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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1211 - (show annotations)
Wed Nov 7 17:02:03 2012 UTC (7 years, 1 month ago) by ph10
File size: 239394 byte(s)
Error occurred while calculating annotation data.
Fix pcretest -C pcre8/pcre16 bug.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.32
5 ------------
6
7 1. Improved JIT compiler optimizations for first character search and single
8 character iterators.
9
10 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11 Patch by Daniel Richard G.
12
13 3. Single character iterator optimizations in the JIT compiler.
14
15 4. Improved JIT compiler optimizations for character ranges.
16
17 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18 Reported by Giuseppe D'Angelo.
19
20 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21 of a line, was being set incorrectly in cases where .* appeared inside
22 atomic brackets at the start of a pattern, or where there was a subsequent
23 *PRUNE or *SKIP.
24
25 7. Improved instruction cache flush for POWER/PowerPC.
26 Patch by Daniel Richard G.
27
28 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
29 grep:
30
31 (a) There is now no limit to the number of patterns to be matched.
32
33 (b) An error is given if a pattern is too long.
34
35 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
36 are now supported.
37
38 (d) --exclude-from and --include-from (multiple use) have been added.
39
40 (e) Exclusions and inclusions now apply to all files and directories, not
41 just to those obtained from scanning a directory recursively.
42
43 (f) Multiple uses of -f and --file-list are now supported.
44
45 (g) In a Windows environment, the default for -d has been changed from
46 "read" (the GNU grep default) to "skip", because otherwise the presence
47 of a directory in the file list provokes an error.
48
49 (h) The documentation has been revised and clarified in places.
50
51 9. Improve the matching speed of capturing brackets.
52
53 10. Changed the meaning of \X so that it now matches a Unicode extended
54 grapheme cluster.
55
56 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
57 out POSIX threads when JIT support is configured.
58
59 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
60
61 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
62 structure could go wrong in environments where size_t is not the same size
63 as int.
64
65 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
66 set.
67
68 15. The EBCDIC support had decayed; later updates to the code had included
69 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
70 general tidy up of EBCDIC-related issues, and the documentation was also
71 not quite right. There is now a test that can be run on ASCII systems to
72 check some of the EBCDIC-related things (but is it not a full test).
73
74 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
75 in a small tidy to the code.
76
77 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
78
79 18. If the --only-matching (-o) option in pcregrep is specified multiple
80 times, each one causes appropriate output. For example, -o1 -o2 outputs the
81 substrings matched by the 1st and 2nd capturing parentheses. A separating
82 string can be specified by --om-separator (default empty).
83
84 19. Improving the first n character searches.
85
86 20. Turn case lists for horizontal and vertical white space into macros so that
87 they are defined only once.
88
89 21. This set of changes together give more compatible Unicode case-folding
90 behaviour for characters that have more than one other case when UCP
91 support is available.
92
93 (a) The Unicode property table now has offsets into a new table of sets of
94 three or more characters that are case-equivalent. The MultiStage2.py
95 script that generates these tables (the pcre_ucd.c file) now scans
96 CaseFolding.txt instead of UnicodeData.txt for character case
97 information.
98
99 (b) The code for adding characters or ranges of characters to a character
100 class has been abstracted into a generalized function that also handles
101 case-independence. In UTF-mode with UCP support, this uses the new data
102 to handle characters with more than one other case.
103
104 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
105 whose other case is greater than 256 are now correctly matched
106 caselessly. Previously, the high codepoint matched the low one, but not
107 vice versa.
108
109 (d) The processing of \h, \H, \v, and \ in character classes now makes use
110 of the new class addition function, using character lists defined as
111 macros alongside the case definitions of 20 above.
112
113 (e) Caseless back references now work with characters that have more than
114 one other case.
115
116 (f) General caseless matching of characters with more than one other case
117 is supported.
118
119 22. Unicode character properties were updated from Unicode 6.2.0
120
121 23. Improved CMake support under Windows. Patch by Daniel Richard G.
122
123 24. Add support for 32-bit character strings, and UTF-32
124
125 25. Major JIT compiler update (code refactoring and bugfixing).
126 Experimental Sparc 32 support is added.
127
128 26. Applied a modified version of Daniel Richard G's patch to create
129 pcre.h.generic and config.h.generic by "make" instead of in the
130 PrepareRelease script.
131
132 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
133 pcre_compile.c when checking for a zero character.
134
135 28. Introducing a native interface for JIT. Through this interface, the compiled
136 machine code can be directly executed. The purpose of this interface is to
137 provide fast pattern matching, so several sanity checks are not performed.
138 However, feature tests are still performed. The new interface provides
139 1.4x speedup compared to the old one.
140
141 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
142 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
143 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
144
145 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
146 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
147 the "old" RFC 2279). Instead, it ended up passing a negative length to
148 pcre_exec().
149
150 31. Add support for GCC's visibility feature to hide internal functions.
151
152 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
153 "unknown -C option" after outputting 0 or 1.
154
155
156 Version 8.31 06-July-2012
157 -------------------------
158
159 1. Fixing a wrong JIT test case and some compiler warnings.
160
161 2. Removed a bashism from the RunTest script.
162
163 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
164 to unsigned type, result still unsigned" that was given by an MS compiler
165 on encountering the code "-sizeof(xxx)".
166
167 4. Partial matching support is added to the JIT compiler.
