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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1172 - (show annotations)
Wed Oct 24 16:41:07 2012 UTC (6 years, 11 months ago) by ph10
File size: 238196 byte(s)
Error occurred while calculating annotation data.
Create the *.generic files using "make" instead of PrepareRelease.
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
133 Version 8.31 06-July-2012
134 -------------------------
135
136 1. Fixing a wrong JIT test case and some compiler warnings.
137
138 2. Removed a bashism from the RunTest script.
139
140 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
141 to unsigned type, result still unsigned" that was given by an MS compiler
142 on encountering the code "-sizeof(xxx)".
143
144 4. Partial matching support is added to the JIT compiler.
145
146 5. Fixed several bugs concerned with partial matching of items that consist
147 of more than one character:
148
149 (a) /^(..)\1/ did not partially match "aba" because checking references was
150 done on an "all or nothing" basis. This also applied to repeated
151 references.
152
153 (b) \R did not give a hard partial match if \r was found at the end of the
154 subject.
155
156 (c) \X did not give a hard partial match after matching one or more
157 characters at the end of the subject.
158
159 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
160 a partial match for the string "\r".
161
162 (e) When newline was set to CRLF, the metacharacter "." did not recognize
163 a partial match for a CR character at the end of the subject string.
164
165 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
166 running pcretest, the text "(JIT)" added to the output whenever JIT is
167 actually used to run the match.
168
169 7. Individual JIT compile options can be set in pcretest by following -s+[+]
170 or /S+[+] with a digit between 1 and 7.
171
172 8. OP_NOT now supports any UTF character not just single-byte ones.
173
174 9. (*MARK) control verb is now supported by the JIT compiler.
175
176 10. The command "./RunTest list" lists the available tests without actually
177 running any of them. (Because I keep forgetting what they all are.)
178
179 11. Add PCRE_INFO_MAXLOOKBEHIND.
180
181 12. Applied a (slightly modified) user-supplied patch that improves performance
182 when the heap is used for recursion (compiled with --disable-stack-for-
183 recursion). Instead of malloc and free for each heap frame each time a
184 logical recursion happens, frames are retained on a chain and re-used where
185 possible. This sometimes gives as much as 30% improvement.
186
187 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
188 call.
189
190 14. As documented, (*COMMIT) is now confined to within a positive assertion.
191
192 15. It is now possible to link pcretest with libedit as an alternative to
193 libreadline.
194
195 16. (*COMMIT) control verb is now supported by the JIT compiler.
196
197 17. The Unicode data tables have been updated to Unicode 6.1.0.
198
199 18. Added --file-list option to pcregrep.
200
201 19. Added binary file support to pcregrep, including the -a, --binary-files,
202 -I, and --text options.
203
204 20. The madvise function is renamed for posix_madvise for QNX compatibility
205 reasons. Fixed by Giuseppe D'Angelo.
206
207 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
208
209 22. Changed the option for creating symbolic links for 16-bit man pages from
210 -s to -sf so that re-installing does not cause issues.
211
212 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
213
214 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
215 matches in certain environments (the workspace was not being correctly
216 retained). Also added to pcre_dfa_exec() a simple plausibility check on
217 some of the workspace data at the beginning of a restart.
218
219 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
220 was not doing so when it should - probably a typo introduced by SVN 528
221 (change 8.10/14).
222
223 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
224 \w+ when the character tables indicated that \x{c4} was a word character.
225 There were several related cases, all because the tests for doing a table
226 lookup were testing for characters less than 127 instead of 255.
227
228 27. If a pattern contains capturing parentheses that are not used in a match,
229 their slots in the ovector are set to -1. For those that are higher than
230 any matched groups, this happens at the end of processing. In the case when
231 there were back references that the ovector was too small to contain
232 (causing temporary malloc'd memory to be used during matching), and the
233 highest capturing number was not used, memory off the end of the ovector
234 was incorrectly being set to -1. (It was using the size of the temporary
235 memory instead of the true size.)
236
237 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
238 ovector size, it uses memory at the end of the block that it has got.
239
240 29. Check for an overlong MARK name and give an error at compile time. The
241 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
242
243 30. JIT compiler update.
244
245 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
246 Rill for the patch.
247
248 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
249
250 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
251
252 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
253 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
254 was enabled.
255
256 35. Improve JIT code generation for greedy plus quantifier.
257
258 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
259 1 to "aa" instead of to an empty string. The bug affected repeated groups
260 that could potentially match an empty string.
261
262 37. Optimizing single character iterators in JIT.
263
264 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
265 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
266 codepoints that are too big for the mode are faulted, and in a UTF mode,
267 disallowed codepoints are also faulted.
268
269 39. If PCRE was compiled with UTF support, in three places in the DFA
270 matcher there was code that should only have been obeyed in UTF mode, but
271 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
272 processing when bytes with values greater than 127 were present. In 16-bit
273 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
274 both cases the values are those that cannot be the first data item in a UTF
275 character. The three items that might have provoked this were recursions,
276 possessively repeated groups, and atomic groups.
277
278 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
279 and pcregrep, because some OS require shared objects to be explicitly
280 passed to ld, causing the link step to fail if they are not.
281
282 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
283 mode, patterns that started with \h* or \R* might be incorrectly matched.
284
285
286 Version 8.30 04-February-2012
287 -----------------------------
288
289 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
290 name is defined in ctype.h.
291
292 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
293 only in quite long subpatterns.
294
295 3. Removed the function pcre_info(), which has been obsolete and deprecated
296 since it was replaced by pcre_fullinfo() in February 2000.
297
298 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
299 match a (*MARK), and the match failed at the start of the subject, a
300 reference to memory before the start of the subject could occur. This bug
301 was introduced by fix 17 of release 8.21.
302
303 5. A reference to an unset group with zero minimum repetition was giving
304 totally wrong answers (in non-JavaScript-compatibility mode). For example,
305 /(another)?(\1?)test/ matched against "hello world test". This bug was
306 introduced in release 8.13.
307
308 6. Add support for 16-bit character strings (a large amount of work involving
309 many changes and refactorings).
310
311 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
312 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
313 from a file.
314
315 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
316 rounding is not applied in this particular case).
317
318 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
319 if they appear, or are escaped, in patterns.
320
321 10. Get rid of a number of -Wunused-but-set-variable warnings.
322
323 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
324 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
325 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
326 also returns the mark "x". This bug applied to capturing parentheses,
327 non-capturing parentheses, and atomic parentheses. It also applied to some
328 assertions.
329
330 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
331 information out of configure.ac instead of relying on pcre.h.generic, which
332 is not stored in the repository.
333
334 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
335 -lreadline.
336
337 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
338
339 15. Applied Graycode's patch to put the top-level frame on the stack rather
340 than the heap when not using the stack for recursion. This gives a
341 performance improvement in many cases when recursion is not deep.
342
343 16. Experimental code added to "pcretest -C" to output the stack frame size.
344
345
346 Version 8.21 12-Dec-2011
347 ------------------------
348
349 1. Updating the JIT compiler.
350
351 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
352 are added as well.
353
354 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
355 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
356 calling _pcre_jit_exec. Some extra comments are added.
357
358 4. (*MARK) settings inside atomic groups that do not contain any capturing
359 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
360 was introduced by change 18 for 8.20.
361
362 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
363 ECMA-262 standard.
364
365 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
366 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
367 This bug was probably introduced by change 9 of 8.13.
368
369 7. While fixing 6 above, I noticed that a number of other items were being
370 incorrectly rejected as "not fixed length". This arose partly because newer
371 opcodes had not been added to the fixed-length checking code. I have (a)
372 corrected the bug and added tests for these items, and (b) arranged for an
373 error to occur if an unknown opcode is encountered while checking for fixed
374 length instead of just assuming "not fixed length". The items that were
375 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
376 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
377 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
378
379 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
380 being incorrectly compiled and would have given unpredicatble results.
381
382 9. A possessively repeated subpattern with minimum repeat count greater than
383 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
384 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
385 the first (A) could occur when it should not.
386
387 10. Add a cast and remove a redundant test from the code.
388
389 11. JIT should use pcre_malloc/pcre_free for allocation.
390
391 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
392 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
393 is anything other than /usr, -L is still shown).
394
395 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
396
397 14. Perl does not support \N without a following name in a [] class; PCRE now
398 also gives an error.
399
400 15. If a forward reference was repeated with an upper limit of around 2000,
401 it caused the error "internal error: overran compiling workspace". The
402 maximum number of forward references (including repeats) was limited by the
403 internal workspace, and dependent on the LINK_SIZE. The code has been
404 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
405 the default depends on LINK_SIZE. There is a new upper limit (for safety)
406 of around 200,000 forward references. While doing this, I also speeded up
407 the filling in of repeated forward references.
408
409 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
410 incorrectly expecting the subject to contain another "a" after the start.
411
412 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
413 in the match, the SKIP should be ignored. This was not happening; instead
414 the SKIP was being treated as NOMATCH. For patterns such as
415 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
416 tested.
417
418 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
419 now much more compatible with Perl, in particular in cases where the result
420 is a non-match for a non-anchored pattern. For example, if
421 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
422 "m", where previously it did not return a name. A side effect of this
423 change is that for partial matches, the last encountered mark name is
424 returned, as for non matches. A number of tests that were previously not
425 Perl-compatible have been moved into the Perl-compatible test files. The
426 refactoring has had the pleasing side effect of removing one argument from
427 the match() function, thus reducing its stack requirements.
428
429 19. If the /S+ option was used in pcretest to study a pattern using JIT,
430 subsequent uses of /S (without +) incorrectly behaved like /S+.
431
432 21. Retrieve executable code size support for the JIT compiler and fixing
433 some warnings.
434
435 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
436 not work when the shorter character appeared right at the end of the
437 subject string.
438
439 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
440 systems.
441
442 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
443 output it when the /M option is used in pcretest.
444
445 25. The CheckMan script was not being included in the distribution. Also, added
446 an explicit "perl" to run Perl scripts from the PrepareRelease script
447 because this is reportedly needed in Windows.
448
449 26. If study data was being save in a file and studying had not found a set of
450 "starts with" bytes for the pattern, the data written to the file (though
451 never used) was taken from uninitialized memory and so caused valgrind to
452 complain.
453
454 27. Updated RunTest.bat as provided by Sheri Pierce.
455
456 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
457
458 29. Computation of memory usage for the table of capturing group names was
459 giving an unnecessarily large value.
460
461
462 Version 8.20 21-Oct-2011
463 ------------------------
464
465 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
466 a POSIX class. After further experiments with Perl, which convinced me that
467 Perl has bugs and confusions, a closing square bracket is no longer allowed
468 in a POSIX name. This bug also affected patterns with classes that started
469 with full stops.
470
471 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
472 captured substring, but while checking the failing first alternative,
473 substring 1 is temporarily captured. If the output vector supplied to
474 pcre_exec() was not big enough for this capture, the yield of the function
475 was still zero ("insufficient space for captured substrings"). This cannot
476 be totally fixed without adding another stack variable, which seems a lot
477 of expense for a edge case. However, I have improved the situation in cases
478 such as /(a)(b)x|abc/ matched against "abc", where the return code
479 indicates that fewer than the maximum number of slots in the ovector have
480 been set.
481
482 3. Related to (2) above: when there are more back references in a pattern than
483 slots in the output vector, pcre_exec() uses temporary memory during
484 matching, and copies in the captures as far as possible afterwards. It was
485 using the entire output vector, but this conflicts with the specification
486 that only 2/3 is used for passing back captured substrings. Now it uses
487 only the first 2/3, for compatibility. This is, of course, another edge
488 case.
489
490 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
491 main code base, and can be used by building with --enable-jit. When this is
492 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
493 runtime --no-jit option is given.
494
495 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
496 ovector, the return from the function was zero, implying that there were
497 other matches that did not fit. The correct "exactly full" value is now
498 returned.
499
500 6. If a subpattern that was called recursively or as a subroutine contained
501 (*PRUNE) or any other control that caused it to give a non-standard return,
502 invalid errors such as "Error -26 (nested recursion at the same subject
503 position)" or even infinite loops could occur.
504
505 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
506 computing the minimum length on reaching *ACCEPT, and so ended up with the
507 wrong value of 1 rather than 0. Further investigation indicates that
508 computing a minimum subject length in the presence of *ACCEPT is difficult
509 (think back references, subroutine calls), and so I have changed the code
510 so that no minimum is registered for a pattern that contains *ACCEPT.
511
512 8. If (*THEN) was present in the first (true) branch of a conditional group,
513 it was not handled as intended. [But see 16 below.]
514
515 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
516 Sheri Pierce.
517
518 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
519 the first byte in a match must be "a".
520
521 11. Change 17 for 8.13 increased the recursion depth for patterns like
522 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
523 pattern contains any instances of (*THEN). If it does not, the old
524 optimizations are restored. It would be nice to do this on a per-group
525 basis, but at the moment that is not feasible.
526
527 12. In some environments, the output of pcretest -C is CRLF terminated. This
528 broke RunTest's code that checks for the link size. A single white space
529 character after the value is now allowed for.
530
531 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
532 For "fr", it uses the Windows-specific input and output files.
533
534 14. If (*THEN) appeared in a group that was called recursively or as a
535 subroutine, it did not work as intended. [But see next item.]
536
537 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
538 pattern fragments (but not containing any | characters). If A and B are
539 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
540 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
541 D. In other words, Perl considers parentheses that do not contain any |
542 characters to be part of a surrounding alternative, whereas PCRE was
543 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
544 differently. PCRE now behaves in the same way as Perl, except in the case
545 of subroutine/recursion calls such as (?1) which have in any case always
546 been different (but PCRE had them first :-).
547
548 16. Related to 15 above: Perl does not treat the | in a conditional group as
549 creating alternatives. Such a group is treated in the same way as an
550 ordinary group without any | characters when processing (*THEN). PCRE has
551 been changed to match Perl's behaviour.
552
553 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
554 RunGrepTest script failed.
555
556 18. Change 22 for version 13 caused atomic groups to use more stack. This is
557 inevitable for groups that contain captures, but it can lead to a lot of
558 stack use in large patterns. The old behaviour has been restored for atomic
559 groups that do not contain any capturing parentheses.
560
561 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
562 suppress the check for a minimum subject length at run time. (If it was
563 given to pcre_exec() or pcre_dfa_exec() it did work.)
564
565 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
566 fail to work when decoding hex characters in data strings in EBCDIC
567 environments.
568
569 21. It appears that in at least one Mac OS environment, the isxdigit() function
570 is implemented as a macro that evaluates to its argument more than once,
571 contravening the C 90 Standard (I haven't checked a later standard). There
572 was an instance in pcretest which caused it to go wrong when processing
573 \x{...} escapes in subject strings. The has been rewritten to avoid using
574 things like p++ in the argument of isxdigit().
575
576
577 Version 8.13 16-Aug-2011
578 ------------------------
579
580 1. The Unicode data tables have been updated to Unicode 6.0.0.
581
582 2. Two minor typos in pcre_internal.h have been fixed.
583
584 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
585 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
586 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
587
588 4. There were a number of related bugs in the code for matching backrefences
589 caselessly in UTF-8 mode when codes for the characters concerned were
590 different numbers of bytes. For example, U+023A and U+2C65 are an upper
591 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
592 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
593 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
594 2-byte code at the end of the subject (it thought there wasn't enough data
595 left).
596
597 5. Comprehensive information about what went wrong is now returned by
598 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
599 as the output vector has at least 2 elements. The offset of the start of
600 the failing character and a reason code are placed in the vector.
601
602 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
603 now returned is for the first byte of the failing character, instead of the
604 last byte inspected. This is an incompatible change, but I hope it is small
605 enough not to be a problem. It makes the returned offset consistent with
606 pcre_exec() and pcre_dfa_exec().
607
608 7. pcretest now gives a text phrase as well as the error number when
609 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
610 failure, the offset and reason code are output.
611
612 8. When \R was used with a maximizing quantifier it failed to skip backwards
613 over a \r\n pair if the subsequent match failed. Instead, it just skipped
614 back over a single character (\n). This seems wrong (because it treated the
615 two characters as a single entity when going forwards), conflicts with the
616 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
617 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
618 has been changed.
619
620 9. Some internal refactoring has changed the processing so that the handling
621 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
622 time (the PCRE_DOTALL option was changed this way some time ago: version
623 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
624 which was always a bit of a fudge. It also means that there is one less
625 argument for the match() function, which reduces its stack requirements
626 slightly. This change also fixes an incompatibility with Perl: the pattern
627 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
628
629 10. More internal refactoring has drastically reduced the number of recursive
630 calls to match() for possessively repeated groups such as (abc)++ when
631 using pcre_exec().
632
633 11. While implementing 10, a number of bugs in the handling of groups were
634 discovered and fixed:
635
636 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
637 (a|)*(?1) gave a compile-time internal error.
638 ((a|)+)+ did not notice that the outer group could match an empty string.
639 (^a|^)+ was not marked as anchored.
640 (.*a|.*)+ was not marked as matching at start or after a newline.
641
642 12. Yet more internal refactoring has removed another argument from the match()
643 function. Special calls to this function are now indicated by setting a
644 value in a variable in the "match data" data block.
645
646 13. Be more explicit in pcre_study() instead of relying on "default" for
647 opcodes that mean there is no starting character; this means that when new
648 ones are added and accidentally left out of pcre_study(), testing should
649 pick them up.
650
651 14. The -s option of pcretest has been documented for ages as being an old
652 synonym of -m (show memory usage). I have changed it to mean "force study
653 for every regex", that is, assume /S for every regex. This is similar to -i
654 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
655 using it. It makes it easier to run collections of tests with and without
656 study enabled, and thereby test pcre_study() more easily. All the standard
657 tests are now run with and without -s (but some patterns can be marked as
658 "never study" - see 20 below).
659
660 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
661 restoration of the capturing data to the outer values was not happening
662 correctly.
663
664 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
665 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
666 pattern had matched an empty string, and so incorrectly returned a no
667 match.
668
669 17. There was optimizing code for the last branch of non-capturing parentheses,
670 and also for the obeyed branch of a conditional subexpression, which used
671 tail recursion to cut down on stack usage. Unfortunately, now that there is
672 the possibility of (*THEN) occurring in these branches, tail recursion is
673 no longer possible because the return has to be checked for (*THEN). These
674 two optimizations have therefore been removed. [But see 8.20/11 above.]
675
676 18. If a pattern containing \R was studied, it was assumed that \R always
677 matched two bytes, thus causing the minimum subject length to be
678 incorrectly computed because \R can also match just one byte.
679
680 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
681 was incorrectly computed.
682
683 20. If /S is present twice on a test pattern in pcretest input, it now
684 *disables* studying, thereby overriding the use of -s on the command line
685 (see 14 above). This is necessary for one or two tests to keep the output
686 identical in both cases.
687
688 21. When (*ACCEPT) was used in an assertion that matched an empty string and
689 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
690
691 22. When an atomic group that contained a capturing parenthesis was
692 successfully matched, but the branch in which it appeared failed, the
693 capturing was not being forgotten if a higher numbered group was later
694 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
695 group 1 to "a", when in fact it should be unset. This applied to multi-
696 branched capturing and non-capturing groups, repeated or not, and also to
697 positive assertions (capturing in negative assertions does not happen
698 in PCRE) and also to nested atomic groups.
699
700 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
701 subject after a captured substring, to make it easier to tell which of a
702 number of identical substrings has been captured.
703
704 24. The way atomic groups are processed by pcre_exec() has been changed so that
705 if they are repeated, backtracking one repetition now resets captured
706 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
707 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
708 "aaa". Previously, it would have been "a". As part of this code
709 refactoring, the way recursive calls are handled has also been changed.
710
711 25. If an assertion condition captured any substrings, they were not passed
712 back unless some other capturing happened later. For example, if
713 (?(?=(a))a) was matched against "a", no capturing was returned.
714
715 26. When studying a pattern that contained subroutine calls or assertions,
716 the code for finding the minimum length of a possible match was handling
717 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
718 group 1 called group 2 while simultaneously a separate group 2 called group
719 1). A stack overflow occurred in this case. I have fixed this by limiting
720 the recursion depth to 10.
721
722 27. Updated RunTest.bat in the distribution to the version supplied by Tom
723 Fortmann. This supports explicit test numbers on the command line, and has
724 argument validation and error reporting.
725
726 28. An instance of \X with an unlimited repeat could fail if at any point the
727 first character it looked at was a mark character.
