.
3021 This is apparently wrong for the Windows case, so I have removed it.
3022 (It makes no difference on Unix-like systems.)
3023
3024 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3025 of recursive calls to match(). This is different to MATCH_LIMIT because
3026 that limits the total number of calls to match(), not all of which increase
3027 the depth of recursion. Limiting the recursion depth limits the amount of
3028 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3029 when PCRE is compiled, and changed at run time. A patch from Google adds
3030 this functionality to the C++ interface.
3031
3032 18. Changes to the handling of Unicode character properties:
3033
3034 (a) Updated the table to Unicode 4.1.0.
3035
3036 (b) Recognize characters that are not in the table as "Cn" (undefined).
3037
3038 (c) I revised the way the table is implemented to a much improved format
3039 which includes recognition of ranges. It now supports the ranges that
3040 are defined in UnicodeData.txt, and it also amalgamates other
3041 characters into ranges. This has reduced the number of entries in the
3042 table from around 16,000 to around 3,000, thus reducing its size
3043 considerably. I realized I did not need to use a tree structure after
3044 all - a binary chop search is just as efficient. Having reduced the
3045 number of entries, I extended their size from 6 bytes to 8 bytes to
3046 allow for more data.
3047
3048 (d) Added support for Unicode script names via properties such as \p{Han}.
3049
3050 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3051 matching that character.
3052
3053 20. When matching a repeated Unicode property with a minimum greater than zero,
3054 (for example \pL{2,}), PCRE could look past the end of the subject if it
3055 reached it while seeking the minimum number of characters. This could
3056 happen only if some of the characters were more than one byte long, because
3057 there is a check for at least the minimum number of bytes.
3058
3059 21. Refactored the implementation of \p and \P so as to be more general, to
3060 allow for more different types of property in future. This has changed the
3061 compiled form incompatibly. Anybody with saved compiled patterns that use
3062 \p or \P will have to recompile them.
3063
3064 22. Added "Any" and "L&" to the supported property types.
3065
3066 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3067 but give a compile time error if the value is greater than 0xff.
3068
3069 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3070 accidentally not being installed or uninstalled.
3071
3072 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3073 made were to insert the current release number. This seemed silly, because
3074 it made things harder for people building PCRE on systems that don't run
3075 "configure". I have turned pcre.h into a distributed file, no longer built
3076 by "configure", with the version identification directly included. There is
3077 no longer a pcre.h.in file.
3078
3079 However, this change necessitated a change to the pcre-config script as
3080 well. It is built from pcre-config.in, and one of the substitutions was the
3081 release number. I have updated configure.ac so that ./configure now finds
3082 the release number by grepping pcre.h.
3083
3084 26. Added the ability to run the tests under valgrind.
3085
3086
3087 Version 6.4 05-Sep-05
3088 ---------------------
3089
3090 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3091 "--" to be printed when multiple files were scanned, even when none of the
3092 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3093 consider it to be a bug, and have restored the previous behaviour.
3094
3095 2. A couple of code tidies to get rid of compiler warnings.
3096
3097 3. The pcretest program used to cheat by referring to symbols in the library
3098 whose names begin with _pcre_. These are internal symbols that are not
3099 really supposed to be visible externally, and in some environments it is
3100 possible to suppress them. The cheating is now confined to including
3101 certain files from the library's source, which is a bit cleaner.
3102
3103 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3104 file's purpose clearer.
3105
3106 5. Reorganized pcre_ucp_findchar().
3107
3108
3109 Version 6.3 15-Aug-05
3110 ---------------------
3111
3112 1. The file libpcre.pc.in did not have general read permission in the tarball.
3113
3114 2. There were some problems when building without C++ support:
3115
3116 (a) If C++ support was not built, "make install" and "make test" still
3117 tried to test it.
3118
3119 (b) There were problems when the value of CXX was explicitly set. Some
3120 changes have been made to try to fix these, and ...
3121
3122 (c) --disable-cpp can now be used to explicitly disable C++ support.
3123
3124 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3125 backslash in a target when C++ was disabled. This confuses some
3126 versions of "make", apparently. Using an intermediate variable solves
3127 this. (Same for CPP_LOBJ.)
3128
3129 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3130 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3131 necessary on certain architectures.
3132
3133 4. Added a setting of -export-symbols-regex to the link command to remove
3134 those symbols that are exported in the C sense, but actually are local
3135 within the library, and not documented. Their names all begin with
3136 "_pcre_". This is not a perfect job, because (a) we have to except some
3137 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3138 available (and never for static libraries). I have made a note to try to
3139 find a way round (a) in the future.
3140
3141
3142 Version 6.2 01-Aug-05
3143 ---------------------
3144
3145 1. There was no test for integer overflow of quantifier values. A construction
3146 such as {1111111111111111} would give undefined results. What is worse, if
3147 a minimum quantifier for a parenthesized subpattern overflowed and became
3148 negative, the calculation of the memory size went wrong. This could have
3149 led to memory overwriting.
3150
3151 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3152
3153 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3154 operating environments where this matters.
3155
3156 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3157 PCRE options from within the C++ wrapper.
3158
3159 5. Named capturing subpatterns were not being correctly counted when a pattern
3160 was compiled. This caused two problems: (a) If there were more than 100
3161 such subpatterns, the calculation of the memory needed for the whole
3162 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3163 back references of the form \12, where the number was greater than 9, were
3164 not recognized as back references, even though there were sufficient
3165 previous subpatterns.
3166
3167 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3168 versions of gcc, e.g. 2.95.4.
3169
3170
3171 Version 6.1 21-Jun-05
3172 ---------------------
3173
3174 1. There was one reference to the variable "posix" in pcretest.c that was not
3175 surrounded by "#if !defined NOPOSIX".
3176
3177 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3178 the cross-check on the old pcre_info() function, for the benefit of the
3179 cut-down version of PCRE that is currently imported into Exim.
3180
3181 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3182 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3183 patterns that start (?i) but I don't think that matters. The use of (?i) is
3184 just an example; this all applies to the other options as well.
3185
3186 4. Since libtool seems to echo the compile commands it is issuing, the output
3187 from "make" can be reduced a bit by putting "@" in front of each libtool
3188 compile command.
3189
3190 5. Patch from the folks at Google for configure.in to be a bit more thorough
3191 in checking for a suitable C++ installation before trying to compile the
3192 C++ stuff. This should fix a reported problem when a compiler was present,
3193 but no suitable headers.
3194
3195 6. The man pages all had just "PCRE" as their title. I have changed them to
3196 be the relevant file name. I have also arranged that these names are
3197 retained in the file doc/pcre.txt, which is a concatenation in text format
3198 of all the man pages except the little individual ones for each function.
3199
3200 7. The NON-UNIX-USE file had not been updated for the different set of source
3201 files that come with release 6. I also added a few comments about the C++
3202 wrapper.
3203
3204
3205 Version 6.0 07-Jun-05
3206 ---------------------
3207
3208 1. Some minor internal re-organization to help with my DFA experiments.
3209
3210 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3211 didn't matter for the library itself when fully configured, but did matter
3212 when compiling without UCP support, or within Exim, where the ucp files are
3213 not imported.
3214
3215 3. Refactoring of the library code to split up the various functions into
3216 different source modules. The addition of the new DFA matching code (see
3217 below) to a single monolithic source would have made it really too
3218 unwieldy, quite apart from causing all the code to be include in a
3219 statically linked application, when only some functions are used. This is
3220 relevant even without the DFA addition now that patterns can be compiled in
3221 one application and matched in another.
3222
3223 The downside of splitting up is that there have to be some external
3224 functions and data tables that are used internally in different modules of
3225 the library but which are not part of the API. These have all had their
3226 names changed to start with "_pcre_" so that they are unlikely to clash
3227 with other external names.
3228
3229 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3230 a different (DFA) algorithm. Although it is slower than the original
3231 function, it does have some advantages for certain types of matching
3232 problem.
3233
3234 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3235 including restarting after a partial match.
3236
3237 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3238 defined when compiling for Windows was sent to me. I have put it into the
3239 code, though I have no means of testing or verifying it.
3240
3241 7. Added the pcre_refcount() auxiliary function.