168
169 5. Fixed several bugs concerned with partial matching of items that consist
170 of more than one character:
171
172 (a) /^(..)\1/ did not partially match "aba" because checking references was
173 done on an "all or nothing" basis. This also applied to repeated
174 references.
175
176 (b) \R did not give a hard partial match if \r was found at the end of the
177 subject.
178
179 (c) \X did not give a hard partial match after matching one or more
180 characters at the end of the subject.
181
182 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
183 a partial match for the string "\r".
184
185 (e) When newline was set to CRLF, the metacharacter "." did not recognize
186 a partial match for a CR character at the end of the subject string.
187
188 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
189 running pcretest, the text "(JIT)" added to the output whenever JIT is
190 actually used to run the match.
191
192 7. Individual JIT compile options can be set in pcretest by following -s+[+]
193 or /S+[+] with a digit between 1 and 7.
194
195 8. OP_NOT now supports any UTF character not just single-byte ones.
196
197 9. (*MARK) control verb is now supported by the JIT compiler.
198
199 10. The command "./RunTest list" lists the available tests without actually
200 running any of them. (Because I keep forgetting what they all are.)
201
202 11. Add PCRE_INFO_MAXLOOKBEHIND.
203
204 12. Applied a (slightly modified) user-supplied patch that improves performance
205 when the heap is used for recursion (compiled with --disable-stack-for-
206 recursion). Instead of malloc and free for each heap frame each time a
207 logical recursion happens, frames are retained on a chain and re-used where
208 possible. This sometimes gives as much as 30% improvement.
209
210 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
211 call.
212
213 14. As documented, (*COMMIT) is now confined to within a positive assertion.
214
215 15. It is now possible to link pcretest with libedit as an alternative to
216 libreadline.
217
218 16. (*COMMIT) control verb is now supported by the JIT compiler.
219
220 17. The Unicode data tables have been updated to Unicode 6.1.0.
221
222 18. Added --file-list option to pcregrep.
223
224 19. Added binary file support to pcregrep, including the -a, --binary-files,
225 -I, and --text options.
226
227 20. The madvise function is renamed for posix_madvise for QNX compatibility
228 reasons. Fixed by Giuseppe D'Angelo.
229
230 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
231
232 22. Changed the option for creating symbolic links for 16-bit man pages from
233 -s to -sf so that re-installing does not cause issues.
234
235 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
236
237 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
238 matches in certain environments (the workspace was not being correctly
239 retained). Also added to pcre_dfa_exec() a simple plausibility check on
240 some of the workspace data at the beginning of a restart.
241
242 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
243 was not doing so when it should - probably a typo introduced by SVN 528
244 (change 8.10/14).
245
246 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
247 \w+ when the character tables indicated that \x{c4} was a word character.
248 There were several related cases, all because the tests for doing a table
249 lookup were testing for characters less than 127 instead of 255.
250
251 27. If a pattern contains capturing parentheses that are not used in a match,
252 their slots in the ovector are set to -1. For those that are higher than
253 any matched groups, this happens at the end of processing. In the case when
254 there were back references that the ovector was too small to contain
255 (causing temporary malloc'd memory to be used during matching), and the
256 highest capturing number was not used, memory off the end of the ovector
257 was incorrectly being set to -1. (It was using the size of the temporary
258 memory instead of the true size.)
259
260 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
261 ovector size, it uses memory at the end of the block that it has got.
262
263 29. Check for an overlong MARK name and give an error at compile time. The
264 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
265
266 30. JIT compiler update.
267
268 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
269 Rill for the patch.
270
271 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
272
273 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
274
275 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
276 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
277 was enabled.
278
279 35. Improve JIT code generation for greedy plus quantifier.
280
281 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
282 1 to "aa" instead of to an empty string. The bug affected repeated groups
283 that could potentially match an empty string.
284
285 37. Optimizing single character iterators in JIT.
286
287 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
288 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
289 codepoints that are too big for the mode are faulted, and in a UTF mode,
290 disallowed codepoints are also faulted.
291
292 39. If PCRE was compiled with UTF support, in three places in the DFA
293 matcher there was code that should only have been obeyed in UTF mode, but
294 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
295 processing when bytes with values greater than 127 were present. In 16-bit
296 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
297 both cases the values are those that cannot be the first data item in a UTF
298 character. The three items that might have provoked this were recursions,
299 possessively repeated groups, and atomic groups.
300
301 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
302 and pcregrep, because some OS require shared objects to be explicitly
303 passed to ld, causing the link step to fail if they are not.
304
305 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
306 mode, patterns that started with \h* or \R* might be incorrectly matched.
307
308
309 Version 8.30 04-February-2012
310 -----------------------------
311
312 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
313 name is defined in ctype.h.
314
315 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
316 only in quite long subpatterns.
317
318 3. Removed the function pcre_info(), which has been obsolete and deprecated
319 since it was replaced by pcre_fullinfo() in February 2000.
320
321 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
322 match a (*MARK), and the match failed at the start of the subject, a
323 reference to memory before the start of the subject could occur. This bug
324 was introduced by fix 17 of release 8.21.
325
326 5. A reference to an unset group with zero minimum repetition was giving
327 totally wrong answers (in non-JavaScript-compatibility mode). For example,
328 /(another)?(\1?)test/ matched against "hello world test". This bug was
329 introduced in release 8.13.