728
729 29. Some minor code refactoring concerning Unicode properties and scripts
730 should reduce the stack requirement of match() slightly.
731
732 30. Added the '=' option to pcretest to check the setting of unused capturing
733 slots at the end of the pattern, which are documented as being -1, but are
734 not included in the return count.
735
736 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
737 compiled something random. Now it gives a compile-time error (as does
738 Perl).
739
740 32. A *MARK encountered during the processing of a positive assertion is now
741 recorded and passed back (compatible with Perl).
742
743 33. If --only-matching or --colour was set on a pcregrep call whose pattern
744 had alternative anchored branches, the search for a second match in a line
745 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
746 matched the line "0102" twice. The same bug affected patterns that started
747 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
748 twice.
749
750 34. Previously, PCRE did not allow quantification of assertions. However, Perl
751 does, and because of capturing effects, quantifying parenthesized
752 assertions may at times be useful. Quantifiers are now allowed for
753 parenthesized assertions.
754
755 35. A minor code tidy in pcre_compile() when checking options for \R usage.
756
757 36. \g was being checked for fancy things in a character class, when it should
758 just be a literal "g".
759
760 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
761 appearance of a nested POSIX class supersedes an apparent external class.
762 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
763 unescaped square brackets may also appear as part of class names. For
764 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
765 more like Perl. (But see 8.20/1 above.)
766
767 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
768 was because it thought it was \N{name}, which is not supported).
769
770 39. Add minix to OS list not supporting the -S option in pcretest.
771
772 40. PCRE tries to detect cases of infinite recursion at compile time, but it
773 cannot analyze patterns in sufficient detail to catch mutual recursions
774 such as ((?1))((?2)). There is now a runtime test that gives an error if a
775 subgroup is called recursively as a subpattern for a second time at the
776 same position in the subject string. In previous releases this might have
777 been caught by the recursion limit, or it might have run out of stack.
778
779 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
780 happen only once. PCRE was, however incorrectly giving a compile time error
781 "recursive call could loop indefinitely" because it cannot analyze the
782 pattern in sufficient detail. The compile time test no longer happens when
783 PCRE is compiling a conditional subpattern, but actual runaway loops are
784 now caught at runtime (see 40 above).
785
786 42. It seems that Perl allows any characters other than a closing parenthesis
787 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
788 has been changed to be the same.
789
790 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
791 as not to get warnings when autogen.sh is called. Also changed
792 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
793
794 44. To help people who use pcregrep to scan files containing exceedingly long
795 lines, the following changes have been made:
796
797 (a) The default value of the buffer size parameter has been increased from
798 8K to 20K. (The actual buffer used is three times this size.)
799
800 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
801 PCRE is built.
802
803 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
804 to be set at run time.
805
806 (d) Numerical values in pcregrep options can be followed by K or M, for
807 example --buffer-size=50K.
808
809 (e) If a line being scanned overflows pcregrep's buffer, an error is now
810 given and the return code is set to 2.
811
812 45. Add a pointer to the latest mark to the callout data block.
813
814 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
815 partial match of an empty string instead of no match. This was specific to
816 the use of ".".
817
818 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
819 complete match instead of a partial match. This bug was dependent on both
820 the PCRE_UTF8 and PCRE_DOTALL options being set.
821
822 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
823 starting byte set, because \b was not being ignored.
824
825
826 Version 8.12 15-Jan-2011
827 ------------------------
828
829 1. Fixed some typos in the markup of the man pages, and wrote a script that
830 checks for such things as part of the documentation building process.
831
832 2. On a big-endian 64-bit system, pcregrep did not correctly process the
833 --match-limit and --recursion-limit options (added for 8.11). In
834 particular, this made one of the standard tests fail. (The integer value
835 went into the wrong half of a long int.)
836
837 3. If the --colour option was given to pcregrep with -v (invert match), it
838 did strange things, either producing crazy output, or crashing. It should,
839 of course, ignore a request for colour when reporting lines that do not
840 match.
841
842 4. Another pcregrep bug caused similar problems if --colour was specified with
843 -M (multiline) and the pattern match finished with a line ending.
844
845 5. In pcregrep, when a pattern that ended with a literal newline sequence was
846 matched in multiline mode, the following line was shown as part of the
847 match. This seems wrong, so I have changed it.
848
849 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
850 the check for further matches in the same line (so they could be coloured)
851 to overrun the end of the current line. If another match was found, it was
852 incorrectly shown (and then shown again when found in the next line).
853
854 7. If pcregrep was compiled under Windows, there was a reference to the
855 function pcregrep_exit() before it was defined. I am assuming this was
856 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
857 reported by a user. I've moved the definition above the reference.
858
859
860 Version 8.11 10-Dec-2010
861 ------------------------
862
863 1. (*THEN) was not working properly if there were untried alternatives prior
864 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
865 backtracked to try for "b" instead of moving to the next alternative branch
866 at the same level (in this case, to look for "c"). The Perl documentation
867 is clear that when (*THEN) is backtracked onto, it goes to the "next
868 alternative in the innermost enclosing group".
869
870 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
871 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
872 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
873 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
874 (*THEN).
875
876 3. If \s appeared in a character class, it removed the VT character from
877 the class, even if it had been included by some previous item, for example
878 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
879 of \s, but is part of the POSIX "space" class.)
880
881 4. A partial match never returns an empty string (because you can always
882 match an empty string at the end of the subject); however the checking for
883 an empty string was starting at the "start of match" point. This has been
884 changed to the "earliest inspected character" point, because the returned
885 data for a partial match starts at this character. This means that, for
886 example, /(?<=abc)def/ gives a partial match for the subject "abc"
887 (previously it gave "no match").
888
889 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
890 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
891 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
892 has an implication that the given string is incomplete (because a partial
893 match is preferred over a full match). For this reason, these items now
894 give a partial match in this situation. [Aside: previously, the one case
895 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
896 match rather than a full match, which was wrong by the old rules, but is
897 now correct.]
898
899 6. There was a bug in the handling of #-introduced comments, recognized when
900 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
901 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
902 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
903 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
904 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
905 places in pcre_compile().
906
907 7. Related to (6) above, when pcre_compile() was skipping #-introduced
908 comments when looking ahead for named forward references to subpatterns,
909 the only newline sequence it recognized was NL. It now handles newlines
910 according to the set newline convention.
911
912 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
913 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
914 cater for a lack of strerror(). These oversights have been fixed.
915
916 9. Added --match-limit and --recursion-limit to pcregrep.
917
918 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
919
920 11. When the -o option was used, pcregrep was setting a return code of 1, even
921 when matches were found, and --line-buffered was not being honoured.
922
923 12. Added an optional parentheses number to the -o and --only-matching options
924 of pcregrep.
925
926 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
927 can match an empty string. The code to do it in pcretest and pcredemo
928 needed fixing:
929
930 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
931 only one byte after an empty string match just before CRLF (this case
932 just got forgotten; "any" and "anycrlf" were OK).
933
934 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
935 mode when an empty string match preceded an ASCII character followed by
936 a non-ASCII character. (The code for advancing by one character rather
937 than one byte was nonsense.)
938
939 (c) The pcredemo.c sample program did not have any code at all to handle
940 the cases when CRLF is a valid newline sequence.
941
942 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
943 as a starting offset was within the subject string. There is now a new
944 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
945 negative or greater than the length of the string. In order to test this,
946 pcretest is extended to allow the setting of negative starting offsets.
947
948 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
949 starting offset points to the beginning of a UTF-8 character was
950 unnecessarily clumsy. I tidied it up.
951
952 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
953 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
954
955 17. Nobody had reported that the --include_dir option, which was added in
956 release 7.7 should have been called --include-dir (hyphen, not underscore)
957 for compatibility with GNU grep. I have changed it to --include-dir, but
958 left --include_dir as an undocumented synonym, and the same for
959 --exclude-dir, though that is not available in GNU grep, at least as of
960 release 2.5.4.
961
962 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
963 characters from a string of bytes) have been redefined so as not to use
964 loops, in order to improve performance in some environments. At the same
965 time, I abstracted some of the common code into auxiliary macros to save
966 repetition (this should not affect the compiled code).
967
968 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
969 compile-time error is now given if \c is not followed by an ASCII
970 character, that is, a byte less than 128. (In EBCDIC mode, the code is
971 different, and any byte value is allowed.)
972
973 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
974 START_OPTIMIZE option, which is now allowed at compile time - but just
975 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
976 to pcregrep and other applications that have no direct access to PCRE
977 options. The new /Y option in pcretest sets this option when calling
978 pcre_compile().
979
980 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
981 back references. Groups containing recursive back references were forced to
982 be atomic by that change, but in the case of named groups, the amount of
983 memory required was incorrectly computed, leading to "Failed: internal
984 error: code overflow". This has been fixed.
985
986 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
987 pcretest.c, to avoid build problems in some Borland environments.
988
989
990 Version 8.10 25-Jun-2010
991 ------------------------
992
993 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
994 THEN.
995
996 2. (*ACCEPT) was not working when inside an atomic group.
997
998 3. Inside a character class, \B is treated as a literal by default, but
999 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1000 causes the error). The code is unchanged, but I tidied the documentation.
1001
1002 4. Inside a character class, PCRE always treated \R and \X as literals,
1003 whereas Perl faults them if its -w option is set. I have changed PCRE so
1004 that it faults them when PCRE_EXTRA is set.
1005
1006 5. Added support for \N, which always matches any character other than
1007 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1008
1009 6. When compiling pcregrep with newer versions of gcc which may have
1010 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1011 declared with attribute warn_unused_result" were given. Just casting the
1012 result to (void) does not stop the warnings; a more elaborate fudge is
1013 needed. I've used a macro to implement this.
1014
1015 7. Minor change to pcretest.c to avoid a compiler warning.
1016
1017 8. Added four artifical Unicode properties to help with an option to make
1018 \s etc use properties (see next item). The new properties are: Xan
1019 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1020
1021 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1022 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1023 this option. Modified pcretest to add /W to test this facility. Added
1024 REG_UCP to make it available via the POSIX interface.
1025
1026 10. Added --line-buffered to pcregrep.
1027
1028 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1029 studied, and the match started with a letter with a code point greater than
1030 127 whose first byte was different to the first byte of the other case of
1031 the letter, the other case of this starting letter was not recognized
1032 (#976).
1033
1034 12. If a pattern that was studied started with a repeated Unicode property
1035 test, for example, \p{Nd}+, there was the theoretical possibility of
1036 setting up an incorrect bitmap of starting bytes, but fortunately it could
1037 not have actually happened in practice until change 8 above was made (it
1038 added property types that matched character-matching opcodes).
1039
1040 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1041 possible starting bytes for non-anchored patterns.
1042
1043 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1044 \R, and also a number of cases that involve Unicode properties, both
1045 explicit and implicit when PCRE_UCP is set.
1046
1047 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1048 input, it could crash or give wrong results if characters with values
1049 greater than 0xc0 were present in the subject string. (Detail: it assumed
1050 UTF-8 input when processing these items.)
1051
1052 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1053 size_t is 64-bit (#991).
1054
1055 17. Added a check for running out of memory when PCRE is compiled with
1056 --disable-stack-for-recursion (#990).
1057
1058 18. If the last data line in a file for pcretest does not have a newline on
1059 the end, a newline was missing in the output.
1060
1061 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1062 less than 128) in its various bitmaps. However, there is a facility for
1063 generating tables according to the current locale when PCRE is compiled. It
1064 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1065 characters, are recognized by isspace() and therefore were getting set in
1066 these tables, and indeed these tables seem to approximate to ISO 8859. This
1067 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1068 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1069 which of course cannot start UTF-8 characters. I have changed the code so
1070 that only real ASCII characters (less than 128) and the correct starting
1071 bytes for UTF-8 encodings are set for characters greater than 127 when in
1072 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1073 altogether.)
1074
1075 20. Added the /T option to pcretest so as to be able to run tests with non-
1076 standard character tables, thus making it possible to include the tests
1077 used for 19 above in the standard set of tests.
1078
1079 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1080 reference to a subpattern the other side of a comment that contains an
1081 opening parenthesis caused either an internal compiling error, or a
1082 reference to the wrong subpattern.
1083
1084
1085 Version 8.02 19-Mar-2010
1086 ------------------------
1087
1088 1. The Unicode data tables have been updated to Unicode 5.2.0.
1089
1090 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1091 configured.
1092
1093 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1094 original author of that file, following a query about its status.
1095
1096 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1097 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1098
1099 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1100 quantifier applied to a forward-referencing subroutine call, could compile
1101 incorrect code or give the error "internal error: previously-checked
1102 referenced subpattern not found".
1103
1104 6. Both MS Visual Studio and Symbian OS have problems with initializing
1105 variables to point to external functions. For these systems, therefore,
1106 pcre_malloc etc. are now initialized to local functions that call the
1107 relevant global functions.
1108
1109 7. There were two entries missing in the vectors called coptable and poptable
1110 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1111 I've fixed the data, and added a kludgy way of testing at compile time that
1112 the lengths are correct (equal to the number of opcodes).
1113
1114 8. Following on from 7, I added a similar kludge to check the length of the
1115 eint vector in pcreposix.c.
1116
1117 9. Error texts for pcre_compile() are held as one long string to avoid too
1118 much relocation at load time. To find a text, the string is searched,
1119 counting zeros. There was no check for running off the end of the string,
1120 which could happen if a new error number was added without updating the
1121 string.
1122
1123 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1124
1125 11. \K was not working if it appeared in an atomic group or in a group that
1126 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1127 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1128 the assertion is positive, but not if it is negative.
1129
1130 12. Change 11 fortuitously reduced the size of the stack frame used in the
1131 "match()" function of pcre_exec.c by one pointer. Forthcoming
1132 implementation of support for (*MARK) will need an extra pointer on the
1133 stack; I have reserved it now, so that the stack frame size does not
1134 decrease.
1135
1136 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1137 item in branch that calls a recursion is a subroutine call - as in the
1138 second branch in the above example - was incorrectly given the compile-
1139 time error "recursive call could loop indefinitely" because pcre_compile()
1140 was not correctly checking the subroutine for matching a non-empty string.
1141
1142 14. The checks for overrunning compiling workspace could trigger after an
1143 overrun had occurred. This is a "should never occur" error, but it can be
1144 triggered by pathological patterns such as hundreds of nested parentheses.
1145 The checks now trigger 100 bytes before the end of the workspace.
1146
1147 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1148
1149
1150 Version 8.01 19-Jan-2010
1151 ------------------------
1152
1153 1. If a pattern contained a conditional subpattern with only one branch (in
1154 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1155 computed the wrong minimum data length (which is of course zero for such
1156 subpatterns). This could cause incorrect "no match" results.
1157
1158 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1159 the pattern is reset in the first branch, pcre_compile() failed with
1160 "internal error: code overflow at offset...". This happened only when
1161 the reset was to the original external option setting. (An optimization
1162 abstracts leading options settings into an external setting, which was the
1163 cause of this.)
1164
1165 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1166 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1167 assertion pattern did not match (meaning that the assertion was true), it
1168 was incorrectly treated as false if the SKIP had been reached during the
1169 matching. This also applied to assertions used as conditions.
1170
1171 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1172 assertion subpattern, including such a pattern used as a condition,
1173 unpredictable results occurred, instead of the error return
1174 PCRE_ERROR_DFA_UITEM.
1175
1176 5. The C++ GlobalReplace function was not working like Perl for the special
1177 situation when an empty string is matched. It now does the fancy magic
1178 stuff that is necessary.
1179
1180 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1181 removed. (These were left over from very, very early versions of PCRE.)
1182
1183 7. Some cosmetic changes to the code to make life easier when compiling it
1184 as part of something else:
1185
1186 (a) Change DEBUG to PCRE_DEBUG.
1187
1188 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1189 called "current" as "current_branch", to prevent a collision with the
1190 Linux macro when compiled as a kernel module.
1191
1192 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1193 prevent a collision with the Linux macro when compiled as a kernel
1194 module.
1195
1196 8. In pcre_compile() there are some checks for integer overflows that used to
1197 cast potentially large values to (double). This has been changed to that
1198 when building, a check for int64_t is made, and if it is found, it is used
1199 instead, thus avoiding the use of floating point arithmetic. (There is no
1200 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1201 double.
1202
1203 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1204 2005 (difference between two addresses compared to an unsigned value).
1205
1206 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1207 custom one, because of the following reported problem in Windows:
1208
1209 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1210 under Win32.
1211 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1212 therefore missing the function definition.
1213 - The compiler thus generates a "C" signature for the test function.
1214 - The linker fails to find the "C" function.
1215 - PCRE fails to configure if asked to do so against libbz2.
1216
1217 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1218 messages were output:
1219
1220 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1221 rerunning libtoolize, to keep the correct libtool macros in-tree.
1222 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1223
1224 I have done both of these things.
1225
1226 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1227 most of the time, it *can* run out if it is given a pattern that contains a
1228 runaway infinite recursion. I updated the discussion in the pcrestack man
1229 page.
1230
1231 13. Now that we have gone to the x.xx style of version numbers, the minor
1232 version may start with zero. Using 08 or 09 is a bad idea because users
1233 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1234 interpreted as invalid octal numbers. I've updated the previous comment in
1235 configure.ac, and also added a check that gives an error if 08 or 09 are
1236 used.
1237
1238 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1239 causing partial matching to fail when the end of the subject matched \W
1240 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1241
1242 15. There were some discrepancies between the declarations in pcre_internal.h
1243 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1244 their definitions. The declarations used "const uschar *" and the
1245 definitions used USPTR. Even though USPTR is normally defined as "const
1246 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1247 reported that: "This difference in casting confuses some C++ compilers, for
1248 example, SunCC recognizes above declarations as different functions and
1249 generates broken code for hbpcre." I have changed the declarations to use
1250 USPTR.
1251
1252 16. GNU libtool is named differently on some systems. The autogen.sh script now
1253 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1254 (FreeBSD).
1255
1256 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1257 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1258 comment: "Figure out how to create a longlong from a string: strtoll and
1259 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1260 instance, but it only takes 2 args instead of 3!"
1261
1262 18. A subtle bug concerned with back references has been fixed by a change of
1263 specification, with a corresponding code fix. A pattern such as
1264 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1265 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1266 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1267 same bug. Such groups have to be quantified to be useful, or contained
1268 inside another quantified group. (If there's no repetition, the reference
1269 can never match.) The problem arises because, having left the group and
1270 moved on to the rest of the pattern, a later failure that backtracks into
1271 the group uses the captured value from the final iteration of the group
1272 rather than the correct earlier one. I have fixed this in PCRE by forcing
1273 any group that contains a reference to itself to be an atomic group; that
1274 is, there cannot be any backtracking into it once it has completed. This is
1275 similar to recursive and subroutine calls.
1276
1277
1278 Version 8.00 19-Oct-09
1279 ----------------------
1280
1281 1. The table for translating pcre_compile() error codes into POSIX error codes
1282 was out-of-date, and there was no check on the pcre_compile() error code
1283 being within the table. This could lead to an OK return being given in
1284 error.
1285
1286 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1287 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1288 in a Windows environment.
1289
1290 3. The pcregrep --count option prints the count for each file even when it is
1291 zero, as does GNU grep. However, pcregrep was also printing all files when
1292 --files-with-matches was added. Now, when both options are given, it prints
1293 counts only for those files that have at least one match. (GNU grep just
1294 prints the file name in this circumstance, but including the count seems
1295 more useful - otherwise, why use --count?) Also ensured that the
1296 combination -clh just lists non-zero counts, with no names.
1297
1298 4. The long form of the pcregrep -F option was incorrectly implemented as
1299 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1300 but it seems right to fix it, and I didn't think it was worth preserving
1301 the old behaviour.
1302
1303 5. The command line items --regex=pattern and --regexp=pattern were not
1304 recognized by pcregrep, which required --regex pattern or --regexp pattern
1305 (with a space rather than an '='). The man page documented the '=' forms,
1306 which are compatible with GNU grep; these now work.
1307
1308 6. No libpcreposix.pc file was created for pkg-config; there was just
1309 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1310
1311 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1312 when UCP support is not needed, by modifying the Python script that
1313 generates it from Unicode data files. This should not matter if the module
1314 is correctly used as a library, but I received one complaint about 50K of
1315 unwanted data. My guess is that the person linked everything into his
1316 program rather than using a library. Anyway, it does no harm.