3242
3243 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3244 match before or at the first newline in the subject string. In pcretest,
3245 the /f option on a pattern can be used to set this.
3246
3247 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3248 would behave wrongly. This has been present in PCRE since release 4.0.
3249
3250 10. A number of changes to the pcregrep command:
3251
3252 (a) Refactored how -x works; insert ^(...)$ instead of setting
3253 PCRE_ANCHORED and checking the length, in preparation for adding
3254 something similar for -w.
3255
3256 (b) Added the -w (match as a word) option.
3257
3258 (c) Refactored the way lines are read and buffered so as to have more
3259 than one at a time available.
3260
3261 (d) Implemented a pcregrep test script.
3262
3263 (e) Added the -M (multiline match) option. This allows patterns to match
3264 over several lines of the subject. The buffering ensures that at least
3265 8K, or the rest of the document (whichever is the shorter) is available
3266 for matching (and similarly the previous 8K for lookbehind assertions).
3267
3268 (f) Changed the --help output so that it now says
3269
3270 -w, --word-regex(p)
3271
3272 instead of two lines, one with "regex" and the other with "regexp"
3273 because that confused at least one person since the short forms are the
3274 same. (This required a bit of code, as the output is generated
3275 automatically from a table. It wasn't just a text change.)
3276
3277 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3278 option but starts with a hyphen. Could be a pattern or a path name
3279 starting with a hyphen, for instance.
3280
3281 (h) "-" can be given as a file name to represent stdin.
3282
3283 (i) When file names are being printed, "(standard input)" is used for
3284 the standard input, for compatibility with GNU grep. Previously
3285 "<stdin>" was used.
3286
3287 (j) The option --label=xxx can be used to supply a name to be used for
3288 stdin when file names are being printed. There is no short form.
3289
3290 (k) Re-factored the options decoding logic because we are going to add
3291 two more options that take data. Such options can now be given in four
3292 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3293
3294 (l) Added the -A, -B, and -C options for requesting that lines of context
3295 around matches be printed.
3296
3297 (m) Added the -L option to print the names of files that do not contain
3298 any matching lines, that is, the complement of -l.
3299
3300 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3301 continue to scan other files.
3302
3303 (o) The -s option was incorrectly implemented. For compatibility with other
3304 greps, it now suppresses the error message for a non-existent or non-
3305 accessible file (but not the return code). There is a new option called
3306 -q that suppresses the output of matching lines, which was what -s was
3307 previously doing.
3308
3309 (p) Added --include and --exclude options to specify files for inclusion
3310 and exclusion when recursing.
3311
3312 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3313 Hopefully, it now does.
3314
3315 12. Missing cast in pcre_study().
3316
3317 13. Added an "uninstall" target to the makefile.
3318
3319 14. Replaced "extern" in the function prototypes in Makefile.in with
3320 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3321 world, but is set differently for Windows.
3322
3323 15. Added a second compiling function called pcre_compile2(). The only
3324 difference is that it has an extra argument, which is a pointer to an
3325 integer error code. When there is a compile-time failure, this is set
3326 non-zero, in addition to the error test pointer being set to point to an
3327 error message. The new argument may be NULL if no error number is required
3328 (but then you may as well call pcre_compile(), which is now just a
3329 wrapper). This facility is provided because some applications need a
3330 numeric error indication, but it has also enabled me to tidy up the way
3331 compile-time errors are handled in the POSIX wrapper.
3332
3333 16. Added VPATH=.libs to the makefile; this should help when building with one
3334 prefix path and installing with another. (Or so I'm told by someone who
3335 knows more about this stuff than I do.)
3336
3337 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3338 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3339 match everything, including newlines. This is not POSIX-compatible, but
3340 somebody wanted the feature. From pcretest it can be activated by using
3341 both the P and the s flags.
3342
3343 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3344
3345 19. libpcre.pc was being incorrectly installed as executable.
3346
3347 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3348 it now also looks for '\r' so that it will work unmodified on Windows.
3349
3350 21. Added Google's contributed C++ wrapper to the distribution.
3351
3352 22. Added some untidy missing memory free() calls in pcretest, to keep
3353 Electric Fence happy when testing.
3354
3355
3356
3357 Version 5.0 13-Sep-04
3358 ---------------------
3359
3360 1. Internal change: literal characters are no longer packed up into items
3361 containing multiple characters in a single byte-string. Each character
3362 is now matched using a separate opcode. However, there may be more than one
3363 byte in the character in UTF-8 mode.
3364
3365 2. The pcre_callout_block structure has two new fields: pattern_position and
3366 next_item_length. These contain the offset in the pattern to the next match
3367 item, and its length, respectively.
3368
3369 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3370 insertion of callouts before each pattern item. Added the /C option to
3371 pcretest to make use of this.
3372
3373 4. On the advice of a Windows user, the lines
3374
3375 #if defined(_WIN32) || defined(WIN32)
3376 _setmode( _fileno( stdout ), 0x8000 );
3377 #endif /* defined(_WIN32) || defined(WIN32) */
3378
3379 have been added to the source of pcretest. This apparently does useful
3380 magic in relation to line terminators.
3381
3382 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3383 for the benefit of those environments where the "b" makes a difference.
3384
3385 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3386 to know about it. I have put a hack into configure.in that adds in code
3387 to set GCC=yes if CC=icc. This seems to end up at a point in the
3388 generated configure script that is early enough to affect the setting of
3389 compiler options, which is what is needed, but I have no means of testing
3390 whether it really works. (The user who reported this had patched the
3391 generated configure script, which of course I cannot do.)
3392
3393 LATER: After change 22 below (new libtool files), the configure script
3394 seems to know about icc (and also ecc). Therefore, I have commented out
3395 this hack in configure.in.
3396
3397 7. Added support for pkg-config (2 patches were sent in).
3398
3399 8. Negated POSIX character classes that used a combination of internal tables
3400 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3401 [[:^ascii]]. Typically, they would match almost any characters. The other
3402 POSIX classes were not broken in this way.
3403
3404 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3405 to find the match, as PCRE was deluded into thinking that the match had to
3406 start at the start point or following a newline. The same bug applied to
3407 patterns with negative forward assertions or any backward assertions
3408 preceding ".*" at the start, unless the pattern required a fixed first
3409 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3410
3411 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3412 starting at the last subject character, bytes beyond the end of the subject
3413 string were read.
3414
3415 11. Renamed the variable "class" as "classbits" to make life easier for C++
3416 users. (Previously there was a macro definition, but it apparently wasn't
3417 enough.)
3418
3419 12. Added the new field "tables" to the extra data so that tables can be passed
3420 in at exec time, or the internal tables can be re-selected. This allows
3421 a compiled regex to be saved and re-used at a later time by a different
3422 program that might have everything at different addresses.
3423
3424 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3425 -R library as well as a -L library.
3426
3427 14. The debugging options of pcretest (-d on the command line or D on a
3428 pattern) showed incorrect output for anything following an extended class
3429 that contained multibyte characters and which was followed by a quantifier.
3430
3431 15. Added optional support for general category Unicode character properties
3432 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3433 support. It adds about 90K to the size of the library. The meanings of the
3434 inbuilt class escapes such as \d and \s have NOT been changed.
3435
3436 16. Updated pcredemo.c to include calls to free() to release the memory for the
3437 compiled pattern.
3438
3439 17. The generated file chartables.c was being created in the source directory
3440 instead of in the building directory. This caused the build to fail if the
3441 source directory was different from the building directory, and was
3442 read-only.
3443
3444 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3445 file. No doubt somebody will tell me if they don't make sense... Also added
3446 Dan Mooney's comments about building on OpenVMS.
3447
3448 19. Added support for partial matching via the PCRE_PARTIAL option for
3449 pcre_exec() and the \P data escape in pcretest.
3450
3451 20. Extended pcretest with 3 new pattern features:
3452
3453 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3454 write the compiled pattern to the file whose name is "rest-of-line".
3455 This is a straight binary dump of the data, with the saved pointer to
3456 the character tables forced to be NULL. The study data, if any, is
3457 written too. After writing, pcretest reads a new pattern.
3458
3459 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3460 compiled pattern from the given file. There must not be any
3461 occurrences of "<" in the file name (pretty unlikely); if there are,
3462 pcretest will instead treat the initial "<" as a pattern delimiter.