330
331 6. Add support for 16-bit character strings (a large amount of work involving
332 many changes and refactorings).
333
334 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
335 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
336 from a file.
337
338 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
339 rounding is not applied in this particular case).
340
341 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
342 if they appear, or are escaped, in patterns.
343
344 10. Get rid of a number of -Wunused-but-set-variable warnings.
345
346 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
347 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
348 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
349 also returns the mark "x". This bug applied to capturing parentheses,
350 non-capturing parentheses, and atomic parentheses. It also applied to some
351 assertions.
352
353 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
354 information out of configure.ac instead of relying on pcre.h.generic, which
355 is not stored in the repository.
356
357 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
358 -lreadline.
359
360 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
361
362 15. Applied Graycode's patch to put the top-level frame on the stack rather
363 than the heap when not using the stack for recursion. This gives a
364 performance improvement in many cases when recursion is not deep.
365
366 16. Experimental code added to "pcretest -C" to output the stack frame size.
367
368
369 Version 8.21 12-Dec-2011
370 ------------------------
371
372 1. Updating the JIT compiler.
373
374 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
375 are added as well.
376
377 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
378 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
379 calling _pcre_jit_exec. Some extra comments are added.
380
381 4. (*MARK) settings inside atomic groups that do not contain any capturing
382 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
383 was introduced by change 18 for 8.20.
384
385 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
386 ECMA-262 standard.
387
388 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
389 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
390 This bug was probably introduced by change 9 of 8.13.
391
392 7. While fixing 6 above, I noticed that a number of other items were being
393 incorrectly rejected as "not fixed length". This arose partly because newer
394 opcodes had not been added to the fixed-length checking code. I have (a)
395 corrected the bug and added tests for these items, and (b) arranged for an
396 error to occur if an unknown opcode is encountered while checking for fixed
397 length instead of just assuming "not fixed length". The items that were
398 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
399 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
400 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
401
402 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
403 being incorrectly compiled and would have given unpredicatble results.
404
405 9. A possessively repeated subpattern with minimum repeat count greater than
406 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
407 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
408 the first (A) could occur when it should not.
409
410 10. Add a cast and remove a redundant test from the code.
411
412 11. JIT should use pcre_malloc/pcre_free for allocation.
413
414 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
415 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
416 is anything other than /usr, -L is still shown).
417
418 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
419
420 14. Perl does not support \N without a following name in a [] class; PCRE now
421 also gives an error.
422
423 15. If a forward reference was repeated with an upper limit of around 2000,
424 it caused the error "internal error: overran compiling workspace". The
425 maximum number of forward references (including repeats) was limited by the
426 internal workspace, and dependent on the LINK_SIZE. The code has been
427 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
428 the default depends on LINK_SIZE. There is a new upper limit (for safety)
429 of around 200,000 forward references. While doing this, I also speeded up
430 the filling in of repeated forward references.
431
432 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
433 incorrectly expecting the subject to contain another "a" after the start.
434
435 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
436 in the match, the SKIP should be ignored. This was not happening; instead
437 the SKIP was being treated as NOMATCH. For patterns such as
438 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
439 tested.
440
441 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
442 now much more compatible with Perl, in particular in cases where the result
443 is a non-match for a non-anchored pattern. For example, if
444 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
445 "m", where previously it did not return a name. A side effect of this
446 change is that for partial matches, the last encountered mark name is
447 returned, as for non matches. A number of tests that were previously not
448 Perl-compatible have been moved into the Perl-compatible test files. The
449 refactoring has had the pleasing side effect of removing one argument from
450 the match() function, thus reducing its stack requirements.
451
452 19. If the /S+ option was used in pcretest to study a pattern using JIT,
453 subsequent uses of /S (without +) incorrectly behaved like /S+.
454
455 21. Retrieve executable code size support for the JIT compiler and fixing
456 some warnings.
457
458 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
459 not work when the shorter character appeared right at the end of the
460 subject string.
461
462 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
463 systems.
464
465 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
466 output it when the /M option is used in pcretest.
467
468 25. The CheckMan script was not being included in the distribution. Also, added
469 an explicit "perl" to run Perl scripts from the PrepareRelease script
470 because this is reportedly needed in Windows.
471
472 26. If study data was being save in a file and studying had not found a set of
473 "starts with" bytes for the pattern, the data written to the file (though
474 never used) was taken from uninitialized memory and so caused valgrind to
475 complain.
476
477 27. Updated RunTest.bat as provided by Sheri Pierce.
478
479 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
480
481 29. Computation of memory usage for the table of capturing group names was
482 giving an unnecessarily large value.
483
484
485 Version 8.20 21-Oct-2011
486 ------------------------
487
488 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
489 a POSIX class. After further experiments with Perl, which convinced me that
490 Perl has bugs and confusions, a closing square bracket is no longer allowed
491 in a POSIX name. This bug also affected patterns with classes that started
492 with full stops.
493
494 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
495 captured substring, but while checking the failing first alternative,
496 substring 1 is temporarily captured. If the output vector supplied to
497 pcre_exec() was not big enough for this capture, the yield of the function
498 was still zero ("insufficient space for captured substrings"). This cannot
499 be totally fixed without adding another stack variable, which seems a lot
500 of expense for a edge case. However, I have improved the situation in cases
501 such as /(a)(b)x|abc/ matched against "abc", where the return code
502 indicates that fewer than the maximum number of slots in the ovector have
503 been set.