1317
1318 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1319 was a minimum greater than 1 for a wide character in a possessive
1320 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1321 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1322 character. Chaos in the form of incorrect output or a compiling loop could
1323 result.
1324
1325 9. The restrictions on what a pattern can contain when partial matching is
1326 requested for pcre_exec() have been removed. All patterns can now be
1327 partially matched by this function. In addition, if there are at least two
1328 slots in the offset vector, the offset of the earliest inspected character
1329 for the match and the offset of the end of the subject are set in them when
1330 PCRE_ERROR_PARTIAL is returned.
1331
1332 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1333 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1334 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1335 and may be more useful for multi-segment matching.
1336
1337 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1338 used to be given if ever the end of the subject was reached; now it is
1339 given only if matching could not proceed because another character was
1340 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1341 string "Z", which now yields "no match" instead of "partial match". In the
1342 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1343 final character ended with (*FAIL).
1344
1345 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1346 if the pattern had a "must contain" character that was already found in the
1347 earlier partial match, unless partial matching was again requested. For
1348 example, with the pattern /dog.(body)?/, the "must contain" character is
1349 "g". If the first part-match was for the string "dog", restarting with
1350 "sbody" failed. This bug has been fixed.
1351
1352 13. The string returned by pcre_dfa_exec() after a partial match has been
1353 changed so that it starts at the first inspected character rather than the
1354 first character of the match. This makes a difference only if the pattern
1355 starts with a lookbehind assertion or \b or \B (\K is not supported by
1356 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1357 matching functions compatible, and I think it's the right thing to do.
1358
1359 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1360 so that the demonstration program is easily available in environments where
1361 PCRE has not been installed from source.
1362
1363 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1364 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1365 library.
1366
1367 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1368 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1369 is not the first non-POSIX option to be added. Clearly some people find
1370 these options useful.
1371
1372 17. If a caller to the POSIX matching function regexec() passes a non-zero
1373 value for nmatch with a NULL value for pmatch, the value of
1374 nmatch is forced to zero.
1375
1376 18. RunGrepTest did not have a test for the availability of the -u option of
1377 the diff command, as RunTest does. It now checks in the same way as
1378 RunTest, and also checks for the -b option.
1379
1380 19. If an odd number of negated classes containing just a single character
1381 interposed, within parentheses, between a forward reference to a named
1382 subpattern and the definition of the subpattern, compilation crashed with
1383 an internal error, complaining that it could not find the referenced
1384 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1385 [The bug was that it was starting one character too far in when skipping
1386 over the character class, thus treating the ] as data rather than
1387 terminating the class. This meant it could skip too much.]
1388
1389 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1390 /g option in pcretest when the pattern contains \K, which makes it possible
1391 to have an empty string match not at the start, even when the pattern is
1392 anchored. Updated pcretest and pcredemo to use this option.
1393
1394 21. If the maximum number of capturing subpatterns in a recursion was greater
1395 than the maximum at the outer level, the higher number was returned, but
1396 with unset values at the outer level. The correct (outer level) value is
1397 now given.
1398
1399 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1400 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1401 make it do so. The string so far is captured, making this feature
1402 compatible with Perl.
1403
1404 23. The tests have been re-organized, adding tests 11 and 12, to make it
1405 possible to check the Perl 5.10 features against Perl 5.10.
1406
1407 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1408 pattern matches a fixed length string. PCRE did not allow this; now it
1409 does. Neither allows recursion.
1410
1411 25. I finally figured out how to implement a request to provide the minimum
1412 length of subject string that was needed in order to match a given pattern.
1413 (It was back references and recursion that I had previously got hung up
1414 on.) This code has now been added to pcre_study(); it finds a lower bound
1415 to the length of subject needed. It is not necessarily the greatest lower
1416 bound, but using it to avoid searching strings that are too short does give
1417 some useful speed-ups. The value is available to calling programs via
1418 pcre_fullinfo().
1419
1420 26. While implementing 25, I discovered to my embarrassment that pcretest had
1421 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1422 study optimizations had never been tested with that matching function.
1423 Oops. What is worse, even when it was passed study data, there was a bug in
1424 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1425 were also very few tests of studied patterns with pcre_dfa_exec().
1426
1427 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1428 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1429 on the other side of the coin, they are no longer allowed to have different
1430 names, because these cannot be distinguished in PCRE, and this has caused
1431 confusion. (This is a difference from Perl.)
1432
1433 28. When duplicate subpattern names are present (necessarily with different
1434 numbers, as required by 27 above), and a test is made by name in a
1435 conditional pattern, either for a subpattern having been matched, or for
1436 recursion in such a pattern, all the associated numbered subpatterns are
1437 tested, and the overall condition is true if the condition is true for any
1438 one of them. This is the way Perl works, and is also more like the way
1439 testing by number works.
1440
1441
1442 Version 7.9 11-Apr-09
1443 ---------------------
1444
1445 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1446 (pcretest), all targets were linked against these libraries. This included
1447 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1448 libraries. This caused unwanted dependencies to be created. This problem
1449 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1450 pcretest is linked with readline.
1451
1452 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1453 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1454 moved outside it again, because FALSE and TRUE are already defined in AIX,
1455 but BOOL is not.
1456
1457 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1458 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1459
1460 4. The pcregrep documentation said spaces were inserted as well as colons (or
1461 hyphens) following file names and line numbers when outputting matching
1462 lines. This is not true; no spaces are inserted. I have also clarified the
1463 wording for the --colour (or --color) option.
1464
1465 5. In pcregrep, when --colour was used with -o, the list of matching strings
1466 was not coloured; this is different to GNU grep, so I have changed it to be
1467 the same.
1468
1469 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1470 each matching line was coloured. Now it goes on to look for further matches
1471 of any of the test patterns, which is the same behaviour as GNU grep.
1472
1473 7. A pattern that could match an empty string could cause pcregrep to loop; it
1474 doesn't make sense to accept an empty string match in pcregrep, so I have
1475 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1476 seems to be how GNU grep behaves.
1477
1478 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1479 start or after a newline", because the conditional assertion was not being
1480 correctly handled. The rule now is that both the assertion and what follows
1481 in the first alternative must satisfy the test.
1482
1483 9. If auto-callout was enabled in a pattern with a conditional group whose
1484 condition was an assertion, PCRE could crash during matching, both with
1485 pcre_exec() and pcre_dfa_exec().
1486
1487 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1488 used for matching.
1489
1490 11. Unicode property support in character classes was not working for
1491 characters (bytes) greater than 127 when not in UTF-8 mode.
1492
1493 12. Added the -M command line option to pcretest.
1494
1495 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1496
1497 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1498
1499 16. Added comments and documentation about mis-use of no_arg in the C++
1500 wrapper.
1501
1502 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1503 from Martin Jerabek that uses macro names for all relevant character and
1504 string constants.
1505
1506 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1507 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1508 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1509 these, but not everybody uses configure.
1510
1511 19. A conditional group that had only one branch was not being correctly
1512 recognized as an item that could match an empty string. This meant that an
1513 enclosing group might also not be so recognized, causing infinite looping
1514 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1515 with the subject "ab", where knowledge that the repeated group can match
1516 nothing is needed in order to break the loop.
1517
1518 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1519 exec(), but without supplying a callout function, matching went wrong.
1520
1521 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1522 leak if the size of the offset vector was greater than 30. When the vector
1523 is smaller, the saved offsets during recursion go onto a local stack
1524 vector, but for larger vectors malloc() is used. It was failing to free
1525 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1526 error, in fact).
1527
1528 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1529 heapframe that is used only when UTF-8 support is enabled. This caused no
1530 problem, but was untidy.
1531
1532 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1533 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1534 included within another project.
1535
1536 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1537 slightly modified by me:
1538
1539 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1540 not building pcregrep.
1541
1542 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1543 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1544
1545 25. Forward references, both numeric and by name, in patterns that made use of
1546 duplicate group numbers, could behave incorrectly or give incorrect errors,
1547 because when scanning forward to find the reference group, PCRE was not
1548 taking into account the duplicate group numbers. A pattern such as
1549 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1550
1551 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1552 the feature of a custom pointer more persuasive (as requested by a user).
1553
1554 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1555 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1556 pre-defined.
1557
1558 28. Added support for (*UTF8) at the start of a pattern.
1559
1560 29. Arrange for flags added by the "release type" setting in CMake to be shown
1561 in the configuration summary.
1562
1563
1564 Version 7.8 05-Sep-08
1565 ---------------------
1566
1567 1. Replaced UCP searching code with optimized version as implemented for Ad
1568 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1569 stage table and inline lookup instead of a function, giving speed ups of 2
1570 to 5 times on some simple patterns that I tested. Permission was given to
1571 distribute the MultiStage2.py script that generates the tables (it's not in
1572 the tarball, but is in the Subversion repository).
1573
1574 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1575 scripts.
1576
1577 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1578 a group with a zero qualifier. The result of the study could be incorrect,
1579 or the function might crash, depending on the pattern.
1580
1581 4. Caseless matching was not working for non-ASCII characters in back
1582 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1583 It now works when Unicode Property Support is available.
1584
1585 5. In pcretest, an escape such as \x{de} in the data was always generating
1586 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1587 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1588 truncation.
1589
1590 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1591
1592 7. Added two (int) casts to pcregrep when printing the difference of two
1593 pointers, in case they are 64-bit values.
1594
1595 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1596 test 2 if it fails.
1597
1598 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1599 and a #define of that name to empty if it is not externally set. This is to
1600 allow users of MSVC to set it if necessary.
1601
1602 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1603 the convenience functions in the pcre_get.c source file.
1604
1605 11. An option change at the start of a pattern that had top-level alternatives
1606 could cause overwriting and/or a crash. This command provoked a crash in
1607 some environments:
1608
1609 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1610
1611 This potential security problem was recorded as CVE-2008-2371.
1612
1613 12. For a pattern where the match had to start at the beginning or immediately
1614 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1615 pcre_dfa_exec() could read past the end of the passed subject if there was
1616 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1617 pcretest so that it places the subject at the end of its malloc-ed buffer.
1618
1619 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1620 exec() might read past the end of the data buffer in UTF-8 mode.
1621
1622 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1623 the data contained the byte 0x85 as part of a UTF-8 character within its
1624 first line. This applied both to normal and DFA matching.
1625
1626 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1627 /^[^d]*?$/8 failed to match "abc".
1628
1629 16. Added a missing copyright notice to pcrecpp_internal.h.
1630
1631 17. Make it more clear in the documentation that values returned from
1632 pcre_exec() in ovector are byte offsets, not character counts.
1633
1634 18. Tidied a few places to stop certain compilers from issuing warnings.
1635
1636 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1637 supplied by Stefan Weber. I made a further small update for 7.8 because
1638 there is a change of source arrangements: the pcre_searchfuncs.c module is
1639 replaced by pcre_ucd.c.
1640
1641
1642 Version 7.7 07-May-08
1643 ---------------------
1644
1645 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1646 a string to a long long, pretend we don't even have a long long." This is
1647 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1648
1649 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1650 pre-7.6 versions, which defined a global no_arg variable instead of putting
1651 it in the RE class. (See also #8 below.)
1652
1653 3. Remove a line of dead code, identified by coverity and reported by Nuno
1654 Lopes.
1655
1656 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1657
1658 (1) The include/exclude patterns were being applied to the whole pathnames
1659 of files, instead of just to the final components.
1660
1661 (2) If there was more than one level of directory, the subdirectories were
1662 skipped unless they satisfied the include/exclude conditions. This is
1663 inconsistent with GNU grep (and could even be seen as contrary to the
1664 pcregrep specification - which I improved to make it absolutely clear).
1665 The action now is always to scan all levels of directory, and just
1666 apply the include/exclude patterns to regular files.
1667
1668 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1669 --exclude_dir in the tests to avoid scanning .svn directories.
1670
1671 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1672 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1673 doesn't support NULs in patterns.
1674
1675 7. Added some missing "const"s to declarations of static tables in
1676 pcre_compile.c and pcre_dfa_exec.c.
1677
1678 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1679 caused by fix #2 above. (Subsequently also a second patch to fix the
1680 first patch. And a third patch - this was a messy problem.)
1681
1682 9. Applied Craig's patch to remove the use of push_back().
1683
1684 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1685 matching function regexec().
1686
1687 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1688 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1689 references. PCRE supports relative numbers with this syntax (I don't think
1690 Oniguruma does).
1691
1692 12. Previously, a group with a zero repeat such as (...){0} was completely
1693 omitted from the compiled regex. However, this means that if the group
1694 was called as a subroutine from elsewhere in the pattern, things went wrong
1695 (an internal error was given). Such groups are now left in the compiled
1696 pattern, with a new opcode that causes them to be skipped at execution
1697 time.
1698
1699 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1700 to the way PCRE behaves:
1701
1702 (a) A lone ] character is dis-allowed (Perl treats it as data).
1703
1704 (b) A back reference to an unmatched subpattern matches an empty string
1705 (Perl fails the current match path).
1706
1707 (c) A data ] in a character class must be notated as \] because if the
1708 first data character in a class is ], it defines an empty class. (In
1709 Perl it is not possible to have an empty class.) The empty class []
1710 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1711 The negative empty class [^] matches any one character, independently
1712 of the DOTALL setting.
1713
1714 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1715 non-existent subpattern following a character class starting with ']' and
1716 containing () gave an internal compiling error instead of "reference to
1717 non-existent subpattern". Fortunately, when the pattern did exist, the
1718 compiled code was correct. (When scanning forwards to check for the
1719 existencd of the subpattern, it was treating the data ']' as terminating
1720 the class, so got the count wrong. When actually compiling, the reference
1721 was subsequently set up correctly.)
1722
1723 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1724 it was being rejected as not supported by pcre_dfa_exec(), even though
1725 other assertions are supported. I have made pcre_dfa_exec() support
1726 (*FAIL).
1727
1728 16. The implementation of 13c above involved the invention of a new opcode,
1729 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1730 cannot be changed at match time, I realized I could make a small
1731 improvement to matching performance by compiling OP_ALLANY instead of
1732 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1733 on the OP_ANY path.
1734
1735 17. Compiling pcretest on Windows with readline support failed without the
1736 following two fixes: (1) Make the unistd.h include conditional on
1737 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1738
1739 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1740 ncurses library to be included for pcretest when ReadLine support is
1741 requested, but also to allow for it to be overridden. This patch came from
1742 Daniel Bergström.
1743
1744 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1745 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1746 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1747 spotting this.
1748
1749
1750 Version 7.6 28-Jan-08
1751 ---------------------
1752
1753 1. A character class containing a very large number of characters with
1754 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1755 overflow.
1756
1757 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1758 HAVE_LONG_LONG is not defined.
1759
1760 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1761 bring them up to date and include new features. This patch includes:
1762
1763 - Fixed PH's badly added libz and libbz2 support.
1764 - Fixed a problem with static linking.
1765 - Added pcredemo. [But later removed - see 7 below.]
1766 - Fixed dftables problem and added an option.
1767 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1768 HAVE_LONG_LONG.
1769 - Added readline support for pcretest.
1770 - Added an listing of the option settings after cmake has run.
1771
1772 4. A user submitted a patch to Makefile that makes it easy to create
1773 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1774 Makefile.am that cause it to include this special target, without
1775 affecting anything else. Note that the same mingw target plus all
1776 the other distribution libraries and programs are now supported
1777 when configuring with CMake (see 6 below) instead of with
1778 Configure/Make.
1779
1780 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1781 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1782 exported in the Windows port". It has not yet been confirmed that the patch
1783 solves the problem, but it does no harm.
1784
1785 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1786 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1787 with CMake, and also correct the comment about stack recursion.
1788
1789 7. Remove the automatic building of pcredemo from the ./configure system and
1790 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1791 of a program that users should build themselves after PCRE is installed, so
1792 building it automatically is not really right. What is more, it gave
1793 trouble in some build environments.
1794
1795 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1796
1797
1798 Version 7.5 10-Jan-08
1799 ---------------------
1800
1801 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1802 values in parens when parsing an RE using the C++ wrapper."
1803
1804 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1805 Characters greater than 255 were excluded from the class instead of being
1806 included.
1807
1808 3. The same bug as (2) above applied to negated POSIX classes such as
1809 [:^space:].
1810
1811 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1812 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1813 I have changed it.
1814
1815 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1816 first named subpattern) and a construct such as (?&a) would reference the
1817 first named subpattern whose name started with "a" (in other words, the
1818 length check was missing). Both these problems are fixed. "Subpattern name
1819 expected" is now given for (?&) (a zero-length name), and this patch also
1820 makes it give the same error for \k'' (previously it complained that that
1821 was a reference to a non-existent subpattern).
1822
1823 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1824 this is right because (?- can be followed by option settings as well as by
1825 digits. I have, however, made the messages clearer.
1826
1827 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1828 than the number used in the conditional) now cause a compile-time error.
1829 This is actually not compatible with Perl, which accepts such patterns, but
1830 treats the conditional as always being FALSE (as PCRE used to), but it
1831 seems to me that giving a diagnostic is better.
1832
1833 8. Change "alphameric" to the more common word "alphanumeric" in comments
1834 and messages.
1835
1836 9. Fix two occurrences of "backslash" in comments that should have been
1837 "backspace".
1838
1839 10. Remove two redundant lines of code that can never be obeyed (their function
1840 was moved elsewhere).
1841
1842 11. The program that makes PCRE's Unicode character property table had a bug
1843 which caused it to generate incorrect table entries for sequences of
1844 characters that have the same character type, but are in different scripts.
1845 It amalgamated them into a single range, with the script of the first of
1846 them. In other words, some characters were in the wrong script. There were
1847 thirteen such cases, affecting characters in the following ranges:
1848
1849 U+002b0 - U+002c1
1850 U+0060c - U+0060d
1851 U+0061e - U+00612
1852 U+0064b - U+0065e
1853 U+0074d - U+0076d
1854 U+01800 - U+01805
1855 U+01d00 - U+01d77
1856 U+01d9b - U+01dbf
1857 U+0200b - U+0200f
1858 U+030fc - U+030fe
1859 U+03260 - U+0327f
1860 U+0fb46 - U+0fbb1
1861 U+10450 - U+1049d
1862
1863 12. The -o option (show only the matching part of a line) for pcregrep was not
1864 compatible with GNU grep in that, if there was more than one match in a
1865 line, it showed only the first of them. It now behaves in the same way as
1866 GNU grep.
1867
1868 13. If the -o and -v options were combined for pcregrep, it printed a blank
1869 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1870 does the same. The return code can be used to tell if there were any
1871 non-matching lines.
1872
1873 14. Added --file-offsets and --line-offsets to pcregrep.
1874
1875 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1876 infinitely looping recursion. The bug was that positive lookaheads were not
1877 being skipped when checking for a possible empty match (negative lookaheads
1878 and both kinds of lookbehind were skipped).
1879
1880 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1881 inclusion of <windows.h> to before rather than after the definition of
1882 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1883
1884 17. Specifying a possessive quantifier with a specific limit for a Unicode
1885 character property caused pcre_compile() to compile bad code, which led at
1886 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1887 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1888 caused the error; without that there was no problem.
1889
1890 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1891
1892 19. Added --enable-pcretest-libreadline.
1893
1894 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1895 RE::GlobalReplace(). As a result, the number of replacements returned was
1896 double what it should be. I removed one of the increments, but Craig sent a
1897 later patch that removed the other one (the right fix) and added unit tests
1898 that check the return values (which was not done before).
1899
1900 21. Several CMake things:
1901
1902 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1903 the names libpcre and libpcreposix, not just pcre and pcreposix.
1904
1905 (2) The above change means that pcretest and pcregrep are now correctly
1906 linked with the newly-built libraries, not previously installed ones.
1907
1908 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1909
1910 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1911 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1912 UTF-8 newline character). The key issue is that the pattern starts .*;
1913 this means that the match must be either at the beginning, or after a
1914 newline. The bug was in the code for advancing after a failed match and
1915 checking that the new position followed a newline. It was not taking
1916 account of UTF-8 characters correctly.
1917
1918 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1919 character classes. PCRE was not treating the sequence [:...:] as a
1920 character class unless the ... were all letters. Perl, however, seems to
1921 allow any characters between [: and :], though of course it rejects as
1922 unknown any "names" that contain non-letters, because all the known class
1923 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1924 for example, whereas PCRE did not - it did not recognize a POSIX character
1925 class. This seemed a bit dangerous, so the code has been changed to be
1926 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1927 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1928 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1929 Perl does, and where it didn't before.