3463 After reading in the pattern, pcretest goes on to read data lines as
3464 usual.
3465
3466 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3467 and 16-bit fields in a compiled pattern, to simulate a pattern that
3468 was compiled on a host of opposite endianness.
3469
3470 21. The pcre-exec() function can now cope with patterns that were compiled on
3471 hosts of opposite endianness, with this restriction:
3472
3473 As for any compiled expression that is saved and used later, the tables
3474 pointer field cannot be preserved; the extra_data field in the arguments
3475 to pcre_exec() should be used to pass in a tables address if a value
3476 other than the default internal tables were used at compile time.
3477
3478 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3479 now diagnosed as an error. Previously, most of the time, a negative number
3480 would have been treated as zero, but if in addition "ovector" was passed as
3481 NULL, a crash could occur.
3482
3483 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3484 new versions from the libtool 1.5 distribution (the last one is a copy of
3485 a file called libtool.m4). This seems to have fixed the need to patch
3486 "configure" to support Darwin 1.3 (which I used to do). However, I still
3487 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3488 workstation).
3489
3490 24. Changed the PCRE licence to be the more standard "BSD" licence.
3491
3492
3493 Version 4.5 01-Dec-03
3494 ---------------------
3495
3496 1. There has been some re-arrangement of the code for the match() function so
3497 that it can be compiled in a version that does not call itself recursively.
3498 Instead, it keeps those local variables that need separate instances for
3499 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3500 needs to "recurse". Keeping track of where control must go is done by means
3501 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3502 hide most of the details from the main code, and operates only if
3503 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3504 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3505 operating.
3506
3507 To make it easier for callers to provide specially tailored get/free
3508 functions for this usage, two new functions, pcre_stack_malloc, and
3509 pcre_stack_free, are used. They are always called in strict stacking order,
3510 and the size of block requested is always the same.
3511
3512 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3513 PCRE has been compiled to use the stack or the heap for recursion. The
3514 -C option of pcretest uses this to show which version is compiled.
3515
3516 A new data escape \S, is added to pcretest; it causes the amounts of store
3517 obtained and freed by both kinds of malloc/free at match time to be added
3518 to the output.
3519
3520 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3521 what's available on my current Linux desktop machine.
3522
3523 3. When matching a UTF-8 string, the test for a valid string at the start has
3524 been extended. If start_offset is not zero, PCRE now checks that it points
3525 to a byte that is the start of a UTF-8 character. If not, it returns
3526 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3527 this is necessary because there may be backward assertions in the pattern.
3528 When matching the same subject several times, it may save resources to use
3529 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3530
3531 4. The code for checking the validity of UTF-8 strings has been tightened so
3532 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3533 containing "overlong sequences".
3534
3535 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3536 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3537 should have been "&", but it just so happened that all the cases this let
3538 through by mistake were picked up later in the function.
3539
3540 6. I had used a variable called "isblank" - this is a C99 function, causing
3541 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3542
3543 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3544 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3545 so the version is shown; (c) stop "make test" echoing "./RunTest".
3546
3547 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3548
3549 9. The prototype for memmove() for systems that don't have it was using
3550 size_t, but the inclusion of the header that defines size_t was later. I've
3551 moved the #includes for the C headers earlier to avoid this.
3552
3553 10. Added some adjustments to the code to make it easier to compiler on certain
3554 special systems:
3555
3556 (a) Some "const" qualifiers were missing.
3557 (b) Added the macro EXPORT before all exported functions; by default this
3558 is defined to be empty.
3559 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3560 that it reads its output file name as an argument instead of writing
3561 to the standard output and assuming this can be redirected.
3562
3563 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3564 class containing characters with values greater than 255, PCRE compilation
3565 went into a loop.
3566
3567 12. A recursive reference to a subpattern that was within another subpattern
3568 that had a minimum quantifier of zero caused PCRE to crash. For example,
3569 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3570 recursion. If the recursively-called subpattern itself had a zero repeat,
3571 that was OK.
3572
3573 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3574 buffer into which it was copied (for escape processing) was still set at
3575 1024, so long lines caused crashes.
3576
3577 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3578 "internal error: code overflow...". This applied to any character class
3579 that was followed by a possessive quantifier.
3580
3581 15. Modified the Makefile to add libpcre.la as a prerequisite for
3582 libpcreposix.la because I was told this is needed for a parallel build to
3583 work.
3584
3585 16. If a pattern that contained .* following optional items at the start was
3586 studied, the wrong optimizing data was generated, leading to matching
3587 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3588 matching string must start with a or b or c. The correct conclusion for
3589 this pattern is that a match can start with any character.
3590
3591
3592 Version 4.4 13-Aug-03
3593 ---------------------
3594
3595 1. In UTF-8 mode, a character class containing characters with values between
3596 127 and 255 was not handled correctly if the compiled pattern was studied.
3597 In fixing this, I have also improved the studying algorithm for such
3598 classes (slightly).
3599
3600 2. Three internal functions had redundant arguments passed to them. Removal
3601 might give a very teeny performance improvement.
3602
3603 3. Documentation bug: the value of the capture_top field in a callout is *one
3604 more than* the number of the hightest numbered captured substring.
3605
3606 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3607 in incorrectly linking with a previously installed version. They now link
3608 explicitly with libpcre.la.
3609
3610 5. configure.in no longer needs to recognize Cygwin specially.
3611
3612 6. A problem in pcre.in for Windows platforms is fixed.
3613
3614 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3615 pcretest, it used to include the size of the study block as part of its
3616 output. Unfortunately, the structure contains a field that has a different
3617 size on different hardware architectures. This meant that the tests that
3618 showed this size failed. As the block is currently always of a fixed size,
3619 this information isn't actually particularly useful in pcretest output, so
3620 I have just removed it.
3621
3622 8. Three pre-processor statements accidentally did not start in column 1.
3623 Sadly, there are *still* compilers around that complain, even though
3624 standard C has not required this for well over a decade. Sigh.
3625
3626 9. In pcretest, the code for checking callouts passed small integers in the
3627 callout_data field, which is a void * field. However, some picky compilers
3628 complained about the casts involved for this on 64-bit systems. Now
3629 pcretest passes the address of the small integer instead, which should get
3630 rid of the warnings.
3631
3632 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3633 both compile and run time, and gives an error if an invalid UTF-8 sequence
3634 is found. There is a option for disabling this check in cases where the
3635 string is known to be correct and/or the maximum performance is wanted.
3636
3637 11. In response to a bug report, I changed one line in Makefile.in from
3638
3639 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3640 to
3641 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3642
3643 to look similar to other lines, but I have no way of telling whether this
3644 is the right thing to do, as I do not use Windows. No doubt I'll get told
3645 if it's wrong...
3646
3647
3648 Version 4.3 21-May-03
3649 ---------------------
3650
3651 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3652 Makefile.
3653
3654 2. Some refactoring to improve the quality of the code:
3655
3656 (i) The utf8_table... variables are now declared "const".
3657
3658 (ii) The code for \cx, which used the "case flipping" table to upper case
3659 lower case letters, now just substracts 32. This is ASCII-specific,
3660 but the whole concept of \cx is ASCII-specific, so it seems
3661 reasonable.
3662
3663 (iii) PCRE was using its character types table to recognize decimal and
3664 hexadecimal digits in the pattern. This is silly, because it handles
3665 only 0-9, a-f, and A-F, but the character types table is locale-
3666 specific, which means strange things might happen. A private
3667 table is now used for this - though it costs 256 bytes, a table is
3668 much faster than multiple explicit tests. Of course, the standard
3669 character types table is still used for matching digits in subject
3670 strings against \d.
3671
3672 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3673 ending in _t are). So I've renamed it as ESC_tee.
3674
3675 3. The first argument for regexec() in the POSIX wrapper should have been
3676 defined as "const".
3677
3678 4. Changed pcretest to use malloc() for its buffers so that they can be
3679 Electric Fenced for debugging.
3680
3681 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3682 to read one or more bytes before the start of the subject string. Often this
3683 had no effect on PCRE's behaviour, but in some circumstances it could
3684 provoke a segmentation fault.
3685
3686 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3687 to try to read one or more bytes before the start of the subject string.