504
505 3. Related to (2) above: when there are more back references in a pattern than
506 slots in the output vector, pcre_exec() uses temporary memory during
507 matching, and copies in the captures as far as possible afterwards. It was
508 using the entire output vector, but this conflicts with the specification
509 that only 2/3 is used for passing back captured substrings. Now it uses
510 only the first 2/3, for compatibility. This is, of course, another edge
511 case.
512
513 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
514 main code base, and can be used by building with --enable-jit. When this is
515 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
516 runtime --no-jit option is given.
517
518 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
519 ovector, the return from the function was zero, implying that there were
520 other matches that did not fit. The correct "exactly full" value is now
521 returned.
522
523 6. If a subpattern that was called recursively or as a subroutine contained
524 (*PRUNE) or any other control that caused it to give a non-standard return,
525 invalid errors such as "Error -26 (nested recursion at the same subject
526 position)" or even infinite loops could occur.
527
528 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
529 computing the minimum length on reaching *ACCEPT, and so ended up with the
530 wrong value of 1 rather than 0. Further investigation indicates that
531 computing a minimum subject length in the presence of *ACCEPT is difficult
532 (think back references, subroutine calls), and so I have changed the code
533 so that no minimum is registered for a pattern that contains *ACCEPT.
534
535 8. If (*THEN) was present in the first (true) branch of a conditional group,
536 it was not handled as intended. [But see 16 below.]
537
538 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
539 Sheri Pierce.
540
541 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
542 the first byte in a match must be "a".
543
544 11. Change 17 for 8.13 increased the recursion depth for patterns like
545 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
546 pattern contains any instances of (*THEN). If it does not, the old
547 optimizations are restored. It would be nice to do this on a per-group
548 basis, but at the moment that is not feasible.
549
550 12. In some environments, the output of pcretest -C is CRLF terminated. This
551 broke RunTest's code that checks for the link size. A single white space
552 character after the value is now allowed for.
553
554 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
555 For "fr", it uses the Windows-specific input and output files.
556
557 14. If (*THEN) appeared in a group that was called recursively or as a
558 subroutine, it did not work as intended. [But see next item.]
559
560 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
561 pattern fragments (but not containing any | characters). If A and B are
562 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
563 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
564 D. In other words, Perl considers parentheses that do not contain any |
565 characters to be part of a surrounding alternative, whereas PCRE was
566 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
567 differently. PCRE now behaves in the same way as Perl, except in the case
568 of subroutine/recursion calls such as (?1) which have in any case always
569 been different (but PCRE had them first :-).
570
571 16. Related to 15 above: Perl does not treat the | in a conditional group as
572 creating alternatives. Such a group is treated in the same way as an
573 ordinary group without any | characters when processing (*THEN). PCRE has
574 been changed to match Perl's behaviour.
575
576 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
577 RunGrepTest script failed.
578
579 18. Change 22 for version 13 caused atomic groups to use more stack. This is
580 inevitable for groups that contain captures, but it can lead to a lot of
581 stack use in large patterns. The old behaviour has been restored for atomic
582 groups that do not contain any capturing parentheses.
583
584 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
585 suppress the check for a minimum subject length at run time. (If it was
586 given to pcre_exec() or pcre_dfa_exec() it did work.)
587
588 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
589 fail to work when decoding hex characters in data strings in EBCDIC
590 environments.
591
592 21. It appears that in at least one Mac OS environment, the isxdigit() function
593 is implemented as a macro that evaluates to its argument more than once,
594 contravening the C 90 Standard (I haven't checked a later standard). There
595 was an instance in pcretest which caused it to go wrong when processing
596 \x{...} escapes in subject strings. The has been rewritten to avoid using
597 things like p++ in the argument of isxdigit().
598
599
600 Version 8.13 16-Aug-2011
601 ------------------------
602
603 1. The Unicode data tables have been updated to Unicode 6.0.0.
604
605 2. Two minor typos in pcre_internal.h have been fixed.
606
607 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
608 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
609 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
610
611 4. There were a number of related bugs in the code for matching backrefences
612 caselessly in UTF-8 mode when codes for the characters concerned were
613 different numbers of bytes. For example, U+023A and U+2C65 are an upper
614 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
615 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
616 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
617 2-byte code at the end of the subject (it thought there wasn't enough data
618 left).
619
620 5. Comprehensive information about what went wrong is now returned by
621 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
622 as the output vector has at least 2 elements. The offset of the start of
623 the failing character and a reason code are placed in the vector.
624
625 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
626 now returned is for the first byte of the failing character, instead of the
627 last byte inspected. This is an incompatible change, but I hope it is small
628 enough not to be a problem. It makes the returned offset consistent with
629 pcre_exec() and pcre_dfa_exec().
630
631 7. pcretest now gives a text phrase as well as the error number when
632 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
633 failure, the offset and reason code are output.
634
635 8. When \R was used with a maximizing quantifier it failed to skip backwards
636 over a \r\n pair if the subsequent match failed. Instead, it just skipped
637 back over a single character (\n). This seems wrong (because it treated the
638 two characters as a single entity when going forwards), conflicts with the
639 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
640 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
641 has been changed.