1930
1931 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1932 Windows environments %n is disabled by default.
1933
1934
1935 Version 7.4 21-Sep-07
1936 ---------------------
1937
1938 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1939 means that a class such as [\s] counted as "explicit reference to CR or
1940 LF". That isn't really right - the whole point of the change was to try to
1941 help when there was an actual mention of one of the two characters. So now
1942 the change happens only if \r or \n (or a literal CR or LF) character is
1943 encountered.
1944
1945 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1946 of both had grown to the point where there were only 3 bits left.
1947 Fortunately, there was spare space in the data structure, and so I have
1948 moved the internal flags into a new 16-bit field to free up more option
1949 bits.
1950
1951 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1952 but did not set the internal JCHANGED flag - either of these is enough to
1953 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1954 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1955 start sets both bits.
1956
1957 4. Added options (at build time, compile time, exec time) to change \R from
1958 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1959
1960 5. doc/pcresyntax.html was missing from the distribution.
1961
1962 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1963 compatibility, even though it is no longer used.
1964
1965 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1966 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1967 windows.h file is present (where different names are used). [This was
1968 reversed later after testing - see 16 below.]
1969
1970 8. Changed all #include <config.h> to #include "config.h". There were also
1971 some further <pcre.h> cases that I changed to "pcre.h".
1972
1973 9. When pcregrep was used with the --colour option, it missed the line ending
1974 sequence off the lines that it output.
1975
1976 10. It was pointed out to me that arrays of string pointers cause lots of
1977 relocations when a shared library is dynamically loaded. A technique of
1978 using a single long string with a table of offsets can drastically reduce
1979 these. I have refactored PCRE in four places to do this. The result is
1980 dramatic:
1981
1982 Originally: 290
1983 After changing UCP table: 187
1984 After changing error message table: 43
1985 After changing table of "verbs" 36
1986 After changing table of Posix names 22
1987
1988 Thanks to the folks working on Gregex for glib for this insight.
1989
1990 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1991 unicode-properties was also set.
1992
1993 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1994
1995 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1996 checked only for CRLF.
1997
1998 14. Added casts to pcretest.c to avoid compiler warnings.
1999
2000 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2001
2002 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2003 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2004 entirely. This removes changes made in 7 above.
2005
2006 17. The CMake files have been updated, and there is now more information about
2007 building with CMake in the NON-UNIX-USE document.
2008
2009
2010 Version 7.3 28-Aug-07
2011 ---------------------
2012
2013 1. In the rejigging of the build system that eventually resulted in 7.1, the
2014 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2015 brackets there is not right, since it causes compilers to look for an
2016 installed pcre.h, not the version that is in the source that is being
2017 compiled (which of course may be different). I have changed it back to:
2018
2019 #include "pcre.h"
2020
2021 I have a vague recollection that the change was concerned with compiling in
2022 different directories, but in the new build system, that is taken care of
2023 by the VPATH setting the Makefile.
2024
2025 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2026 when the subject happened to end in the byte 0x85 (e.g. if the last
2027 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2028 characters but of course it shouldn't be taken as a newline when it is part
2029 of another character. The bug was that, for an unlimited repeat of . in
2030 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2031 characters when looking for a newline.
2032
2033 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2034
2035 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2036 in debug output.
2037
2038 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2039 long printing in the pcrecpp unittest when running under MinGW.
2040
2041 6. ESC_K was left out of the EBCDIC table.
2042
2043 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2044 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2045 limit also applies to "virtual nesting" when a pattern is recursive, and in
2046 this case 1000 isn't so big. I have been able to remove this limit at the
2047 expense of backing off one optimization in certain circumstances. Normally,
2048 when pcre_exec() would call its internal match() function recursively and
2049 immediately return the result unconditionally, it uses a "tail recursion"
2050 feature to save stack. However, when a subpattern that can match an empty
2051 string has an unlimited repetition quantifier, it no longer makes this
2052 optimization. That gives it a stack frame in which to save the data for
2053 checking that an empty string has been matched. Previously this was taken
2054 from the 1000-entry workspace that had been reserved. So now there is no
2055 explicit limit, but more stack is used.
2056
2057 8. Applied Daniel's patches to solve problems with the import/export magic
2058 syntax that is required for Windows, and which was going wrong for the
2059 pcreposix and pcrecpp parts of the library. These were overlooked when this
2060 problem was solved for the main library.
2061
2062 9. There were some crude static tests to avoid integer overflow when computing
2063 the size of patterns that contain repeated groups with explicit upper
2064 limits. As the maximum quantifier is 65535, the maximum group length was
2065 set at 30,000 so that the product of these two numbers did not overflow a
2066 32-bit integer. However, it turns out that people want to use groups that
2067 are longer than 30,000 bytes (though not repeat them that many times).
2068 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2069 made it possible to implement the integer overflow checks in a much more
2070 dynamic way, which I have now done. The artificial limitation on group
2071 length has been removed - we now have only the limit on the total length of
2072 the compiled pattern, which depends on the LINK_SIZE setting.
2073
2074 10. Fixed a bug in the documentation for get/copy named substring when
2075 duplicate names are permitted. If none of the named substrings are set, the
2076 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2077 empty string.
2078
2079 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2080 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2081 because the ] is interpreted as the first data character and the
2082 terminating ] is not found. PCRE has been made compatible with Perl in this
2083 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2084 cause memory overwriting.
2085
2086 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2087 string has been matched (to stop an infinite loop). It was not recognizing
2088 a conditional subpattern that could match an empty string if that
2089 subpattern was within another subpattern. For example, it looped when
2090 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2091 condition was not nested. This bug has been fixed.
2092
2093 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2094 past the start of the subject in the presence of bytes with the top bit
2095 set, for example "\x8aBCD".
2096
2097 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2098 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2099
2100 14. Optimized (?!) to (*FAIL).
2101
2102 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2103 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2104 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2105 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2106 does: it's just the validity check that is more restrictive.
2107
2108 16. Inserted checks for integer overflows during escape sequence (backslash)
2109 processing, and also fixed erroneous offset values for syntax errors during
2110 backslash processing.
2111
2112 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2113 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2114
2115 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2116 caused an overrun.
2117
2118 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2119 something other than just ASCII characters) inside a group that had an
2120 unlimited repeat caused a loop at compile time (while checking to see
2121 whether the group could match an empty string).
2122
2123 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2124 [\P{Any}] did so. (Error in the code for printing property names.)
2125
2126 21. An orphan \E inside a character class could cause a crash.
2127
2128 22. A repeated capturing bracket such as (A)? could cause a wild memory
2129 reference during compilation.
2130
2131 23. There are several functions in pcre_compile() that scan along a compiled
2132 expression for various reasons (e.g. to see if it's fixed length for look
2133 behind). There were bugs in these functions when a repeated \p or \P was
2134 present in the pattern. These operators have additional parameters compared
2135 with \d, etc, and these were not being taken into account when moving along
2136 the compiled data. Specifically:
2137
2138 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2139 length.
2140
2141 (b) An item such as \pL+ within a repeated group could cause crashes or
2142 loops.
2143
2144 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2145 "reference to non-existent subpattern" error.
2146
2147 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2148
2149 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2150 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2151
2152 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2153
2154 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2155 character were causing crashes (broken optimization).
2156
2157 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2158 \p or \P) caused a compile-time loop.
2159
2160 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2161 break. For example, the unstudied pattern [\r\n]A does not match the string
2162 "\r\nA" because change 7.0/46 below moves the current point on by two
2163 characters after failing to match at the start. However, the pattern \nA
2164 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2165 the same is true. There doesn't seem any very clean way out of this, but
2166 what I have chosen to do makes the common cases work: PCRE now takes note
2167 of whether there can be an explicit match for \r or \n anywhere in the
2168 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2169 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2170 pattern has explicit CR or LF references.
2171
2172 29. Added (*CR) etc for changing newline setting at start of pattern.
2173
2174
2175 Version 7.2 19-Jun-07
2176 ---------------------
2177
2178 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2179 which is apparently normally available under Windows.
2180
2181 2. Re-jig the pcregrep tests with different newline settings in an attempt
2182 to make them independent of the local environment's newline setting.
2183
2184 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2185
2186 4. Some of the "internals" tests were previously cut out when the link size
2187 was not 2, because the output contained actual offsets. The recent new
2188 "Z" feature of pcretest means that these can be cut out, making the tests
2189 usable with all link sizes.
2190
2191 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2192 stack recursion. This gives a massive performance boost under BSD, but just
2193 a small improvement under Linux. However, it saves one field in the frame
2194 in all cases.
2195
2196 6. Added more features from the forthcoming Perl 5.10:
2197
2198 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2199 recursion call. It refers to the nth most recently opened parentheses.
2200
2201 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2202 to be opened parentheses.
2203
2204 (c) Conditions that refer to capturing parentheses can be specified
2205 relatively, for example, (?(-2)... or (?(+3)...
2206
2207 (d) \K resets the start of the current match so that everything before
2208 is not part of it.
2209
2210 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2211
2212 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2213 reference syntax.
2214
2215 (g) (?| introduces a group in which the numbering of parentheses in each
2216 alternative starts with the same number.
2217
2218 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2219
2220 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2221 PCRE_INFO_JCHANGED.
2222
2223 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2224 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2225 for detecting groups that can match an empty string.
2226
2227 9. A pattern with a very large number of alternatives (more than several
2228 hundred) was running out of internal workspace during the pre-compile
2229 phase, where pcre_compile() figures out how much memory will be needed. A
2230 bit of new cunning has reduced the workspace needed for groups with
2231 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2232 workspace instead of running out of the 4096 that are available.
2233
2234 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2235
2236 11. Applied patch from Google to remove an optimization that didn't quite work.
2237 The report of the bug said:
2238
2239 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2240 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2241 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2242
2243 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2244 it matched the wrong number of bytes.
2245
2246
2247 Version 7.1 24-Apr-07
2248 ---------------------
2249
2250 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2251 that is more "standard", making use of automake and other Autotools. There
2252 is some re-arrangement of the files and adjustment of comments consequent
2253 on this.
2254
2255 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2256 for recursive directory scanning broke on some systems because the files
2257 are not scanned in any specific order and on different systems the order
2258 was different. A call to "sort" has been inserted into RunGrepTest for the
2259 approprate test as a short-term fix. In the longer term there may be an
2260 alternative.
2261
2262 3. I had an email from Eric Raymond about problems translating some of PCRE's
2263 man pages to HTML (despite the fact that I distribute HTML pages, some
2264 people do their own conversions for various reasons). The problems
2265 concerned the use of low-level troff macros .br and .in. I have therefore
2266 removed all such uses from the man pages (some were redundant, some could
2267 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2268 HTML has been updated to handle .nf/.fi and to complain if it encounters
2269 .br or .in.
2270
2271 4. Updated comments in configure.ac that get placed in config.h.in and also
2272 arranged for config.h to be included in the distribution, with the name
2273 config.h.generic, for the benefit of those who have to compile without
2274 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2275
2276 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2277 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2278 makevp.bat for the current PCRE, using the additional files
2279 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2280
2281 6. A Windows user reported a minor discrepancy with test 2, which turned out
2282 to be caused by a trailing space on an input line that had got lost in his
2283 copy. The trailing space was an accident, so I've just removed it.
2284
2285 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2286 that is needed.
2287
2288 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2289 as "const" (a) because they are and (b) because it helps the PHP
2290 maintainers who have recently made a script to detect big data structures
2291 in the php code that should be moved to the .rodata section. I remembered
2292 to update Builducptable as well, so it won't revert if ucptable.h is ever
2293 re-created.
2294
2295 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2296 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2297 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2298 support is not required. This saves 1.5-2K of code, which is important in
2299 some applications.
2300
2301 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2302 so as not to refer to the tables, even though these functions will never be
2303 called when UTF-8 support is disabled. Otherwise there are problems with a
2304 shared library.
2305
2306 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2307
2308 (a) It was defining its arguments as char * instead of void *.
2309
2310 (b) It was assuming that all moves were upwards in memory; this was true
2311 a long time ago when I wrote it, but is no longer the case.
2312
2313 The emulated memove() is provided for those environments that have neither
2314 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2315 is clearly not the case, as these two bugs were recently reported.
2316
2317 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2318 and Detrail to create the HTML documentation, the .txt form of the man
2319 pages, and it removes trailing spaces from listed files. It also creates
2320 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2321 case, it wraps all the #defines with #ifndefs. This script should be run
2322 before "make dist".
2323
2324 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2325 with Unicode property support.
2326
2327 (a) For a maximizing quantifier, if the two different cases of the
2328 character were of different lengths in their UTF-8 codings (there are
2329 some cases like this - I found 11), and the matching function had to
2330 back up over a mixture of the two cases, it incorrectly assumed they
2331 were both the same length.
2332
2333 (b) When PCRE was configured to use the heap rather than the stack for
2334 recursion during matching, it was not correctly preserving the data for
2335 the other case of a UTF-8 character when checking ahead for a match
2336 while processing a minimizing repeat. If the check also involved
2337 matching a wide character, but failed, corruption could cause an
2338 erroneous result when trying to check for a repeat of the original
2339 character.
2340
2341 13. Some tidying changes to the testing mechanism:
2342
2343 (a) The RunTest script now detects the internal link size and whether there
2344 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2345 values substituted by "configure". (The RunGrepTest script already did
2346 this for UTF-8.) The configure.ac script no longer substitutes the
2347 relevant variables.
2348
2349 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2350 with length and offset values. This means that the output is different
2351 for different internal link sizes. Test 2 is skipped for link sizes
2352 other than 2 because of this, bypassing the problem. Unfortunately,
2353 there was also a test in test 3 (the locale tests) that used /B and
2354 failed for link sizes other than 2. Rather than cut the whole test out,
2355 I have added a new /Z option to pcretest that replaces the length and
2356 offset values with spaces. This is now used to make test 3 independent
2357 of link size. (Test 2 will be tidied up later.)
2358
2359 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2360 segmentation fault instead of returning the appropriate error message.
2361
2362 15. In multiline mode when the newline sequence was set to "any", the pattern
2363 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2364 This doesn't seem right; it now treats the CRLF combination as the line
2365 ending, and so does not match in that case. It's only a pattern such as ^$
2366 that would hit this one: something like ^ABC$ would have failed after \r
2367 and then tried again after \r\n.
2368
2369 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2370 in an attempt to make files that differ only in their line terminators
2371 compare equal. This works on Linux.
2372
2373 17. Under certain error circumstances pcregrep might try to free random memory
2374 as it exited. This is now fixed, thanks to valgrind.
2375
2376 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2377 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2378 was because its rules for how to advance for /g after matching an empty
2379 string at the end of a line did not allow for this case. They now check for
2380 it specially.
2381
2382 20. pcretest is supposed to handle patterns and data of any length, by
2383 extending its buffers when necessary. It was getting this wrong when the
2384 buffer for a data line had to be extended.
2385
2386 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2387 CRLF as a newline sequence.
2388
2389 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2390 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2391 I have nevertheless tidied it up.
2392
2393 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2394
2395 24. Added a man page for pcre-config.
2396
2397
2398 Version 7.0 19-Dec-06
2399 ---------------------
2400
2401 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2402 moving to gcc 4.1.1.
2403
2404 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2405 sys/time.h, which is documented as needed for this function. It doesn't
2406 seem to matter on Linux, but it showed up on some releases of OS X.
2407
2408 3. It seems that there are systems where bytes whose values are greater than
2409 127 match isprint() in the "C" locale. The "C" locale should be the
2410 default when a C program starts up. In most systems, only ASCII printing
2411 characters match isprint(). This difference caused the output from pcretest
2412 to vary, making some of the tests fail. I have changed pcretest so that:
2413
2414 (a) When it is outputting text in the compiled version of a pattern, bytes
2415 other than 32-126 are always shown as hex escapes.
2416
2417 (b) When it is outputting text that is a matched part of a subject string,
2418 it does the same, unless a different locale has been set for the match
2419 (using the /L modifier). In this case, it uses isprint() to decide.
2420
2421 4. Fixed a major bug that caused incorrect computation of the amount of memory
2422 required for a compiled pattern when options that changed within the
2423 pattern affected the logic of the preliminary scan that determines the
2424 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2425 that the computed length was too small. The symptoms of this bug were
2426 either the PCRE error "internal error: code overflow" from pcre_compile(),
2427 or a glibc crash with a message such as "pcretest: free(): invalid next
2428 size (fast)". Examples of patterns that provoked this bug (shown in
2429 pcretest format) are:
2430
2431 /(?-x: )/x
2432 /(?x)(?-x: \s*#\s*)/
2433 /((?i)[\x{c0}])/8
2434 /(?i:[\x{c0}])/8
2435
2436 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2437 is now done differently.
2438
2439 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2440 wrapper classes; (b) implement a new function in the C++ scanner that is
2441 more efficient than the old way of doing things because it avoids levels of
2442 recursion in the regex matching; (c) add a paragraph to the documentation
2443 for the FullMatch() function.
2444
2445 6. The escape sequence \n was being treated as whatever was defined as
2446 "newline". Not only was this contrary to the documentation, which states
2447 that \n is character 10 (hex 0A), but it also went horribly wrong when
2448 "newline" was defined as CRLF. This has been fixed.
2449
2450 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2451 was being set to -1 for the "end of line" case (supposedly a value that no
2452 character can have). Though this value is never used (the check for end of
2453 line is "zero bytes in current character"), it caused compiler complaints.
2454 I've changed it to 0xffffffff.
2455
2456 8. In pcre_version.c, the version string was being built by a sequence of
2457 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2458 string (as it is for production releases) called a macro with an empty
2459 argument. The C standard says the result of this is undefined. The gcc
2460 compiler treats it as an empty string (which was what was wanted) but it is
2461 reported that Visual C gives an error. The source has been hacked around to
2462 avoid this problem.
2463
2464 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2465 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2466 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2467 of them did).
2468
2469 10. Originally, pcretest opened its input and output without "b"; then I was
2470 told that "b" was needed in some environments, so it was added for release
2471 5.0 to both the input and output. (It makes no difference on Unix-like
2472 systems.) Later I was told that it is wrong for the input on Windows. I've
2473 now abstracted the modes into two macros, to make it easier to fiddle with
2474 them, and removed "b" from the input mode under Windows.
2475
2476 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2477
2478 12. Added -help and --help to pcretest as an official way of being reminded
2479 of the options.
2480
2481 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2482 and pcrecpp.cc because they annoy compilers at high warning levels.
2483
2484 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2485
2486 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2487 scripts are not C programs :-) and which was not noticed because it works
2488 on Linux.
2489
2490 16. pcretest is supposed to handle any length of pattern and data line (as one
2491 line or as a continued sequence of lines) by extending its input buffer if
2492 necessary. This feature was broken for very long pattern lines, leading to
2493 a string of junk being passed to pcre_compile() if the pattern was longer
2494 than about 50K.
2495
2496 17. I have done a major re-factoring of the way pcre_compile() computes the
2497 amount of memory needed for a compiled pattern. Previously, there was code
2498 that made a preliminary scan of the pattern in order to do this. That was
2499 OK when PCRE was new, but as the facilities have expanded, it has become
2500 harder and harder to keep it in step with the real compile phase, and there
2501 have been a number of bugs (see for example, 4 above). I have now found a
2502 cunning way of running the real compile function in a "fake" mode that
2503 enables it to compute how much memory it would need, while actually only
2504 ever using a few hundred bytes of working memory and without too many
2505 tests of the mode. This should make future maintenance and development
2506 easier. A side effect of this work is that the limit of 200 on the nesting
2507 depth of parentheses has been removed (though this was never a serious
2508 limitation, I suspect). However, there is a downside: pcre_compile() now
2509 runs more slowly than before (30% or more, depending on the pattern). I
2510 hope this isn't a big issue. There is no effect on runtime performance.
2511
2512 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2513 newline (only possible for the last line of a file) and it was a
2514 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2515
2516 19. Added additional timing features to pcretest. (1) The -tm option now times
2517 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2518 separate command line item, by a number that specifies the number of
2519 repeats to use when timing. The default is 50000; this gives better
2520 precision, but takes uncomfortably long for very large patterns.
2521
2522 20. Extended pcre_study() to be more clever in cases where a branch of a
2523 subpattern has no definite first character. For example, (a*|b*)[cd] would
2524 previously give no result from pcre_study(). Now it recognizes that the
2525 first character must be a, b, c, or d.