3688
3689 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3690 UTF-8 support could misbehave in various ways if the subject string
3691 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3692 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3693 back over UTF-8 characters.)
3694
3695
3696 Version 4.2 14-Apr-03
3697 ---------------------
3698
3699 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3700
3701 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3702 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3703 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3704 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3705 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3706 and BUILD_EXEEXT
3707 Note: automatic setting of the BUILD variables is not yet working
3708 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3709 compile-time but not at link-time
3710 [LINK]: use for linking executables only
3711 make different versions for Windows and non-Windows
3712 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3713 libraries
3714 [LINK_FOR_BUILD]: new variable
3715 [OBJEXT]: use throughout
3716 [EXEEXT]: use throughout
3717 <winshared>: new target
3718 <wininstall>: new target
3719 <dftables.o>: use native compiler
3720 <dftables>: use native linker
3721 <install>: handle Windows platform correctly
3722 <clean>: ditto
3723 <check>: ditto
3724 copy DLL to top builddir before testing
3725
3726 As part of these changes, -no-undefined was removed again. This was reported
3727 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3728 in any case.
3729
3730 3. Some tidies to get rid of compiler warnings:
3731
3732 . In the match_data structure, match_limit was an unsigned long int, whereas
3733 match_call_count was an int. I've made them both unsigned long ints.
3734
3735 . In pcretest the fact that a const uschar * doesn't automatically cast to
3736 a void * provoked a warning.
3737
3738 . Turning on some more compiler warnings threw up some "shadow" variables
3739 and a few more missing casts.
3740
3741 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3742 option, a class that contained a single character with a value between 128
3743 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3744
3745 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3746 option, a class that contained several characters, but with at least one
3747 whose value was between 128 and 255 caused PCRE to crash.
3748
3749
3750 Version 4.1 12-Mar-03
3751 ---------------------
3752
3753 1. Compiling with gcc -pedantic found a couple of places where casts were
3754 needed, and a string in dftables.c that was longer than standard compilers are
3755 required to support.
3756
3757 2. Compiling with Sun's compiler found a few more places where the code could
3758 be tidied up in order to avoid warnings.
3759
3760 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3761 first of these names is deprecated in the latest Autoconf in favour of the name
3762 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3763 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3764 analogy I have changed it to CFLAGS_FOR_BUILD.
3765
3766 4. Added -no-undefined to the linking command in the Makefile, because this is
3767 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3768 linking step for the pcreposix library.
3769
3770 5. PCRE was failing to diagnose the case of two named groups with the same
3771 name.
3772
3773 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3774 literal character that is needed in the subject for a match, and scans along to
3775 ensure that it is present before embarking on the full matching process. This
3776 saves time in cases of nested unlimited repeats that are never going to match.
3777 Problem: the scan can take a lot of time if the subject is very long (e.g.
3778 megabytes), thus penalizing straightforward matches. It is now done only if the
3779 amount of subject to be scanned is less than 1000 bytes.
3780
3781 7. A lesser problem with the same optimization is that it was recording the
3782 first character of an anchored pattern as "needed", thus provoking a search
3783 right along the subject, even when the first match of the pattern was going to
3784 fail. The "needed" character is now not set for anchored patterns, unless it
3785 follows something in the pattern that is of non-fixed length. Thus, it still
3786 fulfils its original purpose of finding quick non-matches in cases of nested
3787 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3788
3789
3790 Version 4.0 17-Feb-03
3791 ---------------------
3792
3793 1. If a comment in an extended regex that started immediately after a meta-item
3794 extended to the end of string, PCRE compiled incorrect data. This could lead to
3795 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3796
3797 2. Moved to autoconf 2.53 and libtool 1.4.2.
3798
3799 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3800 the special perltest8 script is no longer needed - all the tests can be run
3801 from a single perltest script.
3802
3803 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3804 by \s. It has now been removed in PCRE. This means it isn't recognized as
3805 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3806 class [:space:] *does* include VT, thereby creating a mess.
3807
3808 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3809 space and tab.
3810
3811 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3812 its new features into the main test script, reducing the number of scripts.
3813
3814 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3815 were backward compatible, and made the (?i) apply to the whole pattern, as if
3816 /i were given. Now it behaves more logically, and applies the option setting
3817 only to what follows. PCRE has been changed to follow suit. However, if it
3818 finds options settings right at the start of the pattern, it extracts them into
3819 the global options, as before. Thus, they show up in the info data.
3820
3821 8. Added support for the \Q...\E escape sequence. Characters in between are
3822 treated as literals. This is slightly different from Perl in that $ and @ are
3823 also handled as literals inside the quotes. In Perl, they will cause variable
3824 interpolation. Note the following examples:
3825
3826 Pattern PCRE matches Perl matches
3827
3828 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3829 \Qabc\$xyz\E abc\$xyz abc\$xyz
3830 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3831
3832 For compatibility with Perl, \Q...\E sequences are recognized inside character
3833 classes as well as outside them.
3834
3835 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3836 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3837 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3838 signed/unsigned warnings.
3839
3840 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3841 option for pcretest, so I've replaced it by a simple function that does just
3842 that job.
3843
3844 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3845 "pcregrep -".
3846
3847 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3848 Java package. This provides some syntactic sugar for simple cases of what my
3849 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3850 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3851 item, you can use this simplified notation. Note that only makes sense with
3852 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3853 greediness, whatever the setting of the PCRE_UNGREEDY option.
3854
3855 13. A change of greediness default within a pattern was not taking effect at
3856 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3857 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3858 was abstracted outside.
3859
3860 14. PCRE now supports the \G assertion. It is true when the current matching
3861 position is at the start point of the match. This differs from \A when the
3862 starting offset is non-zero. Used with the /g option of pcretest (or similar
3863 code), it works in the same way as it does for Perl's /g option. If all
3864 alternatives of a regex begin with \G, the expression is anchored to the start
3865 match position, and the "anchored" flag is set in the compiled expression.
3866
3867 15. Some bugs concerning the handling of certain option changes within patterns
3868 have been fixed. These applied to options other than (?ims). For example,
3869 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3870 been the other way round. Some of this was related to change 7 above.
3871
3872 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3873 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3874 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3875 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3876
3877 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3878 mode. Unlike ".", it always matches newline, whatever the setting of
3879 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3880 assertions. Perl allows it, but it doesn't (in general) work because it can't
3881 calculate the length of the lookbehind. At least, that's the case for Perl
3882 5.8.0 - I've been told they are going to document that it doesn't work in
3883 future.
3884
3885 18. Added an error diagnosis for escapes that PCRE does not support: these are
3886 \L, \l, \N, \P, \p, \U, \u, and \X.
3887
3888 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3889 reading past the end of the pattern in cases such as /[abcd/.
3890
3891 20. PCRE was getting more memory than necessary for patterns with classes that
3892 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3893
3894 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3895 compiling PCRE for use with Virtual Pascal.
3896
3897 22. Small fix to the Makefile to make it work properly if the build is done
3898 outside the source tree.
3899
3900 23. Added a new extension: a condition to go with recursion. If a conditional
3901 subpattern starts with (?(R) the "true" branch is used if recursion has
3902 happened, whereas the "false" branch is used only at the top level.
3903
3904 24. When there was a very long string of literal characters (over 255 bytes
3905 without UTF support, over 250 bytes with UTF support), the computation of how
3906 much memory was required could be incorrect, leading to segfaults or other
3907 strange effects.
3908
3909 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3910 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3911 there was a subsequent back reference to those brackets. This meant that, for
3912 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3913 possible to check for precisely this case. All we can do is abandon the
3914 optimization if .* occurs inside capturing brackets when there are any back
3915 references whatsoever. (See below for a better fix that came later.)
3916
3917 26. The handling of the optimization for finding the first character of a
3918 non-anchored pattern, and for finding a character that is required later in the
3919 match were failing in some cases. This didn't break the matching; it just
3920 failed to optimize when it could. The way this is done has been re-implemented.