642
643 9. Some internal refactoring has changed the processing so that the handling
644 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
645 time (the PCRE_DOTALL option was changed this way some time ago: version
646 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
647 which was always a bit of a fudge. It also means that there is one less
648 argument for the match() function, which reduces its stack requirements
649 slightly. This change also fixes an incompatibility with Perl: the pattern
650 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
651
652 10. More internal refactoring has drastically reduced the number of recursive
653 calls to match() for possessively repeated groups such as (abc)++ when
654 using pcre_exec().
655
656 11. While implementing 10, a number of bugs in the handling of groups were
657 discovered and fixed:
658
659 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
660 (a|)*(?1) gave a compile-time internal error.
661 ((a|)+)+ did not notice that the outer group could match an empty string.
662 (^a|^)+ was not marked as anchored.
663 (.*a|.*)+ was not marked as matching at start or after a newline.
664
665 12. Yet more internal refactoring has removed another argument from the match()
666 function. Special calls to this function are now indicated by setting a
667 value in a variable in the "match data" data block.
668
669 13. Be more explicit in pcre_study() instead of relying on "default" for
670 opcodes that mean there is no starting character; this means that when new
671 ones are added and accidentally left out of pcre_study(), testing should
672 pick them up.
673
674 14. The -s option of pcretest has been documented for ages as being an old
675 synonym of -m (show memory usage). I have changed it to mean "force study
676 for every regex", that is, assume /S for every regex. This is similar to -i
677 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
678 using it. It makes it easier to run collections of tests with and without
679 study enabled, and thereby test pcre_study() more easily. All the standard
680 tests are now run with and without -s (but some patterns can be marked as
681 "never study" - see 20 below).
682
683 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
684 restoration of the capturing data to the outer values was not happening
685 correctly.
686
687 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
688 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
689 pattern had matched an empty string, and so incorrectly returned a no
690 match.
691
692 17. There was optimizing code for the last branch of non-capturing parentheses,
693 and also for the obeyed branch of a conditional subexpression, which used
694 tail recursion to cut down on stack usage. Unfortunately, now that there is
695 the possibility of (*THEN) occurring in these branches, tail recursion is
696 no longer possible because the return has to be checked for (*THEN). These
697 two optimizations have therefore been removed. [But see 8.20/11 above.]
698
699 18. If a pattern containing \R was studied, it was assumed that \R always
700 matched two bytes, thus causing the minimum subject length to be
701 incorrectly computed because \R can also match just one byte.
702
703 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
704 was incorrectly computed.
705
706 20. If /S is present twice on a test pattern in pcretest input, it now
707 *disables* studying, thereby overriding the use of -s on the command line
708 (see 14 above). This is necessary for one or two tests to keep the output
709 identical in both cases.
710
711 21. When (*ACCEPT) was used in an assertion that matched an empty string and
712 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
713
714 22. When an atomic group that contained a capturing parenthesis was
715 successfully matched, but the branch in which it appeared failed, the
716 capturing was not being forgotten if a higher numbered group was later
717 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
718 group 1 to "a", when in fact it should be unset. This applied to multi-
719 branched capturing and non-capturing groups, repeated or not, and also to
720 positive assertions (capturing in negative assertions does not happen
721 in PCRE) and also to nested atomic groups.
722
723 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
724 subject after a captured substring, to make it easier to tell which of a
725 number of identical substrings has been captured.
726
727 24. The way atomic groups are processed by pcre_exec() has been changed so that
728 if they are repeated, backtracking one repetition now resets captured
729 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
730 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
731 "aaa". Previously, it would have been "a". As part of this code
732 refactoring, the way recursive calls are handled has also been changed.
733
734 25. If an assertion condition captured any substrings, they were not passed
735 back unless some other capturing happened later. For example, if
736 (?(?=(a))a) was matched against "a", no capturing was returned.
737
738 26. When studying a pattern that contained subroutine calls or assertions,
739 the code for finding the minimum length of a possible match was handling
740 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
741 group 1 called group 2 while simultaneously a separate group 2 called group
742 1). A stack overflow occurred in this case. I have fixed this by limiting
743 the recursion depth to 10.
744
745 27. Updated RunTest.bat in the distribution to the version supplied by Tom
746 Fortmann. This supports explicit test numbers on the command line, and has
747 argument validation and error reporting.
748
749 28. An instance of \X with an unlimited repeat could fail if at any point the
750 first character it looked at was a mark character.
751
752 29. Some minor code refactoring concerning Unicode properties and scripts
753 should reduce the stack requirement of match() slightly.
754
755 30. Added the '=' option to pcretest to check the setting of unused capturing
756 slots at the end of the pattern, which are documented as being -1, but are
757 not included in the return count.
758
759 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
760 compiled something random. Now it gives a compile-time error (as does
761 Perl).
762
763 32. A *MARK encountered during the processing of a positive assertion is now
764 recorded and passed back (compatible with Perl).
765
766 33. If --only-matching or --colour was set on a pcregrep call whose pattern
767 had alternative anchored branches, the search for a second match in a line
768 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
769 matched the line "0102" twice. The same bug affected patterns that started
770 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
771 twice.
772
773 34. Previously, PCRE did not allow quantification of assertions. However, Perl
774 does, and because of capturing effects, quantifying parenthesized
775 assertions may at times be useful. Quantifiers are now allowed for
776 parenthesized assertions.