2526
2527 21. There was an incorrect error "recursive call could loop indefinitely" if
2528 a subpattern (or the entire pattern) that was being tested for matching an
2529 empty string contained only one non-empty item after a nested subpattern.
2530 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2531 incorrectly, because the \d was being skipped in the check.
2532
2533 22. The pcretest program now has a new pattern option /B and a command line
2534 option -b, which is equivalent to adding /B to every pattern. This causes
2535 it to show the compiled bytecode, without the additional information that
2536 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2537 is the same as /B/I).
2538
2539 23. A new optimization is now able automatically to treat some sequences such
2540 as a*b as a*+b. More specifically, if something simple (such as a character
2541 or a simple class like \d) has an unlimited quantifier, and is followed by
2542 something that cannot possibly match the quantified thing, the quantifier
2543 is automatically "possessified".
2544
2545 24. A recursive reference to a subpattern whose number was greater than 39
2546 went wrong under certain circumstances in UTF-8 mode. This bug could also
2547 have affected the operation of pcre_study().
2548
2549 25. Realized that a little bit of performance could be had by replacing
2550 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2551
2552 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2553
2554 27. Possessive quantifiers such as a++ were previously implemented by turning
2555 them into atomic groups such as ($>a+). Now they have their own opcodes,
2556 which improves performance. This includes the automatically created ones
2557 from 23 above.
2558
2559 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2560 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2561 the first matched character to be a colon. This applied both to named and
2562 numbered groups.
2563
2564 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2565
2566 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2567 building PCRE on Windows easier, so I have included it in the distribution.
2568
2569 31. There is now a check in pcretest against a ridiculously large number being
2570 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2571 loop, the loop is abandoned.
2572
2573 32. Forward references to subpatterns in conditions such as (?(2)...) where
2574 subpattern 2 is defined later cause pcre_compile() to search forwards in
2575 the pattern for the relevant set of parentheses. This search went wrong
2576 when there were unescaped parentheses in a character class, parentheses
2577 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2578
2579 33. "Subroutine" calls and backreferences were previously restricted to
2580 referencing subpatterns earlier in the regex. This restriction has now
2581 been removed.
2582
2583 34. Added a number of extra features that are going to be in Perl 5.10. On the
2584 whole, these are just syntactic alternatives for features that PCRE had
2585 previously implemented using the Python syntax or my own invention. The
2586 other formats are all retained for compatibility.
2587
2588 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2589 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2590 also .NET compatible.
2591
2592 (b) A recursion or subroutine call to a named group can now be defined as
2593 (?&name) as well as (?P>name).
2594
2595 (c) A backreference to a named group can now be defined as \k<name> or
2596 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2597 5.10, are also .NET compatible.
2598
2599 (d) A conditional reference to a named group can now use the syntax
2600 (?(<name>) or (?('name') as well as (?(name).
2601
2602 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2603 groups (named and numbered) that are never evaluated inline, but can be
2604 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2605 is always false. There may be only one alternative in such a group.
2606
2607 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2608 as the simple (?(R). The condition is true only if the most recent
2609 recursion is that of the given number or name. It does not search out
2610 through the entire recursion stack.
2611
2612 (g) The escape \gN or \g{N} has been added, where N is a positive or
2613 negative number, specifying an absolute or relative reference.
2614
2615 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2616 some "unreachable code" warnings.
2617
2618 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2619 things, this adds five new scripts.
2620
2621 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2622 There were also incompatibilities regarding the handling of \Q..\E inside
2623 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2624 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2625
2626 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2627 matches an empty string, and forcibly breaks the loop. There were bugs in
2628 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2629 against aaaa the result was just "a" rather than "aaaa", for example. Two
2630 separate and independent bugs (that affected different cases) have been
2631 fixed.
2632
2633 39. Refactored the code to abolish the use of different opcodes for small
2634 capturing bracket numbers. This is a tidy that I avoided doing when I
2635 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2636 The new approach is not only tidier, it makes it possible to reduce the
2637 memory needed to fix the previous bug (38).
2638
2639 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2640 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2641 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2642 mode.
2643
2644 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2645 report.
2646
2647 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2648 copy construction and assignment in the C++ wrapper.
2649
2650 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2651 couple of bugs that could have given wrong results in the "--newline=crlf"
2652 case.
2653
2654 44. Added a number of casts and did some reorganization of signed/unsigned int
2655 variables following suggestions from Dair Grant. Also renamed the variable
2656 "this" as "item" because it is a C++ keyword.
2657
2658 45. Arranged for dftables to add
2659
2660 #include "pcre_internal.h"
2661
2662 to pcre_chartables.c because without it, gcc 4.x may remove the array
2663 definition from the final binary if PCRE is built into a static library and
2664 dead code stripping is activated.
2665
2666 46. For an unanchored pattern, if a match attempt fails at the start of a
2667 newline sequence, and the newline setting is CRLF or ANY, and the next two
2668 characters are CRLF, advance by two characters instead of one.
2669
2670
2671 Version 6.7 04-Jul-06
2672 ---------------------
2673
2674 1. In order to handle tests when input lines are enormously long, pcretest has
2675 been re-factored so that it automatically extends its buffers when
2676 necessary. The code is crude, but this _is_ just a test program. The
2677 default size has been increased from 32K to 50K.
2678
2679 2. The code in pcre_study() was using the value of the re argument before
2680 testing it for NULL. (Of course, in any sensible call of the function, it
2681 won't be NULL.)
2682
2683 3. The memmove() emulation function in pcre_internal.h, which is used on
2684 systems that lack both memmove() and bcopy() - that is, hardly ever -
2685 was missing a "static" storage class specifier.
2686
2687 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2688 containing an extended class (one that cannot be represented by a bitmap
2689 because it contains high-valued characters or Unicode property items, e.g.
2690 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2691 pattern, but PCRE should not loop if you do not (it no longer does).
2692 [Detail: two cases were found: (a) a repeated subpattern containing an
2693 extended class; (b) a recursive reference to a subpattern that followed a
2694 previous extended class. It wasn't skipping over the extended class
2695 correctly when UTF-8 mode was not set.]
2696
2697 5. A negated single-character class was not being recognized as fixed-length
2698 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2699 compile error "lookbehind assertion is not fixed length".
2700
2701 6. The RunPerlTest auxiliary script was showing an unexpected difference
2702 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2703 write a Perl script that can interpret lines of an input file either as
2704 byte characters or as UTF-8, which is what "perltest" was being required to
2705 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2706 can't do is switch easily at run time between having the "use utf8;" pragma
2707 or not. In the end, I fudged it by using the RunPerlTest script to insert
2708 "use utf8;" explicitly for the UTF-8 tests.
2709
2710 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2711 the end of the subject string, contrary to the documentation and to what
2712 Perl does. This was true of both matching functions. Now it matches only at
2713 the start of the subject and immediately after *internal* newlines.
2714
2715 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2716 a pointer to an int instead of a pointer to an unsigned long int. This
2717 caused problems on 64-bit systems.
2718
2719 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2720 instance of the 'standard' template library not being so standard".
2721
2722 10. There was no check on the number of named subpatterns nor the maximum
2723 length of a subpattern name. The product of these values is used to compute
2724 the size of the memory block for a compiled pattern. By supplying a very
2725 long subpattern name and a large number of named subpatterns, the size
2726 computation could be caused to overflow. This is now prevented by limiting
2727 the length of names to 32 characters, and the number of named subpatterns
2728 to 10,000.
2729
2730 11. Subpatterns that are repeated with specific counts have to be replicated in
2731 the compiled pattern. The size of memory for this was computed from the
2732 length of the subpattern and the repeat count. The latter is limited to
2733 65535, but there was no limit on the former, meaning that integer overflow
2734 could in principle occur. The compiled length of a repeated subpattern is
2735 now limited to 30,000 bytes in order to prevent this.
2736
2737 12. Added the optional facility to have named substrings with the same name.
2738
2739 13. Added the ability to use a named substring as a condition, using the
2740 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2741 are numbers (not recommended). Forward references are permitted.
2742
2743 14. Added forward references in named backreferences (if you see what I mean).
2744
2745 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2746 pattern could run off the end of the subject. For example, the pattern
2747 "(?s)(.{1,5})"8 did this with the subject "ab".
2748
2749 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2750 PCRE_CASELESS was set when matching characters that were quantified with ?
2751 or *.
2752
2753 17. A character class other than a single negated character that had a minimum
2754 but no maximum quantifier - for example [ab]{6,} - was not handled
2755 correctly by pce_dfa_exec(). It would match only one character.
2756
2757 18. A valid (though odd) pattern that looked like a POSIX character
2758 class but used an invalid character after [ (for example [[,abc,]]) caused
2759 pcre_compile() to give the error "Failed: internal error: code overflow" or
2760 in some cases to crash with a glibc free() error. This could even happen if
2761 the pattern terminated after [[ but there just happened to be a sequence of
2762 letters, a binary zero, and a closing ] in the memory that followed.
2763
2764 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2765 over the years. Originally (before any Unicode support), just the bottom 8
2766 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2767 output from "man perlunicode" includes this:
2768
2769 The regular expression compiler produces polymorphic opcodes. That
2770 is, the pattern adapts to the data and automatically switches to
2771 the Unicode character scheme when presented with Unicode data--or
2772 instead uses a traditional byte scheme when presented with byte
2773 data.
2774
2775 Sadly, a wide octal escape does not cause a switch, and in a string with
2776 no other multibyte characters, these octal escapes are treated as before.
2777 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2778 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2779 Unicode string.
2780
2781 I have not perpetrated such confusion in PCRE. Up till now, it took just
2782 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2783 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2784 translate to the appropriate multibyte character.
2785
2786 29. Applied some refactoring to reduce the number of warnings from Microsoft
2787 and Borland compilers. This has included removing the fudge introduced
2788 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2789 a warning about an unused variable.
2790
2791 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2792 characters since release 4.0, because Perl (from release 5.004) does not.
2793 [Or at least, is documented not to: some releases seem to be in conflict
2794 with the documentation.] However, when a pattern was studied with
2795 pcre_study() and all its branches started with \s, PCRE still included VT
2796 as a possible starting character. Of course, this did no harm; it just
2797 caused an unnecessary match attempt.
2798
2799 22. Removed a now-redundant internal flag bit that recorded the fact that case
2800 dependency changed within the pattern. This was once needed for "required
2801 byte" processing, but is no longer used. This recovers a now-scarce options
2802 bit. Also moved the least significant internal flag bit to the most-
2803 significant bit of the word, which was not previously used (hangover from
2804 the days when it was an int rather than a uint) to free up another bit for
2805 the future.
2806
2807 23. Added support for CRLF line endings as well as CR and LF. As well as the
2808 default being selectable at build time, it can now be changed at runtime
2809 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2810 specify that it is scanning data with non-default line endings.
2811
2812 24. Changed the definition of CXXLINK to make it agree with the definition of
2813 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2814
2815 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2816 recursions. This makes a big different to stack usage for some patterns.
2817
2818 26. If a subpattern containing a named recursion or subroutine reference such
2819 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2820 the space required for the compiled pattern went wrong and gave too small a
2821 value. Depending on the environment, this could lead to "Failed: internal
2822 error: code overflow at offset 49" or "glibc detected double free or
2823 corruption" errors.
2824
2825 27. Applied patches from Google (a) to support the new newline modes and (b) to
2826 advance over multibyte UTF-8 characters in GlobalReplace.
2827
2828 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2829 difference for some implementation of PCRE in some Windows version.
2830
2831 29. Added some extra testing facilities to pcretest:
2832
2833 \q<number> in a data line sets the "match limit" value
2834 \Q<number> in a data line sets the "match recursion limt" value
2835 -S <number> sets the stack size, where <number> is in megabytes
2836
2837 The -S option isn't available for Windows.
2838
2839
2840 Version 6.6 06-Feb-06
2841 ---------------------
2842
2843 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2844 in pcreposix.h. I have copied the definition from pcre.h.
2845
2846 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2847 because pcre.h is no longer a built file.
2848
2849 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2850 not normally included in the compiled code.
2851
2852
2853 Version 6.5 01-Feb-06
2854 ---------------------
2855
2856 1. When using the partial match feature with pcre_dfa_exec(), it was not
2857 anchoring the second and subsequent partial matches at the new starting
2858 point. This could lead to incorrect results. For example, with the pattern
2859 /1234/, partially matching against "123" and then "a4" gave a match.
2860
2861 2. Changes to pcregrep:
2862
2863 (a) All non-match returns from pcre_exec() were being treated as failures
2864 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2865 error message is output. Some extra information is given for the
2866 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2867 probably the only errors that are likely to be caused by users (by
2868 specifying a regex that has nested indefinite repeats, for instance).
2869 If there are more than 20 of these errors, pcregrep is abandoned.
2870
2871 (b) A binary zero was treated as data while matching, but terminated the
2872 output line if it was written out. This has been fixed: binary zeroes
2873 are now no different to any other data bytes.
2874
2875 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2876 used to set a locale for matching. The --locale=xxxx long option has
2877 been added (no short equivalent) to specify a locale explicitly on the
2878 pcregrep command, overriding the environment variables.
2879
2880 (d) When -B was used with -n, some line numbers in the output were one less
2881 than they should have been.
2882
2883 (e) Added the -o (--only-matching) option.
2884
2885 (f) If -A or -C was used with -c (count only), some lines of context were
2886 accidentally printed for the final match.
2887
2888 (g) Added the -H (--with-filename) option.
2889
2890 (h) The combination of options -rh failed to suppress file names for files
2891 that were found from directory arguments.
2892
2893 (i) Added the -D (--devices) and -d (--directories) options.
2894
2895 (j) Added the -F (--fixed-strings) option.
2896
2897 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2898
2899 (l) Added the --colo(u)r option.
2900
2901 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2902 is not present by default.
2903
2904 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2905 items such as (?R) or (?1), when the recursion could match a number of
2906 alternatives. If it matched one of the alternatives, but subsequently,
2907 outside the recursion, there was a failure, the code tried to back up into
2908 the recursion. However, because of the way PCRE is implemented, this is not
2909 possible, and the result was an incorrect result from the match.
2910
2911 In order to prevent this happening, the specification of recursion has
2912 been changed so that all such subpatterns are automatically treated as
2913 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2914
2915 4. I had overlooked the fact that, in some locales, there are characters for
2916 which isalpha() is true but neither isupper() nor islower() are true. In
2917 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2918 and ordfeminine) are like this. This affected the treatment of \w and \W
2919 when they appeared in character classes, but not when they appeared outside
2920 a character class. The bit map for "word" characters is now created
2921 separately from the results of isalnum() instead of just taking it from the
2922 upper, lower, and digit maps. (Plus the underscore character, of course.)
2923
2924 5. The above bug also affected the handling of POSIX character classes such as
2925 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2926 permanent tables. Instead, the bit maps for such a class were previously
2927 created as the appropriate unions of the upper, lower, and digit bitmaps.
2928 Now they are created by subtraction from the [[:word:]] class, which has
2929 its own bitmap.
2930
2931 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2932 It is created by subtracting the vertical space characters (\x09, \x0a,
2933 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2934 subtraction was done in the overall bitmap for a character class, meaning
2935 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2936 be recognized. This bug has been fixed.
2937
2938 7. Patches from the folks at Google:
2939
2940 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2941 real life, but is still worth protecting against".
2942
2943 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2944 regular expressions".
2945
2946 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2947 have it.
2948
2949 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2950 "configure" and the latter not, in order to fix a problem somebody had
2951 with compiling the Arg class on HP-UX.
2952
2953 (e) Improve the error-handling of the C++ wrapper a little bit.
2954
2955 (f) New tests for checking recursion limiting.
2956
2957 8. The pcre_memmove() function, which is used only if the environment does not
2958 have a standard memmove() function (and is therefore rarely compiled),
2959 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2960 returning a result (though PCRE never actually uses the result).
2961
2962 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2963 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2964 returned instead of calling malloc() with an overflowing number that would
2965 most likely cause subsequent chaos.
2966
2967 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2968
2969 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2970 with this option is matched, the nmatch and pmatch options of regexec() are
2971 ignored.
2972
2973 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2974 provided in case anyone wants to the the POSIX interface with UTF-8
2975 strings.
2976
2977 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2978 C++ linking (needed for some HP-UX environments).
2979
2980 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2981 (unused parameter) and in the pcre_printint() function (omitted "default"
2982 switch label when the default is to do nothing).
2983
2984 15. Added some code to make it possible, when PCRE is compiled as a C++
2985 library, to replace subject pointers for pcre_exec() with a smart pointer
2986 class, thus making it possible to process discontinuous strings.
2987
2988 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2989 much the same function. They were added by different people who were trying
2990 to make PCRE easy to compile on non-Unix systems. It has been suggested
2991 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2992 for compiling on Windows systems. I have therefore replaced it with
2993 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2994 defaults to "extern" for C or "extern C" for C++, which works fine on
2995 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2996 SCOPE with something explicit in config.h. In addition:
2997
2998 (a) pcreposix.h still had just "extern" instead of either of these macros;
2999 I have replaced it with PCRE_DATA_SCOPE.
3000
3001 (b) Functions such as _pcre_xclass(), which are internal to the library,
3002 but external in the C sense, all had PCRE_EXPORT in their definitions.
3003 This is apparently wrong for the Windows case, so I have removed it.
3004 (It makes no difference on Unix-like systems.)
3005
3006 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3007 of recursive calls to match(). This is different to MATCH_LIMIT because
3008 that limits the total number of calls to match(), not all of which increase
3009 the depth of recursion. Limiting the recursion depth limits the amount of
3010 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3011 when PCRE is compiled, and changed at run time. A patch from Google adds
3012 this functionality to the C++ interface.
3013
3014 18. Changes to the handling of Unicode character properties:
3015
3016 (a) Updated the table to Unicode 4.1.0.
3017
3018 (b) Recognize characters that are not in the table as "Cn" (undefined).
3019
3020 (c) I revised the way the table is implemented to a much improved format
3021 which includes recognition of ranges. It now supports the ranges that
3022 are defined in UnicodeData.txt, and it also amalgamates other
3023 characters into ranges. This has reduced the number of entries in the
3024 table from around 16,000 to around 3,000, thus reducing its size
3025 considerably. I realized I did not need to use a tree structure after
3026 all - a binary chop search is just as efficient. Having reduced the
3027 number of entries, I extended their size from 6 bytes to 8 bytes to
3028 allow for more data.
3029
3030 (d) Added support for Unicode script names via properties such as \p{Han}.
3031
3032 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3033 matching that character.
3034
3035 20. When matching a repeated Unicode property with a minimum greater than zero,
3036 (for example \pL{2,}), PCRE could look past the end of the subject if it
3037 reached it while seeking the minimum number of characters. This could
3038 happen only if some of the characters were more than one byte long, because
3039 there is a check for at least the minimum number of bytes.
3040
3041 21. Refactored the implementation of \p and \P so as to be more general, to
3042 allow for more different types of property in future. This has changed the
3043 compiled form incompatibly. Anybody with saved compiled patterns that use
3044 \p or \P will have to recompile them.
3045
3046 22. Added "Any" and "L&" to the supported property types.
3047
3048 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3049 but give a compile time error if the value is greater than 0xff.
3050
3051 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3052 accidentally not being installed or uninstalled.
3053
3054 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3055 made were to insert the current release number. This seemed silly, because
3056 it made things harder for people building PCRE on systems that don't run
3057 "configure". I have turned pcre.h into a distributed file, no longer built
3058 by "configure", with the version identification directly included. There is
3059 no longer a pcre.h.in file.
3060
3061 However, this change necessitated a change to the pcre-config script as
3062 well. It is built from pcre-config.in, and one of the substitutions was the
3063 release number. I have updated configure.ac so that ./configure now finds
3064 the release number by grepping pcre.h.
3065
3066 26. Added the ability to run the tests under valgrind.
3067
3068
3069 Version 6.4 05-Sep-05
3070 ---------------------
3071
3072 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3073 "--" to be printed when multiple files were scanned, even when none of the
3074 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3075 consider it to be a bug, and have restored the previous behaviour.
3076
3077 2. A couple of code tidies to get rid of compiler warnings.
3078
3079 3. The pcretest program used to cheat by referring to symbols in the library
3080 whose names begin with _pcre_. These are internal symbols that are not
3081 really supposed to be visible externally, and in some environments it is
3082 possible to suppress them. The cheating is now confined to including
3083 certain files from the library's source, which is a bit cleaner.
3084
3085 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3086 file's purpose clearer.