3921
3922 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3923
3924 28. Added a new feature that provides some of the functionality that Perl
3925 provides with (?{...}). The facility is termed a "callout". The way it is done
3926 in PCRE is for the caller to provide an optional function, by setting
3927 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3928 global variable. By default it is unset, which disables all calling out. To get
3929 the function called, the regex must include (?C) at appropriate points. This
3930 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3931 This provides a means of identifying different callout points. When PCRE
3932 reaches such a point in the regex, if pcre_callout has been set, the external
3933 function is called. It is provided with data in a structure called
3934 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3935 matching continues; if it returns a non-zero value, the match at the current
3936 point fails. However, backtracking will occur if possible. [This was changed
3937 later and other features added - see item 49 below.]
3938
3939 29. pcretest is upgraded to test the callout functionality. It provides a
3940 callout function that displays information. By default, it shows the start of
3941 the match and the current position in the text. There are some new data escapes
3942 to vary what happens:
3943
3944 \C+ in addition, show current contents of captured substrings
3945 \C- do not supply a callout function
3946 \C!n return 1 when callout number n is reached
3947 \C!n!m return 1 when callout number n is reached for the mth time
3948
3949 30. If pcregrep was called with the -l option and just a single file name, it
3950 output "<stdin>" if a match was found, instead of the file name.
3951
3952 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3953 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3954 pcre_exec(). This saves a malloc/free per call. The default value of
3955 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3956 when configuring.
3957
3958 32. The default maximum size of a compiled pattern is 64K. There have been a
3959 few cases of people hitting this limit. The code now uses macros to handle the
3960 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3961 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3962 configuring. Tests 2 and 5 work only with 2-byte links because they output
3963 debugging information about compiled patterns.
3964
3965 33. Internal code re-arrangements:
3966
3967 (a) Moved the debugging function for printing out a compiled regex into
3968 its own source file (printint.c) and used #include to pull it into
3969 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3970 separate copies.
3971
3972 (b) Defined the list of op-code names for debugging as a macro in
3973 internal.h so that it is next to the definition of the opcodes.
3974
3975 (c) Defined a table of op-code lengths for simpler skipping along compiled
3976 code. This is again a macro in internal.h so that it is next to the
3977 definition of the opcodes.
3978
3979 34. Added support for recursive calls to individual subpatterns, along the
3980 lines of Robin Houston's patch (but implemented somewhat differently).
3981
3982 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3983 allow it to read and process whole directories in Win32. This code was
3984 contributed by Lionel Fourquaux; it has not been tested by me.
3985
3986 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3987 used to name a group. Names consist of alphanumerics and underscores, and must
3988 be unique. Back references use the syntax (?P=name) and recursive calls use
3989 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3990 numbers. The function pcre_fullinfo() can be used after compilation to extract
3991 a name/number map. There are three relevant calls:
3992
3993 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3994 PCRE_INFO_NAMECOUNT yields the number of entries
3995 PCRE_INFO_NAMETABLE yields a pointer to the map.
3996
3997 The map is a vector of fixed-size entries. The size of each entry depends on
3998 the length of the longest name used. The first two bytes of each entry are the
3999 group number, most significant byte first. There follows the corresponding
4000 name, zero terminated. The names are in alphabetical order.
4001
4002 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4003 case instead of 255. Making it the same both with and without UTF-8 support
4004 means that the same test output works with both.
4005
4006 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4007 calling malloc() with a zero argument.
4008
4009 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4010 optimization. I've improved things by keeping a bitmap of backreferences with
4011 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4012 fact referenced, the optimization can be applied. It is unlikely that a
4013 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4014 the match to follow \n) will appear inside brackets with a number greater than
4015 31, but if it does, any back reference > 31 suppresses the optimization.
4016
4017 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4018 of disabling numbered capturing parentheses. Any opening parenthesis that is
4019 not followed by ? behaves as if it were followed by ?: but named parentheses
4020 can still be used for capturing (and they will acquire numbers in the usual
4021 way).
4022
4023 41. Redesigned the return codes from the match() function into yes/no/error so
4024 that errors can be passed back from deep inside the nested calls. A malloc
4025 failure while inside a recursive subpattern call now causes the
4026 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4027
4028 42. It is now possible to set a limit on the number of times the match()
4029 function is called in a call to pcre_exec(). This facility makes it possible to
4030 limit the amount of recursion and backtracking, though not in a directly
4031 obvious way, because the match() function is used in a number of different
4032 circumstances. The count starts from zero for each position in the subject
4033 string (for non-anchored patterns). The default limit is, for compatibility, a
4034 large number, namely 10 000 000. You can change this in two ways:
4035
4036 (a) When configuring PCRE before making, you can use --with-match-limit=n
4037 to set a default value for the compiled library.
4038
4039 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4040 a different value is set. See 45 below.
4041
4042 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4043
4044 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4045 of things that can be changed at compile time. The first argument specifies
4046 what is wanted and the second points to where the information is to be placed.
4047 The current list of available information is:
4048
4049 PCRE_CONFIG_UTF8
4050
4051 The output is an integer that is set to one if UTF-8 support is available;
4052 otherwise it is set to zero.
4053
4054 PCRE_CONFIG_NEWLINE
4055
4056 The output is an integer that it set to the value of the code that is used for
4057 newline. It is either LF (10) or CR (13).
4058
4059 PCRE_CONFIG_LINK_SIZE
4060
4061 The output is an integer that contains the number of bytes used for internal
4062 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4063
4064 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4065
4066 The output is an integer that contains the threshold above which the POSIX
4067 interface uses malloc() for output vectors. See item 31 above.
4068
4069 PCRE_CONFIG_MATCH_LIMIT
4070
4071 The output is an unsigned integer that contains the default limit of the number
4072 of match() calls in a pcre_exec() execution. See 42 above.
4073
4074 44. pcretest has been upgraded by the addition of the -C option. This causes it
4075 to extract all the available output from the new pcre_config() function, and to
4076 output it. The program then exits immediately.
4077
4078 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4079 order to support additional features. One way would have been to define
4080 pcre_exec2() (for example) with extra arguments, but this would not have been
4081 extensible, and would also have required all calls to the original function to
4082 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4083 is used for passing in "extra" data from pcre_study().
4084
4085 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4086 contains the following fields:
4087
4088 flags a bitmap indicating which of the following fields are set
4089 study_data opaque data from pcre_study()
4090 match_limit a way of specifying a limit on match() calls for a specific
4091 call to pcre_exec()
4092 callout_data data for callouts (see 49 below)
4093
4094 The flag bits are also defined in pcre.h, and are
4095
4096 PCRE_EXTRA_STUDY_DATA
4097 PCRE_EXTRA_MATCH_LIMIT
4098 PCRE_EXTRA_CALLOUT_DATA
4099
4100 The pcre_study() function now returns one of these new pcre_extra blocks, with
4101 the actual study data pointed to by the study_data field, and the
4102 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4103 before. That is, this change is entirely upwards-compatible and requires no
4104 change to existing code.
4105
4106 If you want to pass in additional data to pcre_exec(), you can either place it
4107 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4108 block.
4109
4110 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4111 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4112 times with different match limits, until it finds the minimum value needed for
4113 pcre_exec() to complete. The value is then output. This can be instructive; for
4114 most simple matches the number is quite small, but for pathological cases it
4115 gets very large very quickly.
4116
4117 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4118 returns the size of the data block pointed to by the study_data field in a
4119 pcre_extra block, that is, the value that was passed as the argument to
4120 pcre_malloc() when PCRE was getting memory in which to place the information
4121 created by pcre_study(). The fourth argument should point to a size_t variable.
4122 pcretest has been extended so that this information is shown after a successful
4123 pcre_study() call when information about the compiled regex is being displayed.
4124
4125 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4126 because what follows is always an absolute path. (Later: it turns out that this
4127 is more than cosmetic for MinGW, because it doesn't like empty path
4128 components.)
4129
4130 49. Some changes have been made to the callout feature (see 28 above):
4131
4132 (i) A callout function now has three choices for what it returns:
4133
4134 0 => success, carry on matching
4135 > 0 => failure at this point, but backtrack if possible
4136 < 0 => serious error, return this value from pcre_exec()
4137
4138 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4139 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4140 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4141 use by callout functions. It will never be used by PCRE itself.