777
778 35. A minor code tidy in pcre_compile() when checking options for \R usage.
779
780 36. \g was being checked for fancy things in a character class, when it should
781 just be a literal "g".
782
783 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
784 appearance of a nested POSIX class supersedes an apparent external class.
785 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
786 unescaped square brackets may also appear as part of class names. For
787 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
788 more like Perl. (But see 8.20/1 above.)
789
790 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
791 was because it thought it was \N{name}, which is not supported).
792
793 39. Add minix to OS list not supporting the -S option in pcretest.
794
795 40. PCRE tries to detect cases of infinite recursion at compile time, but it
796 cannot analyze patterns in sufficient detail to catch mutual recursions
797 such as ((?1))((?2)). There is now a runtime test that gives an error if a
798 subgroup is called recursively as a subpattern for a second time at the
799 same position in the subject string. In previous releases this might have
800 been caught by the recursion limit, or it might have run out of stack.
801
802 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
803 happen only once. PCRE was, however incorrectly giving a compile time error
804 "recursive call could loop indefinitely" because it cannot analyze the
805 pattern in sufficient detail. The compile time test no longer happens when
806 PCRE is compiling a conditional subpattern, but actual runaway loops are
807 now caught at runtime (see 40 above).
808
809 42. It seems that Perl allows any characters other than a closing parenthesis
810 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
811 has been changed to be the same.
812
813 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
814 as not to get warnings when autogen.sh is called. Also changed
815 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
816
817 44. To help people who use pcregrep to scan files containing exceedingly long
818 lines, the following changes have been made:
819
820 (a) The default value of the buffer size parameter has been increased from
821 8K to 20K. (The actual buffer used is three times this size.)
822
823 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
824 PCRE is built.
825
826 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
827 to be set at run time.
828
829 (d) Numerical values in pcregrep options can be followed by K or M, for
830 example --buffer-size=50K.
831
832 (e) If a line being scanned overflows pcregrep's buffer, an error is now
833 given and the return code is set to 2.
834
835 45. Add a pointer to the latest mark to the callout data block.
836
837 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
838 partial match of an empty string instead of no match. This was specific to
839 the use of ".".
840
841 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
842 complete match instead of a partial match. This bug was dependent on both
843 the PCRE_UTF8 and PCRE_DOTALL options being set.
844
845 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
846 starting byte set, because \b was not being ignored.
847
848
849 Version 8.12 15-Jan-2011
850 ------------------------
851
852 1. Fixed some typos in the markup of the man pages, and wrote a script that
853 checks for such things as part of the documentation building process.
854
855 2. On a big-endian 64-bit system, pcregrep did not correctly process the
856 --match-limit and --recursion-limit options (added for 8.11). In
857 particular, this made one of the standard tests fail. (The integer value
858 went into the wrong half of a long int.)
859
860 3. If the --colour option was given to pcregrep with -v (invert match), it
861 did strange things, either producing crazy output, or crashing. It should,
862 of course, ignore a request for colour when reporting lines that do not
863 match.
864
865 4. Another pcregrep bug caused similar problems if --colour was specified with
866 -M (multiline) and the pattern match finished with a line ending.
867
868 5. In pcregrep, when a pattern that ended with a literal newline sequence was
869 matched in multiline mode, the following line was shown as part of the
870 match. This seems wrong, so I have changed it.
871
872 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
873 the check for further matches in the same line (so they could be coloured)
874 to overrun the end of the current line. If another match was found, it was
875 incorrectly shown (and then shown again when found in the next line).
876
877 7. If pcregrep was compiled under Windows, there was a reference to the
878 function pcregrep_exit() before it was defined. I am assuming this was
879 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
880 reported by a user. I've moved the definition above the reference.
881
882
883 Version 8.11 10-Dec-2010
884 ------------------------
885
886 1. (*THEN) was not working properly if there were untried alternatives prior
887 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
888 backtracked to try for "b" instead of moving to the next alternative branch
889 at the same level (in this case, to look for "c"). The Perl documentation
890 is clear that when (*THEN) is backtracked onto, it goes to the "next
891 alternative in the innermost enclosing group".
892
893 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
894 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
895 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
896 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
897 (*THEN).
898
899 3. If \s appeared in a character class, it removed the VT character from
900 the class, even if it had been included by some previous item, for example
901 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
902 of \s, but is part of the POSIX "space" class.)
903
904 4. A partial match never returns an empty string (because you can always
905 match an empty string at the end of the subject); however the checking for
906 an empty string was starting at the "start of match" point. This has been
907 changed to the "earliest inspected character" point, because the returned
908 data for a partial match starts at this character. This means that, for
909 example, /(?<=abc)def/ gives a partial match for the subject "abc"
910 (previously it gave "no match").
911
912 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
913 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
914 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
915 has an implication that the given string is incomplete (because a partial
916 match is preferred over a full match). For this reason, these items now
917 give a partial match in this situation. [Aside: previously, the one case
918 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
919 match rather than a full match, which was wrong by the old rules, but is
920 now correct.]
921
922 6. There was a bug in the handling of #-introduced comments, recognized when
923 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
924 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
925 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
926 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
927 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
928 places in pcre_compile().
929
930 7. Related to (6) above, when pcre_compile() was skipping #-introduced
931 comments when looking ahead for named forward references to subpatterns,
932 the only newline sequence it recognized was NL. It now handles newlines
933 according to the set newline convention.