3087
3088 5. Reorganized pcre_ucp_findchar().
3089
3090
3091 Version 6.3 15-Aug-05
3092 ---------------------
3093
3094 1. The file libpcre.pc.in did not have general read permission in the tarball.
3095
3096 2. There were some problems when building without C++ support:
3097
3098 (a) If C++ support was not built, "make install" and "make test" still
3099 tried to test it.
3100
3101 (b) There were problems when the value of CXX was explicitly set. Some
3102 changes have been made to try to fix these, and ...
3103
3104 (c) --disable-cpp can now be used to explicitly disable C++ support.
3105
3106 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3107 backslash in a target when C++ was disabled. This confuses some
3108 versions of "make", apparently. Using an intermediate variable solves
3109 this. (Same for CPP_LOBJ.)
3110
3111 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3112 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3113 necessary on certain architectures.
3114
3115 4. Added a setting of -export-symbols-regex to the link command to remove
3116 those symbols that are exported in the C sense, but actually are local
3117 within the library, and not documented. Their names all begin with
3118 "_pcre_". This is not a perfect job, because (a) we have to except some
3119 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3120 available (and never for static libraries). I have made a note to try to
3121 find a way round (a) in the future.
3122
3123
3124 Version 6.2 01-Aug-05
3125 ---------------------
3126
3127 1. There was no test for integer overflow of quantifier values. A construction
3128 such as {1111111111111111} would give undefined results. What is worse, if
3129 a minimum quantifier for a parenthesized subpattern overflowed and became
3130 negative, the calculation of the memory size went wrong. This could have
3131 led to memory overwriting.
3132
3133 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3134
3135 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3136 operating environments where this matters.
3137
3138 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3139 PCRE options from within the C++ wrapper.
3140
3141 5. Named capturing subpatterns were not being correctly counted when a pattern
3142 was compiled. This caused two problems: (a) If there were more than 100
3143 such subpatterns, the calculation of the memory needed for the whole
3144 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3145 back references of the form \12, where the number was greater than 9, were
3146 not recognized as back references, even though there were sufficient
3147 previous subpatterns.
3148
3149 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3150 versions of gcc, e.g. 2.95.4.
3151
3152
3153 Version 6.1 21-Jun-05
3154 ---------------------
3155
3156 1. There was one reference to the variable "posix" in pcretest.c that was not
3157 surrounded by "#if !defined NOPOSIX".
3158
3159 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3160 the cross-check on the old pcre_info() function, for the benefit of the
3161 cut-down version of PCRE that is currently imported into Exim.
3162
3163 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3164 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3165 patterns that start (?i) but I don't think that matters. The use of (?i) is
3166 just an example; this all applies to the other options as well.
3167
3168 4. Since libtool seems to echo the compile commands it is issuing, the output
3169 from "make" can be reduced a bit by putting "@" in front of each libtool
3170 compile command.
3171
3172 5. Patch from the folks at Google for configure.in to be a bit more thorough
3173 in checking for a suitable C++ installation before trying to compile the
3174 C++ stuff. This should fix a reported problem when a compiler was present,
3175 but no suitable headers.
3176
3177 6. The man pages all had just "PCRE" as their title. I have changed them to
3178 be the relevant file name. I have also arranged that these names are
3179 retained in the file doc/pcre.txt, which is a concatenation in text format
3180 of all the man pages except the little individual ones for each function.
3181
3182 7. The NON-UNIX-USE file had not been updated for the different set of source
3183 files that come with release 6. I also added a few comments about the C++
3184 wrapper.
3185
3186
3187 Version 6.0 07-Jun-05
3188 ---------------------
3189
3190 1. Some minor internal re-organization to help with my DFA experiments.
3191
3192 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3193 didn't matter for the library itself when fully configured, but did matter
3194 when compiling without UCP support, or within Exim, where the ucp files are
3195 not imported.
3196
3197 3. Refactoring of the library code to split up the various functions into
3198 different source modules. The addition of the new DFA matching code (see
3199 below) to a single monolithic source would have made it really too
3200 unwieldy, quite apart from causing all the code to be include in a
3201 statically linked application, when only some functions are used. This is
3202 relevant even without the DFA addition now that patterns can be compiled in
3203 one application and matched in another.
3204
3205 The downside of splitting up is that there have to be some external
3206 functions and data tables that are used internally in different modules of
3207 the library but which are not part of the API. These have all had their
3208 names changed to start with "_pcre_" so that they are unlikely to clash
3209 with other external names.
3210
3211 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3212 a different (DFA) algorithm. Although it is slower than the original
3213 function, it does have some advantages for certain types of matching
3214 problem.
3215
3216 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3217 including restarting after a partial match.
3218
3219 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3220 defined when compiling for Windows was sent to me. I have put it into the
3221 code, though I have no means of testing or verifying it.
3222
3223 7. Added the pcre_refcount() auxiliary function.
3224
3225 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3226 match before or at the first newline in the subject string. In pcretest,
3227 the /f option on a pattern can be used to set this.
3228
3229 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3230 would behave wrongly. This has been present in PCRE since release 4.0.
3231
3232 10. A number of changes to the pcregrep command:
3233
3234 (a) Refactored how -x works; insert ^(...)$ instead of setting
3235 PCRE_ANCHORED and checking the length, in preparation for adding
3236 something similar for -w.
3237
3238 (b) Added the -w (match as a word) option.
3239
3240 (c) Refactored the way lines are read and buffered so as to have more
3241 than one at a time available.
3242
3243 (d) Implemented a pcregrep test script.
3244
3245 (e) Added the -M (multiline match) option. This allows patterns to match
3246 over several lines of the subject. The buffering ensures that at least
3247 8K, or the rest of the document (whichever is the shorter) is available
3248 for matching (and similarly the previous 8K for lookbehind assertions).
3249
3250 (f) Changed the --help output so that it now says
3251
3252 -w, --word-regex(p)
3253
3254 instead of two lines, one with "regex" and the other with "regexp"
3255 because that confused at least one person since the short forms are the
3256 same. (This required a bit of code, as the output is generated
3257 automatically from a table. It wasn't just a text change.)
3258
3259 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3260 option but starts with a hyphen. Could be a pattern or a path name
3261 starting with a hyphen, for instance.
3262
3263 (h) "-" can be given as a file name to represent stdin.
3264
3265 (i) When file names are being printed, "(standard input)" is used for
3266 the standard input, for compatibility with GNU grep. Previously
3267 "<stdin>" was used.
3268
3269 (j) The option --label=xxx can be used to supply a name to be used for
3270 stdin when file names are being printed. There is no short form.
3271
3272 (k) Re-factored the options decoding logic because we are going to add
3273 two more options that take data. Such options can now be given in four
3274 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3275
3276 (l) Added the -A, -B, and -C options for requesting that lines of context
3277 around matches be printed.
3278
3279 (m) Added the -L option to print the names of files that do not contain
3280 any matching lines, that is, the complement of -l.
3281
3282 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3283 continue to scan other files.
3284
3285 (o) The -s option was incorrectly implemented. For compatibility with other
3286 greps, it now suppresses the error message for a non-existent or non-
3287 accessible file (but not the return code). There is a new option called
3288 -q that suppresses the output of matching lines, which was what -s was
3289 previously doing.
3290
3291 (p) Added --include and --exclude options to specify files for inclusion
3292 and exclusion when recursing.
3293
3294 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3295 Hopefully, it now does.
3296
3297 12. Missing cast in pcre_study().
3298
3299 13. Added an "uninstall" target to the makefile.
3300
3301 14. Replaced "extern" in the function prototypes in Makefile.in with
3302 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3303 world, but is set differently for Windows.
3304
3305 15. Added a second compiling function called pcre_compile2(). The only
3306 difference is that it has an extra argument, which is a pointer to an
3307 integer error code. When there is a compile-time failure, this is set
3308 non-zero, in addition to the error test pointer being set to point to an
3309 error message. The new argument may be NULL if no error number is required
3310 (but then you may as well call pcre_compile(), which is now just a
3311 wrapper). This facility is provided because some applications need a
3312 numeric error indication, but it has also enabled me to tidy up the way
3313 compile-time errors are handled in the POSIX wrapper.
3314
3315 16. Added VPATH=.libs to the makefile; this should help when building with one
3316 prefix path and installing with another. (Or so I'm told by someone who
3317 knows more about this stuff than I do.)
3318
3319 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3320 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3321 match everything, including newlines. This is not POSIX-compatible, but
3322 somebody wanted the feature. From pcretest it can be activated by using
3323 both the P and the s flags.
3324
3325 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3326
3327 19. libpcre.pc was being incorrectly installed as executable.
3328
3329 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3330 it now also looks for '\r' so that it will work unmodified on Windows.
3331
3332 21. Added Google's contributed C++ wrapper to the distribution.
3333
3334 22. Added some untidy missing memory free() calls in pcretest, to keep
3335 Electric Fence happy when testing.
3336
3337
3338
3339 Version 5.0 13-Sep-04
3340 ---------------------
3341
3342 1. Internal change: literal characters are no longer packed up into items
3343 containing multiple characters in a single byte-string. Each character
3344 is now matched using a separate opcode. However, there may be more than one
3345 byte in the character in UTF-8 mode.
3346
3347 2. The pcre_callout_block structure has two new fields: pattern_position and
3348 next_item_length. These contain the offset in the pattern to the next match
3349 item, and its length, respectively.
3350
3351 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3352 insertion of callouts before each pattern item. Added the /C option to
3353 pcretest to make use of this.
3354
3355 4. On the advice of a Windows user, the lines
3356
3357 #if defined(_WIN32) || defined(WIN32)
3358 _setmode( _fileno( stdout ), 0x8000 );
3359 #endif /* defined(_WIN32) || defined(WIN32) */
3360
3361 have been added to the source of pcretest. This apparently does useful
3362 magic in relation to line terminators.
3363
3364 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3365 for the benefit of those environments where the "b" makes a difference.
3366
3367 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3368 to know about it. I have put a hack into configure.in that adds in code
3369 to set GCC=yes if CC=icc. This seems to end up at a point in the
3370 generated configure script that is early enough to affect the setting of
3371 compiler options, which is what is needed, but I have no means of testing
3372 whether it really works. (The user who reported this had patched the
3373 generated configure script, which of course I cannot do.)
3374
3375 LATER: After change 22 below (new libtool files), the configure script
3376 seems to know about icc (and also ecc). Therefore, I have commented out
3377 this hack in configure.in.
3378
3379 7. Added support for pkg-config (2 patches were sent in).
3380
3381 8. Negated POSIX character classes that used a combination of internal tables
3382 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3383 [[:^ascii]]. Typically, they would match almost any characters. The other
3384 POSIX classes were not broken in this way.
3385
3386 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3387 to find the match, as PCRE was deluded into thinking that the match had to
3388 start at the start point or following a newline. The same bug applied to
3389 patterns with negative forward assertions or any backward assertions
3390 preceding ".*" at the start, unless the pattern required a fixed first
3391 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3392
3393 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3394 starting at the last subject character, bytes beyond the end of the subject
3395 string were read.
3396
3397 11. Renamed the variable "class" as "classbits" to make life easier for C++
3398 users. (Previously there was a macro definition, but it apparently wasn't
3399 enough.)
3400
3401 12. Added the new field "tables" to the extra data so that tables can be passed
3402 in at exec time, or the internal tables can be re-selected. This allows
3403 a compiled regex to be saved and re-used at a later time by a different
3404 program that might have everything at different addresses.
3405
3406 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3407 -R library as well as a -L library.
3408
3409 14. The debugging options of pcretest (-d on the command line or D on a
3410 pattern) showed incorrect output for anything following an extended class
3411 that contained multibyte characters and which was followed by a quantifier.
3412
3413 15. Added optional support for general category Unicode character properties
3414 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3415 support. It adds about 90K to the size of the library. The meanings of the
3416 inbuilt class escapes such as \d and \s have NOT been changed.
3417
3418 16. Updated pcredemo.c to include calls to free() to release the memory for the
3419 compiled pattern.
3420
3421 17. The generated file chartables.c was being created in the source directory
3422 instead of in the building directory. This caused the build to fail if the
3423 source directory was different from the building directory, and was
3424 read-only.
3425
3426 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3427 file. No doubt somebody will tell me if they don't make sense... Also added
3428 Dan Mooney's comments about building on OpenVMS.
3429
3430 19. Added support for partial matching via the PCRE_PARTIAL option for
3431 pcre_exec() and the \P data escape in pcretest.
3432
3433 20. Extended pcretest with 3 new pattern features:
3434
3435 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3436 write the compiled pattern to the file whose name is "rest-of-line".
3437 This is a straight binary dump of the data, with the saved pointer to
3438 the character tables forced to be NULL. The study data, if any, is
3439 written too. After writing, pcretest reads a new pattern.
3440
3441 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3442 compiled pattern from the given file. There must not be any
3443 occurrences of "<" in the file name (pretty unlikely); if there are,
3444 pcretest will instead treat the initial "<" as a pattern delimiter.
3445 After reading in the pattern, pcretest goes on to read data lines as
3446 usual.
3447
3448 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3449 and 16-bit fields in a compiled pattern, to simulate a pattern that
3450 was compiled on a host of opposite endianness.
3451
3452 21. The pcre-exec() function can now cope with patterns that were compiled on
3453 hosts of opposite endianness, with this restriction:
3454
3455 As for any compiled expression that is saved and used later, the tables
3456 pointer field cannot be preserved; the extra_data field in the arguments
3457 to pcre_exec() should be used to pass in a tables address if a value
3458 other than the default internal tables were used at compile time.
3459
3460 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3461 now diagnosed as an error. Previously, most of the time, a negative number
3462 would have been treated as zero, but if in addition "ovector" was passed as
3463 NULL, a crash could occur.
3464
3465 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3466 new versions from the libtool 1.5 distribution (the last one is a copy of
3467 a file called libtool.m4). This seems to have fixed the need to patch
3468 "configure" to support Darwin 1.3 (which I used to do). However, I still
3469 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3470 workstation).
3471
3472 24. Changed the PCRE licence to be the more standard "BSD" licence.
3473
3474
3475 Version 4.5 01-Dec-03
3476 ---------------------
3477
3478 1. There has been some re-arrangement of the code for the match() function so
3479 that it can be compiled in a version that does not call itself recursively.
3480 Instead, it keeps those local variables that need separate instances for
3481 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3482 needs to "recurse". Keeping track of where control must go is done by means
3483 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3484 hide most of the details from the main code, and operates only if
3485 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3486 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3487 operating.
3488
3489 To make it easier for callers to provide specially tailored get/free
3490 functions for this usage, two new functions, pcre_stack_malloc, and
3491 pcre_stack_free, are used. They are always called in strict stacking order,
3492 and the size of block requested is always the same.
3493
3494 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3495 PCRE has been compiled to use the stack or the heap for recursion. The
3496 -C option of pcretest uses this to show which version is compiled.
3497
3498 A new data escape \S, is added to pcretest; it causes the amounts of store
3499 obtained and freed by both kinds of malloc/free at match time to be added
3500 to the output.
3501
3502 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3503 what's available on my current Linux desktop machine.
3504
3505 3. When matching a UTF-8 string, the test for a valid string at the start has
3506 been extended. If start_offset is not zero, PCRE now checks that it points
3507 to a byte that is the start of a UTF-8 character. If not, it returns
3508 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3509 this is necessary because there may be backward assertions in the pattern.
3510 When matching the same subject several times, it may save resources to use
3511 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3512
3513 4. The code for checking the validity of UTF-8 strings has been tightened so
3514 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3515 containing "overlong sequences".
3516
3517 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3518 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3519 should have been "&", but it just so happened that all the cases this let
3520 through by mistake were picked up later in the function.
3521
3522 6. I had used a variable called "isblank" - this is a C99 function, causing
3523 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3524
3525 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3526 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3527 so the version is shown; (c) stop "make test" echoing "./RunTest".
3528
3529 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3530
3531 9. The prototype for memmove() for systems that don't have it was using
3532 size_t, but the inclusion of the header that defines size_t was later. I've
3533 moved the #includes for the C headers earlier to avoid this.
3534
3535 10. Added some adjustments to the code to make it easier to compiler on certain
3536 special systems:
3537
3538 (a) Some "const" qualifiers were missing.
3539 (b) Added the macro EXPORT before all exported functions; by default this
3540 is defined to be empty.
3541 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3542 that it reads its output file name as an argument instead of writing
3543 to the standard output and assuming this can be redirected.
3544
3545 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3546 class containing characters with values greater than 255, PCRE compilation
3547 went into a loop.
3548
3549 12. A recursive reference to a subpattern that was within another subpattern
3550 that had a minimum quantifier of zero caused PCRE to crash. For example,
3551 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3552 recursion. If the recursively-called subpattern itself had a zero repeat,
3553 that was OK.
3554
3555 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3556 buffer into which it was copied (for escape processing) was still set at
3557 1024, so long lines caused crashes.
3558
3559 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3560 "internal error: code overflow...". This applied to any character class
3561 that was followed by a possessive quantifier.
3562
3563 15. Modified the Makefile to add libpcre.la as a prerequisite for
3564 libpcreposix.la because I was told this is needed for a parallel build to
3565 work.
3566
3567 16. If a pattern that contained .* following optional items at the start was
3568 studied, the wrong optimizing data was generated, leading to matching
3569 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3570 matching string must start with a or b or c. The correct conclusion for
3571 this pattern is that a match can start with any character.
3572
3573
3574 Version 4.4 13-Aug-03
3575 ---------------------
3576
3577 1. In UTF-8 mode, a character class containing characters with values between
3578 127 and 255 was not handled correctly if the compiled pattern was studied.
3579 In fixing this, I have also improved the studying algorithm for such
3580 classes (slightly).
3581
3582 2. Three internal functions had redundant arguments passed to them. Removal
3583 might give a very teeny performance improvement.
3584
3585 3. Documentation bug: the value of the capture_top field in a callout is *one
3586 more than* the number of the hightest numbered captured substring.
3587
3588 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3589 in incorrectly linking with a previously installed version. They now link
3590 explicitly with libpcre.la.
3591
3592 5. configure.in no longer needs to recognize Cygwin specially.
3593
3594 6. A problem in pcre.in for Windows platforms is fixed.
3595
3596 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3597 pcretest, it used to include the size of the study block as part of its
3598 output. Unfortunately, the structure contains a field that has a different
3599 size on different hardware architectures. This meant that the tests that
3600 showed this size failed. As the block is currently always of a fixed size,
3601 this information isn't actually particularly useful in pcretest output, so
3602 I have just removed it.
3603
3604 8. Three pre-processor statements accidentally did not start in column 1.
3605 Sadly, there are *still* compilers around that complain, even though
3606 standard C has not required this for well over a decade. Sigh.
3607
3608 9. In pcretest, the code for checking callouts passed small integers in the
3609 callout_data field, which is a void * field. However, some picky compilers
3610 complained about the casts involved for this on 64-bit systems. Now
3611 pcretest passes the address of the small integer instead, which should get
3612 rid of the warnings.
3613
3614 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3615 both compile and run time, and gives an error if an invalid UTF-8 sequence
3616 is found. There is a option for disabling this check in cases where the
3617 string is known to be correct and/or the maximum performance is wanted.
3618
3619 11. In response to a bug report, I changed one line in Makefile.in from
3620
3621 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3622 to
3623 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3624
3625 to look similar to other lines, but I have no way of telling whether this
3626 is the right thing to do, as I do not use Windows. No doubt I'll get told
3627 if it's wrong...
3628
3629
3630 Version 4.3 21-May-03
3631 ---------------------
3632
3633 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3634 Makefile.
3635
3636 2. Some refactoring to improve the quality of the code:
3637
3638 (i) The utf8_table... variables are now declared "const".
3639
3640 (ii) The code for \cx, which used the "case flipping" table to upper case
3641 lower case letters, now just substracts 32. This is ASCII-specific,
3642 but the whole concept of \cx is ASCII-specific, so it seems
3643 reasonable.
3644
3645 (iii) PCRE was using its character types table to recognize decimal and
3646 hexadecimal digits in the pattern. This is silly, because it handles
3647 only 0-9, a-f, and A-F, but the character types table is locale-
3648 specific, which means strange things might happen. A private
3649 table is now used for this - though it costs 256 bytes, a table is
3650 much faster than multiple explicit tests. Of course, the standard
3651 character types table is still used for matching digits in subject
3652 strings against \d.
3653
3654 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3655 ending in _t are). So I've renamed it as ESC_tee.