4142
4143 (ii) The pcre_extra structure (see 45 above) has a void * field called
4144 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4145 pcre_callout_block structure has a field of the same name. The contents of
4146 the field passed in the pcre_extra structure are passed to the callout
4147 function in the corresponding field in the callout block. This makes it
4148 easier to use the same callout-containing regex from multiple threads. For
4149 testing, the pcretest program has a new data escape
4150
4151 \C*n pass the number n (may be negative) as callout_data
4152
4153 If the callout function in pcretest receives a non-zero value as
4154 callout_data, it returns that value.
4155
4156 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4157 there were some redundant $(CFLAGS) in commands that are now specified as
4158 $(LINK), which already includes $(CFLAGS).
4159
4160 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4161 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4162 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4163 one-byte characters throughout. Note that case-insensitive matching applies
4164 only to characters whose values are less than 256. PCRE doesn't support the
4165 notion of cases for higher-valued characters.
4166
4167 (i) A character class whose characters are all within 0-255 is handled as
4168 a bit map, and the map is inverted for negative classes. Previously, a
4169 character > 255 always failed to match such a class; however it should
4170 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4171
4172 (ii) A negated character class with a single character < 255 is coded as
4173 "not this character" (OP_NOT). This wasn't working properly when the test
4174 character was multibyte, either singly or repeated.
4175
4176 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4177 mode, for example: \x{100}{2,3}.
4178
4179 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4180 singly or repeated) now correctly test multibyte characters. However,
4181 PCRE doesn't recognize any characters with values greater than 255 as
4182 digits, spaces, or word characters. Such characters always match \D, \S,
4183 and \W, and never match \d, \s, or \w.
4184
4185 (v) Classes may now contain characters and character ranges with values
4186 greater than 255. For example: [ab\x{100}-\x{400}].
4187
4188 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4189 PCRE in UTF-8 mode.
4190
4191 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4192 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4193 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4194 value.)
4195
4196 53. The single man page has become too large. I have therefore split it up into
4197 a number of separate man pages. These also give rise to individual HTML pages;
4198 these are now put in a separate directory, and there is an index.html page that
4199 lists them all. Some hyperlinking between the pages has been installed.
4200
4201 54. Added convenience functions for handling named capturing parentheses.
4202
4203 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4204 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4205 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4206 are faulted.
4207
4208 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4209 calling configure. These values are used when compiling the dftables.c program
4210 which is run to generate the source of the default character tables. They
4211 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4212 you will need to set these values.
4213
4214 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4215
4216
4217 Version 3.9 02-Jan-02
4218 ---------------------
4219
4220 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4221
4222 2. If --disable-static was given, the building process failed when trying to
4223 build pcretest and pcregrep. (For some reason it was using libtool to compile
4224 them, which is not right, as they aren't part of the library.)
4225
4226
4227 Version 3.8 18-Dec-01
4228 ---------------------
4229
4230 1. The experimental UTF-8 code was completely screwed up. It was packing the
4231 bytes in the wrong order. How dumb can you get?
4232
4233
4234 Version 3.7 29-Oct-01
4235 ---------------------
4236
4237 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4238 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4239 this didn't happen under Solaris 8, where I normally test things.
4240
4241 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4242 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4243 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4244 this fix an hour or so after the initial 3.7 release.)
4245
4246
4247 Version 3.6 23-Oct-01
4248 ---------------------
4249
4250 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4251 offsets passed as NULL with zero offset count.
4252
4253 2. The config.guess and config.sub files had not been updated when I moved to
4254 the latest autoconf.
4255
4256
4257 Version 3.5 15-Aug-01
4258 ---------------------
4259
4260 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4261 had been forgotten.
4262
4263 2. By using declared but undefined structures, we can avoid using "void"
4264 definitions in pcre.h while keeping the internal definitions of the structures
4265 private.
4266
4267 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4268 user point of view, this means that both static and shared libraries are built
4269 by default, but this can be individually controlled. More of the work of
4270 handling this static/shared cases is now inside libtool instead of PCRE's make
4271 file.
4272
4273 4. The pcretest utility is now installed along with pcregrep because it is
4274 useful for users (to test regexs) and by doing this, it automatically gets
4275 relinked by libtool. The documentation has been turned into a man page, so
4276 there are now .1, .txt, and .html versions in /doc.
4277
4278 5. Upgrades to pcregrep:
4279 (i) Added long-form option names like gnu grep.
4280 (ii) Added --help to list all options with an explanatory phrase.
4281 (iii) Added -r, --recursive to recurse into sub-directories.
4282 (iv) Added -f, --file to read patterns from a file.
4283
4284 6. pcre_exec() was referring to its "code" argument before testing that
4285 argument for NULL (and giving an error if it was NULL).
4286
4287 7. Upgraded Makefile.in to allow for compiling in a different directory from
4288 the source directory.
4289
4290 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4291 options bits, the pointer it was passed was to an int instead of to an unsigned
4292 long int. This mattered only on 64-bit systems.
4293
4294 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4295 generated) instead of pcre.in, which it its source. Also made the same change
4296 in several of the .c files.
4297
4298 10. A new release of gcc defines printf() as a macro, which broke pcretest
4299 because it had an ifdef in the middle of a string argument for printf(). Fixed
4300 by using separate calls to printf().
4301
4302 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4303 script, to force use of CR or LF instead of \n in the source. On non-Unix
4304 systems, the value can be set in config.h.
4305
4306 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4307 absolute limit. Changed the text of the error message to make this clear, and
4308 likewise updated the man page.
4309
4310 13. The limit of 99 on the number of capturing subpatterns has been removed.
4311 The new limit is 65535, which I hope will not be a "real" limit.
4312
4313
4314 Version 3.4 22-Aug-00
4315 ---------------------
4316
4317 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4318
4319 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4320
4321
4322 Version 3.3 01-Aug-00
4323 ---------------------
4324
4325 1. If an octal character was given, but the value was greater than \377, it
4326 was not getting masked to the least significant bits, as documented. This could
4327 lead to crashes in some systems.
4328
4329 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4330 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4331
4332 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4333 These just pass their arguments on to (pcre_free)(), but they are provided
4334 because some uses of PCRE bind it to non-C systems that can call its functions,
4335 but cannot call free() or pcre_free() directly.
4336
4337 4. Add "make test" as a synonym for "make check". Corrected some comments in
4338 the Makefile.
4339
4340 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4341 Makefile.
4342
4343 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4344 command called pgrep for grepping around the active processes.
4345
4346 7. Added the beginnings of support for UTF-8 character strings.
4347
4348 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4349 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4350 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4351 out for the ar command.)
4352
4353
4354 Version 3.2 12-May-00
4355 ---------------------
4356
4357 This is purely a bug fixing release.
4358
4359 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4360 of ZA. This was just one example of several cases that could provoke this bug,
4361 which was introduced by change 9 of version 2.00. The code for breaking
4362 infinite loops after an iteration that matches an empty string was't working
4363 correctly.
4364
4365 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4366 when matched against abbab (for example). After matching an empty string, it
4367 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4368 caused it to match further down the string than it should.
4369
4370 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4371 was there because it doesn't seem to be needed, and it causes trouble on some
4372 systems, as it is not a Standard C header. It has been removed.
4373
4374 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4375 were reported on the Macintosh. The changes were from
4376
4377 while ((c = *(++ptr)) != 0 && c != '\n');
4378 to
4379 while ((c = *(++ptr)) != 0 && c != '\n') ;
4380
4381 Totally extraordinary, but if that's what it takes...
4382
4383 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4384 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4385 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4386 assumes the way PCRE uses memmove() (always moving upwards).
4387
4388 6. PCRE is being used in one environment where strchr() is not available. There
4389 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4390 faster code anyway.
4391
4392
4393 Version 3.1 09-Feb-00
4394 ---------------------
4395
4396 The only change in this release is the fixing of some bugs in Makefile.in for
4397 the "install" target:
4398
4399 (1) It was failing to install pcreposix.h.
4400
4401 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4402
4403
4404 Version 3.0 01-Feb-00
4405 ---------------------
4406
4407 1. Add support for the /+ modifier to perltest (to output $` like it does in
4408 pcretest).
4409
4410 2. Add support for the /g modifier to perltest.
4411
4412 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4413 matches null strings.
4414
4415 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4416 pattern. Perl treats empty patterns specially - it reuses the most recent
4417 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4418 effect.