934
935 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
936 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
937 cater for a lack of strerror(). These oversights have been fixed.
938
939 9. Added --match-limit and --recursion-limit to pcregrep.
940
941 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
942
943 11. When the -o option was used, pcregrep was setting a return code of 1, even
944 when matches were found, and --line-buffered was not being honoured.
945
946 12. Added an optional parentheses number to the -o and --only-matching options
947 of pcregrep.
948
949 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
950 can match an empty string. The code to do it in pcretest and pcredemo
951 needed fixing:
952
953 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
954 only one byte after an empty string match just before CRLF (this case
955 just got forgotten; "any" and "anycrlf" were OK).
956
957 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
958 mode when an empty string match preceded an ASCII character followed by
959 a non-ASCII character. (The code for advancing by one character rather
960 than one byte was nonsense.)
961
962 (c) The pcredemo.c sample program did not have any code at all to handle
963 the cases when CRLF is a valid newline sequence.
964
965 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
966 as a starting offset was within the subject string. There is now a new
967 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
968 negative or greater than the length of the string. In order to test this,
969 pcretest is extended to allow the setting of negative starting offsets.
970
971 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
972 starting offset points to the beginning of a UTF-8 character was
973 unnecessarily clumsy. I tidied it up.
974
975 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
976 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
977
978 17. Nobody had reported that the --include_dir option, which was added in
979 release 7.7 should have been called --include-dir (hyphen, not underscore)
980 for compatibility with GNU grep. I have changed it to --include-dir, but
981 left --include_dir as an undocumented synonym, and the same for
982 --exclude-dir, though that is not available in GNU grep, at least as of
983 release 2.5.4.
984
985 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
986 characters from a string of bytes) have been redefined so as not to use
987 loops, in order to improve performance in some environments. At the same
988 time, I abstracted some of the common code into auxiliary macros to save
989 repetition (this should not affect the compiled code).
990
991 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
992 compile-time error is now given if \c is not followed by an ASCII
993 character, that is, a byte less than 128. (In EBCDIC mode, the code is
994 different, and any byte value is allowed.)
995
996 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
997 START_OPTIMIZE option, which is now allowed at compile time - but just
998 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
999 to pcregrep and other applications that have no direct access to PCRE
1000 options. The new /Y option in pcretest sets this option when calling
1001 pcre_compile().
1002
1003 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1004 back references. Groups containing recursive back references were forced to
1005 be atomic by that change, but in the case of named groups, the amount of
1006 memory required was incorrectly computed, leading to "Failed: internal
1007 error: code overflow". This has been fixed.
1008
1009 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1010 pcretest.c, to avoid build problems in some Borland environments.
1011
1012
1013 Version 8.10 25-Jun-2010
1014 ------------------------
1015
1016 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1017 THEN.
1018
1019 2. (*ACCEPT) was not working when inside an atomic group.
1020
1021 3. Inside a character class, \B is treated as a literal by default, but
1022 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1023 causes the error). The code is unchanged, but I tidied the documentation.
1024
1025 4. Inside a character class, PCRE always treated \R and \X as literals,
1026 whereas Perl faults them if its -w option is set. I have changed PCRE so
1027 that it faults them when PCRE_EXTRA is set.
1028
1029 5. Added support for \N, which always matches any character other than
1030 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1031
1032 6. When compiling pcregrep with newer versions of gcc which may have
1033 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1034 declared with attribute warn_unused_result" were given. Just casting the
1035 result to (void) does not stop the warnings; a more elaborate fudge is
1036 needed. I've used a macro to implement this.
1037
1038 7. Minor change to pcretest.c to avoid a compiler warning.
1039
1040 8. Added four artifical Unicode properties to help with an option to make
1041 \s etc use properties (see next item). The new properties are: Xan
1042 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1043
1044 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1045 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1046 this option. Modified pcretest to add /W to test this facility. Added
1047 REG_UCP to make it available via the POSIX interface.
1048
1049 10. Added --line-buffered to pcregrep.
1050
1051 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1052 studied, and the match started with a letter with a code point greater than
1053 127 whose first byte was different to the first byte of the other case of
1054 the letter, the other case of this starting letter was not recognized
1055 (#976).
1056
1057 12. If a pattern that was studied started with a repeated Unicode property
1058 test, for example, \p{Nd}+, there was the theoretical possibility of
1059 setting up an incorrect bitmap of starting bytes, but fortunately it could
1060 not have actually happened in practice until change 8 above was made (it
1061 added property types that matched character-matching opcodes).
1062
1063 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1064 possible starting bytes for non-anchored patterns.
1065
1066 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1067 \R, and also a number of cases that involve Unicode properties, both
1068 explicit and implicit when PCRE_UCP is set.
1069
1070 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1071 input, it could crash or give wrong results if characters with values
1072 greater than 0xc0 were present in the subject string. (Detail: it assumed
1073 UTF-8 input when processing these items.)
1074
1075 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1076 size_t is 64-bit (#991).
1077
1078 17. Added a check for running out of memory when PCRE is compiled with
1079 --disable-stack-for-recursion (#990).
1080
1081 18. If the last data line in a file for pcretest does not have a newline on
1082 the end, a newline was missing in the output.