3656
3657 3. The first argument for regexec() in the POSIX wrapper should have been
3658 defined as "const".
3659
3660 4. Changed pcretest to use malloc() for its buffers so that they can be
3661 Electric Fenced for debugging.
3662
3663 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3664 to read one or more bytes before the start of the subject string. Often this
3665 had no effect on PCRE's behaviour, but in some circumstances it could
3666 provoke a segmentation fault.
3667
3668 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3669 to try to read one or more bytes before the start of the subject string.
3670
3671 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3672 UTF-8 support could misbehave in various ways if the subject string
3673 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3674 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3675 back over UTF-8 characters.)
3676
3677
3678 Version 4.2 14-Apr-03
3679 ---------------------
3680
3681 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3682
3683 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3684 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3685 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3686 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3687 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3688 and BUILD_EXEEXT
3689 Note: automatic setting of the BUILD variables is not yet working
3690 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3691 compile-time but not at link-time
3692 [LINK]: use for linking executables only
3693 make different versions for Windows and non-Windows
3694 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3695 libraries
3696 [LINK_FOR_BUILD]: new variable
3697 [OBJEXT]: use throughout
3698 [EXEEXT]: use throughout
3699 <winshared>: new target
3700 <wininstall>: new target
3701 <dftables.o>: use native compiler
3702 <dftables>: use native linker
3703 <install>: handle Windows platform correctly
3704 <clean>: ditto
3705 <check>: ditto
3706 copy DLL to top builddir before testing
3707
3708 As part of these changes, -no-undefined was removed again. This was reported
3709 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3710 in any case.
3711
3712 3. Some tidies to get rid of compiler warnings:
3713
3714 . In the match_data structure, match_limit was an unsigned long int, whereas
3715 match_call_count was an int. I've made them both unsigned long ints.
3716
3717 . In pcretest the fact that a const uschar * doesn't automatically cast to
3718 a void * provoked a warning.
3719
3720 . Turning on some more compiler warnings threw up some "shadow" variables
3721 and a few more missing casts.
3722
3723 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3724 option, a class that contained a single character with a value between 128
3725 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3726
3727 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3728 option, a class that contained several characters, but with at least one
3729 whose value was between 128 and 255 caused PCRE to crash.
3730
3731
3732 Version 4.1 12-Mar-03
3733 ---------------------
3734
3735 1. Compiling with gcc -pedantic found a couple of places where casts were
3736 needed, and a string in dftables.c that was longer than standard compilers are
3737 required to support.
3738
3739 2. Compiling with Sun's compiler found a few more places where the code could
3740 be tidied up in order to avoid warnings.
3741
3742 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3743 first of these names is deprecated in the latest Autoconf in favour of the name
3744 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3745 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3746 analogy I have changed it to CFLAGS_FOR_BUILD.
3747
3748 4. Added -no-undefined to the linking command in the Makefile, because this is
3749 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3750 linking step for the pcreposix library.
3751
3752 5. PCRE was failing to diagnose the case of two named groups with the same
3753 name.
3754
3755 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3756 literal character that is needed in the subject for a match, and scans along to
3757 ensure that it is present before embarking on the full matching process. This
3758 saves time in cases of nested unlimited repeats that are never going to match.
3759 Problem: the scan can take a lot of time if the subject is very long (e.g.
3760 megabytes), thus penalizing straightforward matches. It is now done only if the
3761 amount of subject to be scanned is less than 1000 bytes.
3762
3763 7. A lesser problem with the same optimization is that it was recording the
3764 first character of an anchored pattern as "needed", thus provoking a search
3765 right along the subject, even when the first match of the pattern was going to
3766 fail. The "needed" character is now not set for anchored patterns, unless it
3767 follows something in the pattern that is of non-fixed length. Thus, it still
3768 fulfils its original purpose of finding quick non-matches in cases of nested
3769 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3770
3771
3772 Version 4.0 17-Feb-03
3773 ---------------------
3774
3775 1. If a comment in an extended regex that started immediately after a meta-item
3776 extended to the end of string, PCRE compiled incorrect data. This could lead to
3777 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3778
3779 2. Moved to autoconf 2.53 and libtool 1.4.2.
3780
3781 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3782 the special perltest8 script is no longer needed - all the tests can be run
3783 from a single perltest script.
3784
3785 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3786 by \s. It has now been removed in PCRE. This means it isn't recognized as
3787 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3788 class [:space:] *does* include VT, thereby creating a mess.
3789
3790 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3791 space and tab.
3792
3793 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3794 its new features into the main test script, reducing the number of scripts.
3795
3796 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3797 were backward compatible, and made the (?i) apply to the whole pattern, as if
3798 /i were given. Now it behaves more logically, and applies the option setting
3799 only to what follows. PCRE has been changed to follow suit. However, if it
3800 finds options settings right at the start of the pattern, it extracts them into
3801 the global options, as before. Thus, they show up in the info data.
3802
3803 8. Added support for the \Q...\E escape sequence. Characters in between are
3804 treated as literals. This is slightly different from Perl in that $ and @ are
3805 also handled as literals inside the quotes. In Perl, they will cause variable
3806 interpolation. Note the following examples:
3807
3808 Pattern PCRE matches Perl matches
3809
3810 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3811 \Qabc\$xyz\E abc\$xyz abc\$xyz
3812 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3813
3814 For compatibility with Perl, \Q...\E sequences are recognized inside character
3815 classes as well as outside them.
3816
3817 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3818 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3819 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3820 signed/unsigned warnings.
3821
3822 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3823 option for pcretest, so I've replaced it by a simple function that does just
3824 that job.
3825
3826 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3827 "pcregrep -".
3828
3829 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3830 Java package. This provides some syntactic sugar for simple cases of what my
3831 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3832 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3833 item, you can use this simplified notation. Note that only makes sense with
3834 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3835 greediness, whatever the setting of the PCRE_UNGREEDY option.
3836
3837 13. A change of greediness default within a pattern was not taking effect at
3838 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3839 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3840 was abstracted outside.
3841
3842 14. PCRE now supports the \G assertion. It is true when the current matching
3843 position is at the start point of the match. This differs from \A when the
3844 starting offset is non-zero. Used with the /g option of pcretest (or similar
3845 code), it works in the same way as it does for Perl's /g option. If all
3846 alternatives of a regex begin with \G, the expression is anchored to the start
3847 match position, and the "anchored" flag is set in the compiled expression.
3848
3849 15. Some bugs concerning the handling of certain option changes within patterns
3850 have been fixed. These applied to options other than (?ims). For example,
3851 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3852 been the other way round. Some of this was related to change 7 above.
3853
3854 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3855 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3856 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3857 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3858
3859 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3860 mode. Unlike ".", it always matches newline, whatever the setting of
3861 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3862 assertions. Perl allows it, but it doesn't (in general) work because it can't
3863 calculate the length of the lookbehind. At least, that's the case for Perl
3864 5.8.0 - I've been told they are going to document that it doesn't work in
3865 future.
3866
3867 18. Added an error diagnosis for escapes that PCRE does not support: these are
3868 \L, \l, \N, \P, \p, \U, \u, and \X.
3869
3870 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3871 reading past the end of the pattern in cases such as /[abcd/.
3872
3873 20. PCRE was getting more memory than necessary for patterns with classes that
3874 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3875
3876 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3877 compiling PCRE for use with Virtual Pascal.
3878
3879 22. Small fix to the Makefile to make it work properly if the build is done
3880 outside the source tree.
3881
3882 23. Added a new extension: a condition to go with recursion. If a conditional
3883 subpattern starts with (?(R) the "true" branch is used if recursion has
3884 happened, whereas the "false" branch is used only at the top level.
3885
3886 24. When there was a very long string of literal characters (over 255 bytes
3887 without UTF support, over 250 bytes with UTF support), the computation of how
3888 much memory was required could be incorrect, leading to segfaults or other
3889 strange effects.
3890
3891 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3892 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3893 there was a subsequent back reference to those brackets. This meant that, for
3894 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3895 possible to check for precisely this case. All we can do is abandon the
3896 optimization if .* occurs inside capturing brackets when there are any back
3897 references whatsoever. (See below for a better fix that came later.)
3898
3899 26. The handling of the optimization for finding the first character of a
3900 non-anchored pattern, and for finding a character that is required later in the
3901 match were failing in some cases. This didn't break the matching; it just
3902 failed to optimize when it could. The way this is done has been re-implemented.
3903
3904 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3905
3906 28. Added a new feature that provides some of the functionality that Perl
3907 provides with (?{...}). The facility is termed a "callout". The way it is done
3908 in PCRE is for the caller to provide an optional function, by setting
3909 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3910 global variable. By default it is unset, which disables all calling out. To get
3911 the function called, the regex must include (?C) at appropriate points. This
3912 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3913 This provides a means of identifying different callout points. When PCRE
3914 reaches such a point in the regex, if pcre_callout has been set, the external
3915 function is called. It is provided with data in a structure called
3916 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3917 matching continues; if it returns a non-zero value, the match at the current
3918 point fails. However, backtracking will occur if possible. [This was changed
3919 later and other features added - see item 49 below.]
3920
3921 29. pcretest is upgraded to test the callout functionality. It provides a
3922 callout function that displays information. By default, it shows the start of
3923 the match and the current position in the text. There are some new data escapes
3924 to vary what happens:
3925
3926 \C+ in addition, show current contents of captured substrings
3927 \C- do not supply a callout function
3928 \C!n return 1 when callout number n is reached
3929 \C!n!m return 1 when callout number n is reached for the mth time
3930
3931 30. If pcregrep was called with the -l option and just a single file name, it
3932 output "<stdin>" if a match was found, instead of the file name.
3933
3934 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3935 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3936 pcre_exec(). This saves a malloc/free per call. The default value of
3937 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3938 when configuring.
3939
3940 32. The default maximum size of a compiled pattern is 64K. There have been a
3941 few cases of people hitting this limit. The code now uses macros to handle the
3942 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3943 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3944 configuring. Tests 2 and 5 work only with 2-byte links because they output
3945 debugging information about compiled patterns.
3946
3947 33. Internal code re-arrangements:
3948
3949 (a) Moved the debugging function for printing out a compiled regex into
3950 its own source file (printint.c) and used #include to pull it into
3951 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3952 separate copies.
3953
3954 (b) Defined the list of op-code names for debugging as a macro in
3955 internal.h so that it is next to the definition of the opcodes.
3956
3957 (c) Defined a table of op-code lengths for simpler skipping along compiled
3958 code. This is again a macro in internal.h so that it is next to the
3959 definition of the opcodes.
3960
3961 34. Added support for recursive calls to individual subpatterns, along the
3962 lines of Robin Houston's patch (but implemented somewhat differently).
3963
3964 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3965 allow it to read and process whole directories in Win32. This code was
3966 contributed by Lionel Fourquaux; it has not been tested by me.
3967
3968 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3969 used to name a group. Names consist of alphanumerics and underscores, and must
3970 be unique. Back references use the syntax (?P=name) and recursive calls use
3971 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3972 numbers. The function pcre_fullinfo() can be used after compilation to extract
3973 a name/number map. There are three relevant calls:
3974
3975 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3976 PCRE_INFO_NAMECOUNT yields the number of entries
3977 PCRE_INFO_NAMETABLE yields a pointer to the map.
3978
3979 The map is a vector of fixed-size entries. The size of each entry depends on
3980 the length of the longest name used. The first two bytes of each entry are the
3981 group number, most significant byte first. There follows the corresponding
3982 name, zero terminated. The names are in alphabetical order.
3983
3984 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3985 case instead of 255. Making it the same both with and without UTF-8 support
3986 means that the same test output works with both.
3987
3988 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3989 calling malloc() with a zero argument.
3990
3991 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3992 optimization. I've improved things by keeping a bitmap of backreferences with
3993 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3994 fact referenced, the optimization can be applied. It is unlikely that a
3995 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3996 the match to follow \n) will appear inside brackets with a number greater than
3997 31, but if it does, any back reference > 31 suppresses the optimization.
3998
3999 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4000 of disabling numbered capturing parentheses. Any opening parenthesis that is
4001 not followed by ? behaves as if it were followed by ?: but named parentheses
4002 can still be used for capturing (and they will acquire numbers in the usual
4003 way).
4004
4005 41. Redesigned the return codes from the match() function into yes/no/error so
4006 that errors can be passed back from deep inside the nested calls. A malloc
4007 failure while inside a recursive subpattern call now causes the
4008 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4009
4010 42. It is now possible to set a limit on the number of times the match()
4011 function is called in a call to pcre_exec(). This facility makes it possible to
4012 limit the amount of recursion and backtracking, though not in a directly
4013 obvious way, because the match() function is used in a number of different
4014 circumstances. The count starts from zero for each position in the subject
4015 string (for non-anchored patterns). The default limit is, for compatibility, a
4016 large number, namely 10 000 000. You can change this in two ways:
4017
4018 (a) When configuring PCRE before making, you can use --with-match-limit=n
4019 to set a default value for the compiled library.
4020
4021 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4022 a different value is set. See 45 below.
4023
4024 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4025
4026 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4027 of things that can be changed at compile time. The first argument specifies
4028 what is wanted and the second points to where the information is to be placed.
4029 The current list of available information is:
4030
4031 PCRE_CONFIG_UTF8
4032
4033 The output is an integer that is set to one if UTF-8 support is available;
4034 otherwise it is set to zero.
4035
4036 PCRE_CONFIG_NEWLINE
4037
4038 The output is an integer that it set to the value of the code that is used for
4039 newline. It is either LF (10) or CR (13).
4040
4041 PCRE_CONFIG_LINK_SIZE
4042
4043 The output is an integer that contains the number of bytes used for internal
4044 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4045
4046 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4047
4048 The output is an integer that contains the threshold above which the POSIX
4049 interface uses malloc() for output vectors. See item 31 above.
4050
4051 PCRE_CONFIG_MATCH_LIMIT
4052
4053 The output is an unsigned integer that contains the default limit of the number
4054 of match() calls in a pcre_exec() execution. See 42 above.
4055
4056 44. pcretest has been upgraded by the addition of the -C option. This causes it
4057 to extract all the available output from the new pcre_config() function, and to
4058 output it. The program then exits immediately.
4059
4060 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4061 order to support additional features. One way would have been to define
4062 pcre_exec2() (for example) with extra arguments, but this would not have been
4063 extensible, and would also have required all calls to the original function to
4064 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4065 is used for passing in "extra" data from pcre_study().
4066
4067 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4068 contains the following fields:
4069
4070 flags a bitmap indicating which of the following fields are set
4071 study_data opaque data from pcre_study()
4072 match_limit a way of specifying a limit on match() calls for a specific
4073 call to pcre_exec()
4074 callout_data data for callouts (see 49 below)
4075
4076 The flag bits are also defined in pcre.h, and are
4077
4078 PCRE_EXTRA_STUDY_DATA
4079 PCRE_EXTRA_MATCH_LIMIT
4080 PCRE_EXTRA_CALLOUT_DATA
4081
4082 The pcre_study() function now returns one of these new pcre_extra blocks, with
4083 the actual study data pointed to by the study_data field, and the
4084 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4085 before. That is, this change is entirely upwards-compatible and requires no
4086 change to existing code.
4087
4088 If you want to pass in additional data to pcre_exec(), you can either place it
4089 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4090 block.
4091
4092 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4093 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4094 times with different match limits, until it finds the minimum value needed for
4095 pcre_exec() to complete. The value is then output. This can be instructive; for
4096 most simple matches the number is quite small, but for pathological cases it
4097 gets very large very quickly.
4098
4099 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4100 returns the size of the data block pointed to by the study_data field in a
4101 pcre_extra block, that is, the value that was passed as the argument to
4102 pcre_malloc() when PCRE was getting memory in which to place the information
4103 created by pcre_study(). The fourth argument should point to a size_t variable.
4104 pcretest has been extended so that this information is shown after a successful
4105 pcre_study() call when information about the compiled regex is being displayed.
4106
4107 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4108 because what follows is always an absolute path. (Later: it turns out that this
4109 is more than cosmetic for MinGW, because it doesn't like empty path
4110 components.)
4111
4112 49. Some changes have been made to the callout feature (see 28 above):
4113
4114 (i) A callout function now has three choices for what it returns:
4115
4116 0 => success, carry on matching
4117 > 0 => failure at this point, but backtrack if possible
4118 < 0 => serious error, return this value from pcre_exec()
4119
4120 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4121 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4122 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4123 use by callout functions. It will never be used by PCRE itself.
4124
4125 (ii) The pcre_extra structure (see 45 above) has a void * field called
4126 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4127 pcre_callout_block structure has a field of the same name. The contents of
4128 the field passed in the pcre_extra structure are passed to the callout
4129 function in the corresponding field in the callout block. This makes it
4130 easier to use the same callout-containing regex from multiple threads. For
4131 testing, the pcretest program has a new data escape
4132
4133 \C*n pass the number n (may be negative) as callout_data
4134
4135 If the callout function in pcretest receives a non-zero value as
4136 callout_data, it returns that value.
4137
4138 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4139 there were some redundant $(CFLAGS) in commands that are now specified as
4140 $(LINK), which already includes $(CFLAGS).
4141
4142 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4143 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4144 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4145 one-byte characters throughout. Note that case-insensitive matching applies
4146 only to characters whose values are less than 256. PCRE doesn't support the
4147 notion of cases for higher-valued characters.
4148
4149 (i) A character class whose characters are all within 0-255 is handled as
4150 a bit map, and the map is inverted for negative classes. Previously, a
4151 character > 255 always failed to match such a class; however it should
4152 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4153
4154 (ii) A negated character class with a single character < 255 is coded as
4155 "not this character" (OP_NOT). This wasn't working properly when the test
4156 character was multibyte, either singly or repeated.
4157
4158 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4159 mode, for example: \x{100}{2,3}.
4160
4161 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4162 singly or repeated) now correctly test multibyte characters. However,
4163 PCRE doesn't recognize any characters with values greater than 255 as
4164 digits, spaces, or word characters. Such characters always match \D, \S,
4165 and \W, and never match \d, \s, or \w.
4166
4167 (v) Classes may now contain characters and character ranges with values
4168 greater than 255. For example: [ab\x{100}-\x{400}].
4169
4170 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4171 PCRE in UTF-8 mode.
4172
4173 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4174 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4175 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4176 value.)
4177
4178 53. The single man page has become too large. I have therefore split it up into
4179 a number of separate man pages. These also give rise to individual HTML pages;
4180 these are now put in a separate directory, and there is an index.html page that
4181 lists them all. Some hyperlinking between the pages has been installed.
4182
4183 54. Added convenience functions for handling named capturing parentheses.
4184
4185 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4186 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4187 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4188 are faulted.
4189
4190 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4191 calling configure. These values are used when compiling the dftables.c program
4192 which is run to generate the source of the default character tables. They
4193 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4194 you will need to set these values.
4195
4196 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4197
4198
4199 Version 3.9 02-Jan-02
4200 ---------------------
4201
4202 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4203
4204 2. If --disable-static was given, the building process failed when trying to
4205 build pcretest and pcregrep. (For some reason it was using libtool to compile
4206 them, which is not right, as they aren't part of the library.)
4207
4208
4209 Version 3.8 18-Dec-01
4210 ---------------------
4211
4212 1. The experimental UTF-8 code was completely screwed up. It was packing the
4213 bytes in the wrong order. How dumb can you get?
4214
4215
4216 Version 3.7 29-Oct-01
4217 ---------------------
4218
4219 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4220 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4221 this didn't happen under Solaris 8, where I normally test things.
4222
4223 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4224 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4225 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4226 this fix an hour or so after the initial 3.7 release.)
4227
4228
4229 Version 3.6 23-Oct-01
4230 ---------------------
4231
4232 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4233 offsets passed as NULL with zero offset count.
4234
4235 2. The config.guess and config.sub files had not been updated when I moved to
4236 the latest autoconf.
4237
4238
4239 Version 3.5 15-Aug-01
4240 ---------------------
4241
4242 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4243 had been forgotten.
4244
4245 2. By using declared but undefined structures, we can avoid using "void"
4246 definitions in pcre.h while keeping the internal definitions of the structures
4247 private.
4248
4249 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4250 user point of view, this means that both static and shared libraries are built
4251 by default, but this can be individually controlled. More of the work of
4252 handling this static/shared cases is now inside libtool instead of PCRE's make
4253 file.