4419
4420 5. The POSIX interface was broken in that it was just handing over the POSIX
4421 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4422 required a bigger vector, with some working space on the end. This means that
4423 the POSIX wrapper now has to get and free some memory, and copy the results.
4424
4425 6. Added some simple autoconf support, placing the test data and the
4426 documentation in separate directories, re-organizing some of the
4427 information files, and making it build pcre-config (a GNU standard). Also added
4428 libtool support for building PCRE as a shared library, which is now the
4429 default.
4430
4431 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4432 09 are not valid octal constants. Single digits will be used for minor values
4433 less than 10.
4434
4435 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4436 existing programs that set these in the POSIX interface can use PCRE without
4437 modification.
4438
4439 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4440 return all that pcre_info() returns, plus additional data. The pcre_info()
4441 function is retained for compatibility, but is considered to be obsolete.
4442
4443 10. Added experimental recursion feature (?R) to handle one common case that
4444 Perl 5.6 will be able to do with (?p{...}).
4445
4446 11. Added support for POSIX character classes like [:alpha:], which Perl is
4447 adopting.
4448
4449
4450 Version 2.08 31-Aug-99
4451 ----------------------
4452
4453 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4454 trying to match at the startoffset position, but instead was moving forward to
4455 the next newline as if a previous match had failed.
4456
4457 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4458 and could get into a loop if a null string was matched other than at the start
4459 of the subject.
4460
4461 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4462 be distinguished at compile time, and for completeness also added PCRE_DATE.
4463
4464 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4465 in GnuWin32 environments.
4466
4467
4468 Version 2.07 29-Jul-99
4469 ----------------------
4470
4471 1. The documentation is now supplied in plain text form and HTML as well as in
4472 the form of man page sources.
4473
4474 2. C++ compilers don't like assigning (void *) values to other pointer types.
4475 In particular this affects malloc(). Although there is no problem in Standard
4476 C, I've put in casts to keep C++ compilers happy.
4477
4478 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4479 should be (const char *).
4480
4481 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4482 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4483 However, I haven't made this a standard facility. The documentation doesn't
4484 mention it, and the Makefile doesn't support it.
4485
4486 5. The Makefile now contains an "install" target, with editable destinations at
4487 the top of the file. The pcretest program is not installed.
4488
4489 6. pgrep -V now gives the PCRE version number and date.
4490
4491 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4492 causing the entire string to be ignored, instead of just the last character.
4493
4494 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4495 non-matching string, it can take a very, very long time, even for strings of
4496 quite modest length, because of the nested recursion. PCRE now does better in
4497 some of these cases. It does this by remembering the last required literal
4498 character in the pattern, and pre-searching the subject to ensure it is present
4499 before running the real match. In other words, it applies a heuristic to detect
4500 some types of certain failure quickly, and in the above example, if presented
4501 with a string that has no trailing " it gives "no match" very quickly.
4502
4503 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4504 other alternatives are tried instead.
4505
4506
4507 Version 2.06 09-Jun-99
4508 ----------------------
4509
4510 1. Change pcretest's output for amount of store used to show just the code
4511 space, because the remainder (the data block) varies in size between 32-bit and
4512 64-bit systems.
4513
4514 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4515 start matching at. This allows lookbehinds to work when searching for multiple
4516 occurrences in a string.
4517
4518 3. Added additional options to pcretest for testing multiple occurrences:
4519
4520 /+ outputs the rest of the string that follows a match
4521 /g loops for multiple occurrences, using the new startoffset argument
4522 /G loops for multiple occurrences by passing an incremented pointer
4523
4524 4. PCRE wasn't doing the "first character" optimization for patterns starting
4525 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4526 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4527 the letter 'x'. On long subject strings, this gives a significant speed-up.
4528
4529
4530 Version 2.05 21-Apr-99
4531 ----------------------
4532
4533 1. Changed the type of magic_number from int to long int so that it works
4534 properly on 16-bit systems.
4535
4536 2. Fixed a bug which caused patterns starting with .* not to work correctly
4537 when the subject string contained newline characters. PCRE was assuming
4538 anchoring for such patterns in all cases, which is not correct because .* will
4539 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4540 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4541 must be retried after every newline in the subject.
4542
4543
4544 Version 2.04 18-Feb-99
4545 ----------------------
4546
4547 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4548 computation of the store needed to hold the pattern was incorrect (too large).
4549 If such patterns were nested a few deep, this could multiply and become a real
4550 problem.
4551
4552 2. Added /M option to pcretest to show the memory requirement of a specific
4553 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4554
4555 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4556 compiled in such a way that the backtracking after subsequent failure was
4557 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4558 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4559
4560
4561 Version 2.03 02-Feb-99
4562 ----------------------
4563
4564 1. Fixed typo and small mistake in man page.
4565
4566 2. Added 4th condition (GPL supersedes if conflict) and created separate
4567 LICENCE file containing the conditions.
4568
4569 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4570 Perl, that is the internal \ allows the delimiter to be included in the
4571 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4572 the final delimiter, add \ to the end of the pattern (to test the error).
4573
4574 4. Added the convenience functions for extracting substrings after a successful
4575 match. Updated pcretest to make it able to test these functions.
4576
4577
4578 Version 2.02 14-Jan-99
4579 ----------------------
4580
4581 1. Initialized the working variables associated with each extraction so that
4582 their saving and restoring doesn't refer to uninitialized store.
4583
4584 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4585 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4586 fix the problem.
4587
4588 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4589 calls, and wasn't printing the correct value for compiling calls. Increased the
4590 default value of LOOPREPEAT, and the number of significant figures in the
4591 times.
4592
4593 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4594
4595 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4596 a building problem on Windows NT with a FAT file system.
4597
4598
4599 Version 2.01 21-Oct-98
4600 ----------------------
4601
4602 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4603 to character tables built by pcre_maketables() in the current locale. If NULL
4604 is passed, the default tables are used.
4605
4606
4607 Version 2.00 24-Sep-98
4608 ----------------------
4609
4610 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4611 it any more.
4612
4613 2. Allow quantification of (?>) groups, and make it work correctly.
4614
4615 3. The first character computation wasn't working for (?>) groups.
4616
4617 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4618 end of the subject) and add 5.005's \z, which really does match only at the
4619 very end of the subject.
4620
4621 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4622
4623 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4624 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4625 localized options. All options to pcre_study() were also removed.
4626
4627 7. Add other new features from 5.005:
4628
4629 $(?<= positive lookbehind
4630 $(?<! negative lookbehind
4631 (?imsx-imsx) added the unsetting capability
4632 such a setting is global if at outer level; local otherwise
4633 (?imsx-imsx:) non-capturing groups with option setting
4634 (?(cond)re|re) conditional pattern matching
4635
4636 A backreference to itself in a repeated group matches the previous
4637 captured string.
4638
4639 8. General tidying up of studying (both automatic and via "study")
4640 consequential on the addition of new assertions.
4641
4642 9. As in 5.005, unlimited repeated groups that could match an empty substring
4643 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4644 runtime if any iteration does actually match an empty substring.
4645
4646 10. Include the RunTest script in the distribution.
4647
4648 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4649 discrepancies, some of which were old and were also with respect to 5.004. They
4650 have now been fixed.
4651
4652
4653 Version 1.09 28-Apr-98
4654 ----------------------
4655
4656 1. A negated single character class followed by a quantifier with a minimum
4657 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4658 program crashes, or just wrong answers. This did not apply to negated classes
4659 containing more than one character, or to minima other than one.
4660
4661
4662 Version 1.08 27-Mar-98
4663 ----------------------
4664
4665 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4666
4667 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4668 latter must appear before anything that relies on it in the pattern.
4669
4670
4671 Version 1.07 16-Feb-98
4672 ----------------------
4673
4674 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4675 repeat of a potentially empty string).
4676
4677
4678 Version 1.06 23-Jan-98
4679 ----------------------
4680
4681 1. Added Markus Oberhumer's little patches for C++.
4682
4683 2. Literal strings longer than 255 characters were broken.
4684
4685
4686 Version 1.05 23-Dec-97
4687 ----------------------
4688
4689 1. Negated character classes containing more than one character were failing if
4690 PCRE_CASELESS was set at run time.
4691
4692
4693 Version 1.04 19-Dec-97
4694 ----------------------
4695
4696 1. Corrected the man page, where some "const" qualifiers had been omitted.
4697
4698 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4699 input syntax.
4700
4701 3. Fixed memory leak which occurred when a regex with back references was
4702 matched with an offsets vector that wasn't big enough. The temporary memory
4703 that is used in this case wasn't being freed if the match failed.
4704
4705 4. Tidied pcretest to ensure it frees memory that it gets.
4706
4707 5. Temporary memory was being obtained in the case where the passed offsets
4708 vector was exactly big enough.
4709
4710 6. Corrected definition of offsetof() from change 5 below.
4711
4712 7. I had screwed up change 6 below and broken the rules for the use of
4713 setjmp(). Now fixed.
4714
4715
4716 Version 1.03 18-Dec-97
4717 ----------------------
4718
4719 1. A erroneous regex with a missing opening parenthesis was correctly
4720 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4721 on some systems.
4722
4723 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4724 it was reported that one broken compiler failed on the former because "code" is
4725 also an independent variable.
4726
4727 3. The erroneous regex a[]b caused an array overrun reference.
4728
4729 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4730 fail on data ending with that character. (It was going on too far, and checking
4731 the next character, typically a binary zero.) This was specific to the
4732 optimized code for single-character negative classes.
4733
4734 5. Added a contributed patch from the TIN world which does the following:
4735
4736 + Add an undef for memmove, in case the the system defines a macro for it.
4737
4738 + Add a definition of offsetof(), in case there isn't one. (I don't know
4739 the reason behind this - offsetof() is part of the ANSI standard - but
4740 it does no harm).
4741
4742 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4743 most of the places where whitespace preceded '#'. I have given up and
4744 allowed the remaining 2 cases to be at the margin.
4745
4746 + Rename some variables in pcre to eliminate shadowing. This seems very
4747 pedantic, but does no harm, of course.
4748
4749 6. Moved the call to setjmp() into its own function, to get rid of warnings
4750 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4751
4752 7. Constructs such as \d{8,} were compiling into the equivalent of
4753 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4754 outcome, but in this particular case used more store than had been allocated,
4755 which caused the bug to be discovered because it threw up an internal error.
4756
4757 8. The debugging code in both pcre and pcretest for outputting the compiled
4758 form of a regex was going wrong in the case of back references followed by
4759 curly-bracketed repeats.
4760
4761
4762 Version 1.02 12-Dec-97
4763 ----------------------
4764
4765 1. Typos in pcre.3 and comments in the source fixed.
4766
4767 2. Applied a contributed patch to get rid of places where it used to remove
4768 'const' from variables, and fixed some signed/unsigned and uninitialized
4769 variable warnings.
4770
4771 3. Added the "runtest" target to Makefile.
4772
4773 4. Set default compiler flag to -O2 rather than just -O.
4774
4775
4776 Version 1.01 19-Nov-97
4777 ----------------------
4778
4779 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4780 like /([ab]*)*/, that is, for classes with more than one character in them.
4781
4782 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4783 as /((?>a*))*/ (a PCRE_EXTRA facility).
4784
4785
4786 Version 1.00 18-Nov-97
4787 ----------------------
4788
4789 1. Added compile-time macros to support systems such as SunOS4 which don't have
4790 memmove() or strerror() but have other things that can be used instead.
4791
4792 2. Arranged that "make clean" removes the executables.
4793
4794
4795 Version 0.99 27-Oct-97
4796 ----------------------
4797
4798 1. Fixed bug in code for optimizing classes with only one character. It was
4799 initializing a 32-byte map regardless, which could cause it to run off the end
4800 of the memory it had got.
4801
4802 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4803
4804
4805 Version 0.98 22-Oct-97
4806 ----------------------
4807
4808 1. Fixed bug in code for handling temporary memory usage when there are more
4809 back references than supplied space in the ovector. This could cause segfaults.
4810
4811
4812 Version 0.97 21-Oct-97
4813 ----------------------
4814
4815 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4816
4817 2. Optimized negated single characters not to use a bit map.
4818
4819 3. Brought error texts together as macro definitions; clarified some of them;
4820 fixed one that was wrong - it said "range out of order" when it meant "invalid
4821 escape sequence".
4822
4823 4. Changed some char * arguments to const char *.
4824
4825 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4826
4827 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4828 pcretest.
4829
4830
4831 Version 0.96 16-Oct-97
4832 ----------------------
4833
4834 1. Added a simple "pgrep" utility to the distribution.
4835
4836 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4837 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4838 where "ddd" means "one or more decimal digits".
4839
4840 3. Fixed serious bug. If a pattern had a back reference, but the call to
4841 pcre_exec() didn't supply a large enough ovector to record the related
4842 identifying subpattern, the match always failed. PCRE now remembers the number
4843 of the largest back reference, and gets some temporary memory in which to save
4844 the offsets during matching if necessary, in order to ensure that
4845 backreferences always work.
4846
4847 4. Increased the compatibility with Perl in a number of ways:
4848
4849 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4850 to request this handling. The option can be set at compile or exec time.
4851
4852 (b) $ matches before a terminating newline by default; an option
4853 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4854 mode). The option can be set at compile or exec time.
4855
4856 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4857 the same as Perl's. If the decimal number it represents is less than 10
4858 or there aren't that many previous left capturing parentheses, an octal
4859 escape is read. Inside a character class, it's always an octal escape,
4860 even if it is a single digit.
4861
4862 (d) An escaped but undefined alphabetic character is taken as a literal,
4863 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4864 escapes.
4865
4866 (e) {0} is now permitted. (The previous item is removed from the compiled
4867 pattern).
4868
4869 5. Changed all the names of code files so that the basic parts are no longer
4870 than 10 characters, and abolished the teeny "globals.c" file.
4871
4872 6. Changed the handling of character classes; they are now done with a 32-byte
4873 bit map always.
4874
4875 7. Added the -d and /D options to pcretest to make it possible to look at the
4876 internals of compilation without having to recompile pcre.
4877
4878
4879 Version 0.95 23-Sep-97
4880 ----------------------
4881
4882 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4883 \x20 at the start of a run of normal characters. These were being treated as
4884 real characters, instead of the source characters being re-checked.
4885
4886
4887 Version 0.94 18-Sep-97
4888 ----------------------
4889
4890 1. The functions are now thread-safe, with the caveat that the global variables
4891 containing pointers to malloc() and free() or alternative functions are the
4892 same for all threads.
4893
4894 2. Get pcre_study() to generate a bitmap of initial characters for non-
4895 anchored patterns when this is possible, and use it if passed to pcre_exec().
4896
4897
4898 Version 0.93 15-Sep-97
4899 ----------------------
4900
4901 1. /(b)|(:+)/ was computing an incorrect first character.
4902
4903 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4904 but not actually doing anything yet.
4905
4906 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4907 as Perl does (e.g. [-az] or [az-]).
4908
4909 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4910 all possible positions.
4911
4912 5. Split up into different modules to avoid including unneeded functions in a
4913 compiled binary. However, compile and exec are still in one module. The "study"
4914 function is split off.
4915
4916 6. The character tables are now in a separate module whose source is generated
4917 by an auxiliary program - but can then be edited by hand if required. There are
4918 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4919 toupper() in the code.
4920
4921 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4922 make them global. Abolish the function for setting them, as the caller can now
4923 set them directly.
4924
4925
4926 Version 0.92 11-Sep-97
4927 ----------------------
4928
4929 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4930 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4931
4932 2. Caseless matching was not working in character classes if the characters in
4933 the pattern were in upper case.
4934
4935 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4936
4937 4. Make PCRE_ANCHORED public and accept as a compile option.
4938
4939 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4940 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4941 pass them.
4942
4943 6. Give an error if bad option bits passed at compile or run time.
4944
4945 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4946 pcretest to cause it to pass that flag.
4947
4948 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4949 options, and the first character, if set.
4950
4951 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4952
4953
4954 Version 0.91 10-Sep-97
4955 ----------------------
4956
4957 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4958 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4959
4960 2. PCRE was looping on encountering an indefinitely repeated back reference to
4961 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4962 Perl does - treats the match as successful.
4963
4964 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5