4254
4255 4. The pcretest utility is now installed along with pcregrep because it is
4256 useful for users (to test regexs) and by doing this, it automatically gets
4257 relinked by libtool. The documentation has been turned into a man page, so
4258 there are now .1, .txt, and .html versions in /doc.
4259
4260 5. Upgrades to pcregrep:
4261 (i) Added long-form option names like gnu grep.
4262 (ii) Added --help to list all options with an explanatory phrase.
4263 (iii) Added -r, --recursive to recurse into sub-directories.
4264 (iv) Added -f, --file to read patterns from a file.
4265
4266 6. pcre_exec() was referring to its "code" argument before testing that
4267 argument for NULL (and giving an error if it was NULL).
4268
4269 7. Upgraded Makefile.in to allow for compiling in a different directory from
4270 the source directory.
4271
4272 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4273 options bits, the pointer it was passed was to an int instead of to an unsigned
4274 long int. This mattered only on 64-bit systems.
4275
4276 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4277 generated) instead of pcre.in, which it its source. Also made the same change
4278 in several of the .c files.
4279
4280 10. A new release of gcc defines printf() as a macro, which broke pcretest
4281 because it had an ifdef in the middle of a string argument for printf(). Fixed
4282 by using separate calls to printf().
4283
4284 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4285 script, to force use of CR or LF instead of \n in the source. On non-Unix
4286 systems, the value can be set in config.h.
4287
4288 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4289 absolute limit. Changed the text of the error message to make this clear, and
4290 likewise updated the man page.
4291
4292 13. The limit of 99 on the number of capturing subpatterns has been removed.
4293 The new limit is 65535, which I hope will not be a "real" limit.
4294
4295
4296 Version 3.4 22-Aug-00
4297 ---------------------
4298
4299 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4300
4301 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4302
4303
4304 Version 3.3 01-Aug-00
4305 ---------------------
4306
4307 1. If an octal character was given, but the value was greater than \377, it
4308 was not getting masked to the least significant bits, as documented. This could
4309 lead to crashes in some systems.
4310
4311 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4312 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4313
4314 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4315 These just pass their arguments on to (pcre_free)(), but they are provided
4316 because some uses of PCRE bind it to non-C systems that can call its functions,
4317 but cannot call free() or pcre_free() directly.
4318
4319 4. Add "make test" as a synonym for "make check". Corrected some comments in
4320 the Makefile.
4321
4322 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4323 Makefile.
4324
4325 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4326 command called pgrep for grepping around the active processes.
4327
4328 7. Added the beginnings of support for UTF-8 character strings.
4329
4330 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4331 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4332 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4333 out for the ar command.)
4334
4335
4336 Version 3.2 12-May-00
4337 ---------------------
4338
4339 This is purely a bug fixing release.
4340
4341 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4342 of ZA. This was just one example of several cases that could provoke this bug,
4343 which was introduced by change 9 of version 2.00. The code for breaking
4344 infinite loops after an iteration that matches an empty string was't working
4345 correctly.
4346
4347 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4348 when matched against abbab (for example). After matching an empty string, it
4349 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4350 caused it to match further down the string than it should.
4351
4352 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4353 was there because it doesn't seem to be needed, and it causes trouble on some
4354 systems, as it is not a Standard C header. It has been removed.
4355
4356 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4357 were reported on the Macintosh. The changes were from
4358
4359 while ((c = *(++ptr)) != 0 && c != '\n');
4360 to
4361 while ((c = *(++ptr)) != 0 && c != '\n') ;
4362
4363 Totally extraordinary, but if that's what it takes...
4364
4365 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4366 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4367 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4368 assumes the way PCRE uses memmove() (always moving upwards).
4369
4370 6. PCRE is being used in one environment where strchr() is not available. There
4371 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4372 faster code anyway.
4373
4374
4375 Version 3.1 09-Feb-00
4376 ---------------------
4377
4378 The only change in this release is the fixing of some bugs in Makefile.in for
4379 the "install" target:
4380
4381 (1) It was failing to install pcreposix.h.
4382
4383 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4384
4385
4386 Version 3.0 01-Feb-00
4387 ---------------------
4388
4389 1. Add support for the /+ modifier to perltest (to output $` like it does in
4390 pcretest).
4391
4392 2. Add support for the /g modifier to perltest.
4393
4394 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4395 matches null strings.
4396
4397 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4398 pattern. Perl treats empty patterns specially - it reuses the most recent
4399 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4400 effect.
4401
4402 5. The POSIX interface was broken in that it was just handing over the POSIX
4403 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4404 required a bigger vector, with some working space on the end. This means that
4405 the POSIX wrapper now has to get and free some memory, and copy the results.
4406
4407 6. Added some simple autoconf support, placing the test data and the
4408 documentation in separate directories, re-organizing some of the
4409 information files, and making it build pcre-config (a GNU standard). Also added
4410 libtool support for building PCRE as a shared library, which is now the
4411 default.
4412
4413 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4414 09 are not valid octal constants. Single digits will be used for minor values
4415 less than 10.
4416
4417 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4418 existing programs that set these in the POSIX interface can use PCRE without
4419 modification.
4420
4421 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4422 return all that pcre_info() returns, plus additional data. The pcre_info()
4423 function is retained for compatibility, but is considered to be obsolete.
4424
4425 10. Added experimental recursion feature (?R) to handle one common case that
4426 Perl 5.6 will be able to do with (?p{...}).
4427
4428 11. Added support for POSIX character classes like [:alpha:], which Perl is
4429 adopting.
4430
4431
4432 Version 2.08 31-Aug-99
4433 ----------------------
4434
4435 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4436 trying to match at the startoffset position, but instead was moving forward to
4437 the next newline as if a previous match had failed.
4438
4439 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4440 and could get into a loop if a null string was matched other than at the start
4441 of the subject.
4442
4443 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4444 be distinguished at compile time, and for completeness also added PCRE_DATE.
4445
4446 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4447 in GnuWin32 environments.
4448
4449
4450 Version 2.07 29-Jul-99
4451 ----------------------
4452
4453 1. The documentation is now supplied in plain text form and HTML as well as in
4454 the form of man page sources.
4455
4456 2. C++ compilers don't like assigning (void *) values to other pointer types.
4457 In particular this affects malloc(). Although there is no problem in Standard
4458 C, I've put in casts to keep C++ compilers happy.
4459
4460 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4461 should be (const char *).
4462
4463 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4464 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4465 However, I haven't made this a standard facility. The documentation doesn't
4466 mention it, and the Makefile doesn't support it.
4467
4468 5. The Makefile now contains an "install" target, with editable destinations at
4469 the top of the file. The pcretest program is not installed.
4470
4471 6. pgrep -V now gives the PCRE version number and date.
4472
4473 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4474 causing the entire string to be ignored, instead of just the last character.
4475
4476 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4477 non-matching string, it can take a very, very long time, even for strings of
4478 quite modest length, because of the nested recursion. PCRE now does better in
4479 some of these cases. It does this by remembering the last required literal
4480 character in the pattern, and pre-searching the subject to ensure it is present
4481 before running the real match. In other words, it applies a heuristic to detect
4482 some types of certain failure quickly, and in the above example, if presented
4483 with a string that has no trailing " it gives "no match" very quickly.
4484
4485 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4486 other alternatives are tried instead.
4487
4488
4489 Version 2.06 09-Jun-99
4490 ----------------------
4491
4492 1. Change pcretest's output for amount of store used to show just the code
4493 space, because the remainder (the data block) varies in size between 32-bit and
4494 64-bit systems.
4495
4496 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4497 start matching at. This allows lookbehinds to work when searching for multiple
4498 occurrences in a string.
4499
4500 3. Added additional options to pcretest for testing multiple occurrences:
4501
4502 /+ outputs the rest of the string that follows a match
4503 /g loops for multiple occurrences, using the new startoffset argument
4504 /G loops for multiple occurrences by passing an incremented pointer
4505
4506 4. PCRE wasn't doing the "first character" optimization for patterns starting
4507 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4508 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4509 the letter 'x'. On long subject strings, this gives a significant speed-up.
4510
4511
4512 Version 2.05 21-Apr-99
4513 ----------------------
4514
4515 1. Changed the type of magic_number from int to long int so that it works
4516 properly on 16-bit systems.
4517
4518 2. Fixed a bug which caused patterns starting with .* not to work correctly
4519 when the subject string contained newline characters. PCRE was assuming
4520 anchoring for such patterns in all cases, which is not correct because .* will
4521 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4522 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4523 must be retried after every newline in the subject.
4524
4525
4526 Version 2.04 18-Feb-99
4527 ----------------------
4528
4529 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4530 computation of the store needed to hold the pattern was incorrect (too large).
4531 If such patterns were nested a few deep, this could multiply and become a real
4532 problem.
4533
4534 2. Added /M option to pcretest to show the memory requirement of a specific
4535 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4536
4537 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4538 compiled in such a way that the backtracking after subsequent failure was
4539 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4540 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4541
4542
4543 Version 2.03 02-Feb-99
4544 ----------------------
4545
4546 1. Fixed typo and small mistake in man page.
4547
4548 2. Added 4th condition (GPL supersedes if conflict) and created separate
4549 LICENCE file containing the conditions.
4550
4551 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4552 Perl, that is the internal \ allows the delimiter to be included in the
4553 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4554 the final delimiter, add \ to the end of the pattern (to test the error).
4555
4556 4. Added the convenience functions for extracting substrings after a successful
4557 match. Updated pcretest to make it able to test these functions.
4558
4559
4560 Version 2.02 14-Jan-99
4561 ----------------------
4562
4563 1. Initialized the working variables associated with each extraction so that
4564 their saving and restoring doesn't refer to uninitialized store.
4565
4566 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4567 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4568 fix the problem.
4569
4570 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4571 calls, and wasn't printing the correct value for compiling calls. Increased the
4572 default value of LOOPREPEAT, and the number of significant figures in the
4573 times.
4574
4575 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4576
4577 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4578 a building problem on Windows NT with a FAT file system.
4579
4580
4581 Version 2.01 21-Oct-98
4582 ----------------------
4583
4584 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4585 to character tables built by pcre_maketables() in the current locale. If NULL
4586 is passed, the default tables are used.
4587
4588
4589 Version 2.00 24-Sep-98
4590 ----------------------
4591
4592 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4593 it any more.
4594
4595 2. Allow quantification of (?>) groups, and make it work correctly.
4596
4597 3. The first character computation wasn't working for (?>) groups.
4598
4599 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4600 end of the subject) and add 5.005's \z, which really does match only at the
4601 very end of the subject.
4602
4603 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4604
4605 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4606 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4607 localized options. All options to pcre_study() were also removed.
4608
4609 7. Add other new features from 5.005:
4610
4611 $(?<= positive lookbehind
4612 $(?<! negative lookbehind
4613 (?imsx-imsx) added the unsetting capability
4614 such a setting is global if at outer level; local otherwise
4615 (?imsx-imsx:) non-capturing groups with option setting
4616 (?(cond)re|re) conditional pattern matching
4617
4618 A backreference to itself in a repeated group matches the previous
4619 captured string.
4620
4621 8. General tidying up of studying (both automatic and via "study")
4622 consequential on the addition of new assertions.
4623
4624 9. As in 5.005, unlimited repeated groups that could match an empty substring
4625 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4626 runtime if any iteration does actually match an empty substring.
4627
4628 10. Include the RunTest script in the distribution.
4629
4630 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4631 discrepancies, some of which were old and were also with respect to 5.004. They
4632 have now been fixed.
4633
4634
4635 Version 1.09 28-Apr-98
4636 ----------------------
4637
4638 1. A negated single character class followed by a quantifier with a minimum
4639 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4640 program crashes, or just wrong answers. This did not apply to negated classes
4641 containing more than one character, or to minima other than one.
4642
4643
4644 Version 1.08 27-Mar-98
4645 ----------------------
4646
4647 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4648
4649 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4650 latter must appear before anything that relies on it in the pattern.
4651
4652
4653 Version 1.07 16-Feb-98
4654 ----------------------
4655
4656 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4657 repeat of a potentially empty string).
4658
4659
4660 Version 1.06 23-Jan-98
4661 ----------------------
4662
4663 1. Added Markus Oberhumer's little patches for C++.
4664
4665 2. Literal strings longer than 255 characters were broken.
4666
4667
4668 Version 1.05 23-Dec-97
4669 ----------------------
4670
4671 1. Negated character classes containing more than one character were failing if
4672 PCRE_CASELESS was set at run time.
4673
4674
4675 Version 1.04 19-Dec-97
4676 ----------------------
4677
4678 1. Corrected the man page, where some "const" qualifiers had been omitted.
4679
4680 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4681 input syntax.
4682
4683 3. Fixed memory leak which occurred when a regex with back references was
4684 matched with an offsets vector that wasn't big enough. The temporary memory
4685 that is used in this case wasn't being freed if the match failed.
4686
4687 4. Tidied pcretest to ensure it frees memory that it gets.
4688
4689 5. Temporary memory was being obtained in the case where the passed offsets
4690 vector was exactly big enough.
4691
4692 6. Corrected definition of offsetof() from change 5 below.
4693
4694 7. I had screwed up change 6 below and broken the rules for the use of
4695 setjmp(). Now fixed.
4696
4697
4698 Version 1.03 18-Dec-97
4699 ----------------------
4700
4701 1. A erroneous regex with a missing opening parenthesis was correctly
4702 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4703 on some systems.
4704
4705 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4706 it was reported that one broken compiler failed on the former because "code" is
4707 also an independent variable.
4708
4709 3. The erroneous regex a[]b caused an array overrun reference.
4710
4711 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4712 fail on data ending with that character. (It was going on too far, and checking
4713 the next character, typically a binary zero.) This was specific to the
4714 optimized code for single-character negative classes.
4715
4716 5. Added a contributed patch from the TIN world which does the following:
4717
4718 + Add an undef for memmove, in case the the system defines a macro for it.
4719
4720 + Add a definition of offsetof(), in case there isn't one. (I don't know
4721 the reason behind this - offsetof() is part of the ANSI standard - but
4722 it does no harm).
4723
4724 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4725 most of the places where whitespace preceded '#'. I have given up and
4726 allowed the remaining 2 cases to be at the margin.
4727
4728 + Rename some variables in pcre to eliminate shadowing. This seems very
4729 pedantic, but does no harm, of course.
4730
4731 6. Moved the call to setjmp() into its own function, to get rid of warnings
4732 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4733
4734 7. Constructs such as \d{8,} were compiling into the equivalent of
4735 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4736 outcome, but in this particular case used more store than had been allocated,
4737 which caused the bug to be discovered because it threw up an internal error.
4738
4739 8. The debugging code in both pcre and pcretest for outputting the compiled
4740 form of a regex was going wrong in the case of back references followed by
4741 curly-bracketed repeats.
4742
4743
4744 Version 1.02 12-Dec-97
4745 ----------------------
4746
4747 1. Typos in pcre.3 and comments in the source fixed.
4748
4749 2. Applied a contributed patch to get rid of places where it used to remove
4750 'const' from variables, and fixed some signed/unsigned and uninitialized
4751 variable warnings.
4752
4753 3. Added the "runtest" target to Makefile.
4754
4755 4. Set default compiler flag to -O2 rather than just -O.
4756
4757
4758 Version 1.01 19-Nov-97
4759 ----------------------
4760
4761 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4762 like /([ab]*)*/, that is, for classes with more than one character in them.
4763
4764 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4765 as /((?>a*))*/ (a PCRE_EXTRA facility).
4766
4767
4768 Version 1.00 18-Nov-97
4769 ----------------------
4770
4771 1. Added compile-time macros to support systems such as SunOS4 which don't have
4772 memmove() or strerror() but have other things that can be used instead.
4773
4774 2. Arranged that "make clean" removes the executables.
4775
4776
4777 Version 0.99 27-Oct-97
4778 ----------------------
4779
4780 1. Fixed bug in code for optimizing classes with only one character. It was
4781 initializing a 32-byte map regardless, which could cause it to run off the end
4782 of the memory it had got.
4783
4784 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4785
4786
4787 Version 0.98 22-Oct-97
4788 ----------------------
4789
4790 1. Fixed bug in code for handling temporary memory usage when there are more
4791 back references than supplied space in the ovector. This could cause segfaults.
4792
4793
4794 Version 0.97 21-Oct-97
4795 ----------------------
4796
4797 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4798
4799 2. Optimized negated single characters not to use a bit map.
4800
4801 3. Brought error texts together as macro definitions; clarified some of them;
4802 fixed one that was wrong - it said "range out of order" when it meant "invalid
4803 escape sequence".
4804
4805 4. Changed some char * arguments to const char *.
4806
4807 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4808
4809 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4810 pcretest.
4811
4812
4813 Version 0.96 16-Oct-97
4814 ----------------------
4815
4816 1. Added a simple "pgrep" utility to the distribution.
4817
4818 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4819 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4820 where "ddd" means "one or more decimal digits".
4821
4822 3. Fixed serious bug. If a pattern had a back reference, but the call to
4823 pcre_exec() didn't supply a large enough ovector to record the related
4824 identifying subpattern, the match always failed. PCRE now remembers the number
4825 of the largest back reference, and gets some temporary memory in which to save
4826 the offsets during matching if necessary, in order to ensure that
4827 backreferences always work.
4828
4829 4. Increased the compatibility with Perl in a number of ways:
4830
4831 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4832 to request this handling. The option can be set at compile or exec time.
4833
4834 (b) $ matches before a terminating newline by default; an option
4835 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4836 mode). The option can be set at compile or exec time.
4837
4838 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4839 the same as Perl's. If the decimal number it represents is less than 10
4840 or there aren't that many previous left capturing parentheses, an octal
4841 escape is read. Inside a character class, it's always an octal escape,
4842 even if it is a single digit.
4843
4844 (d) An escaped but undefined alphabetic character is taken as a literal,
4845 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4846 escapes.
4847
4848 (e) {0} is now permitted. (The previous item is removed from the compiled
4849 pattern).
4850
4851 5. Changed all the names of code files so that the basic parts are no longer
4852 than 10 characters, and abolished the teeny "globals.c" file.
4853
4854 6. Changed the handling of character classes; they are now done with a 32-byte
4855 bit map always.
4856
4857 7. Added the -d and /D options to pcretest to make it possible to look at the
4858 internals of compilation without having to recompile pcre.
4859
4860
4861 Version 0.95 23-Sep-97
4862 ----------------------
4863
4864 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4865 \x20 at the start of a run of normal characters. These were being treated as
4866 real characters, instead of the source characters being re-checked.
4867
4868
4869 Version 0.94 18-Sep-97
4870 ----------------------
4871
4872 1. The functions are now thread-safe, with the caveat that the global variables
4873 containing pointers to malloc() and free() or alternative functions are the
4874 same for all threads.
4875
4876 2. Get pcre_study() to generate a bitmap of initial characters for non-
4877 anchored patterns when this is possible, and use it if passed to pcre_exec().
4878
4879
4880 Version 0.93 15-Sep-97
4881 ----------------------
4882
4883 1. /(b)|(:+)/ was computing an incorrect first character.
4884
4885 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4886 but not actually doing anything yet.
4887
4888 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4889 as Perl does (e.g. [-az] or [az-]).
4890
4891 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4892 all possible positions.
4893
4894 5. Split up into different modules to avoid including unneeded functions in a
4895 compiled binary. However, compile and exec are still in one module. The "study"
4896 function is split off.
4897
4898 6. The character tables are now in a separate module whose source is generated
4899 by an auxiliary program - but can then be edited by hand if required. There are
4900 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4901 toupper() in the code.
4902
4903 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4904 make them global. Abolish the function for setting them, as the caller can now
4905 set them directly.
4906
4907
4908 Version 0.92 11-Sep-97
4909 ----------------------
4910
4911 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4912 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4913
4914 2. Caseless matching was not working in character classes if the characters in
4915 the pattern were in upper case.
4916
4917 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4918
4919 4. Make PCRE_ANCHORED public and accept as a compile option.
4920
4921 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4922 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4923 pass them.
4924
4925 6. Give an error if bad option bits passed at compile or run time.
4926
4927 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4928 pcretest to cause it to pass that flag.
4929
4930 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4931 options, and the first character, if set.
4932
4933 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4934
4935
4936 Version 0.91 10-Sep-97
4937 ----------------------
4938
4939 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4940 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4941
4942 2. PCRE was looping on encountering an indefinitely repeated back reference to
4943 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4944 Perl does - treats the match as successful.
4945
4946 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5