1083
1084 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1085 less than 128) in its various bitmaps. However, there is a facility for
1086 generating tables according to the current locale when PCRE is compiled. It
1087 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1088 characters, are recognized by isspace() and therefore were getting set in
1089 these tables, and indeed these tables seem to approximate to ISO 8859. This
1090 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1091 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1092 which of course cannot start UTF-8 characters. I have changed the code so
1093 that only real ASCII characters (less than 128) and the correct starting
1094 bytes for UTF-8 encodings are set for characters greater than 127 when in
1095 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1096 altogether.)
1097
1098 20. Added the /T option to pcretest so as to be able to run tests with non-
1099 standard character tables, thus making it possible to include the tests
1100 used for 19 above in the standard set of tests.
1101
1102 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1103 reference to a subpattern the other side of a comment that contains an
1104 opening parenthesis caused either an internal compiling error, or a
1105 reference to the wrong subpattern.
1106
1107
1108 Version 8.02 19-Mar-2010
1109 ------------------------
1110
1111 1. The Unicode data tables have been updated to Unicode 5.2.0.
1112
1113 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1114 configured.
1115
1116 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1117 original author of that file, following a query about its status.
1118
1119 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1120 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1121
1122 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1123 quantifier applied to a forward-referencing subroutine call, could compile
1124 incorrect code or give the error "internal error: previously-checked
1125 referenced subpattern not found".
1126
1127 6. Both MS Visual Studio and Symbian OS have problems with initializing
1128 variables to point to external functions. For these systems, therefore,
1129 pcre_malloc etc. are now initialized to local functions that call the
1130 relevant global functions.
1131
1132 7. There were two entries missing in the vectors called coptable and poptable
1133 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1134 I've fixed the data, and added a kludgy way of testing at compile time that
1135 the lengths are correct (equal to the number of opcodes).
1136
1137 8. Following on from 7, I added a similar kludge to check the length of the
1138 eint vector in pcreposix.c.
1139
1140 9. Error texts for pcre_compile() are held as one long string to avoid too
1141 much relocation at load time. To find a text, the string is searched,
1142 counting zeros. There was no check for running off the end of the string,
1143 which could happen if a new error number was added without updating the
1144 string.
1145
1146 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1147
1148 11. \K was not working if it appeared in an atomic group or in a group that
1149 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1150 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1151 the assertion is positive, but not if it is negative.
1152
1153 12. Change 11 fortuitously reduced the size of the stack frame used in the
1154 "match()" function of pcre_exec.c by one pointer. Forthcoming
1155 implementation of support for (*MARK) will need an extra pointer on the
1156 stack; I have reserved it now, so that the stack frame size does not
1157 decrease.
1158
1159 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1160 item in branch that calls a recursion is a subroutine call - as in the
1161 second branch in the above example - was incorrectly given the compile-
1162 time error "recursive call could loop indefinitely" because pcre_compile()
1163 was not correctly checking the subroutine for matching a non-empty string.
1164
1165 14. The checks for overrunning compiling workspace could trigger after an
1166 overrun had occurred. This is a "should never occur" error, but it can be
1167 triggered by pathological patterns such as hundreds of nested parentheses.
1168 The checks now trigger 100 bytes before the end of the workspace.
1169
1170 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1171
1172
1173 Version 8.01 19-Jan-2010
1174 ------------------------
1175
1176 1. If a pattern contained a conditional subpattern with only one branch (in
1177 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1178 computed the wrong minimum data length (which is of course zero for such
1179 subpatterns). This could cause incorrect "no match" results.
1180
1181 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1182 the pattern is reset in the first branch, pcre_compile() failed with
1183 "internal error: code overflow at offset...". This happened only when
1184 the reset was to the original external option setting. (An optimization
1185 abstracts leading options settings into an external setting, which was the
1186 cause of this.)
1187
1188 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1189 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1190 assertion pattern did not match (meaning that the assertion was true), it
1191 was incorrectly treated as false if the SKIP had been reached during the
1192 matching. This also applied to assertions used as conditions.
1193
1194 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1195 assertion subpattern, including such a pattern used as a condition,
1196 unpredictable results occurred, instead of the error return
1197 PCRE_ERROR_DFA_UITEM.
1198
1199 5. The C++ GlobalReplace function was not working like Perl for the special
1200 situation when an empty string is matched. It now does the fancy magic
1201 stuff that is necessary.
1202
1203 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1204 removed. (These were left over from very, very early versions of PCRE.)
1205
1206 7. Some cosmetic changes to the code to make life easier when compiling it
1207 as part of something else:
1208
1209 (a) Change DEBUG to PCRE_DEBUG.
1210
1211 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1212 called "current" as "current_branch", to prevent a collision with the
1213 Linux macro when compiled as a kernel module.
1214
1215 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1216 prevent a collision with the Linux macro when compiled as a kernel
1217 module.
1218
1219 8. In pcre_compile() there are some checks for integer overflows that used to
1220 cast potentially large values to (double). This has been changed to that
1221 when building, a check for int64_t is made, and if it is found, it is used
1222 instead, thus avoiding the use of floating point arithmetic. (There is no
1223 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1224 double.
1225
1226 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1227 2005 (difference between two addresses compared to an unsigned value).
1228
1229 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1230 custom one, because of the following reported problem in Windows:
1231
1232 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1233 under Win32.
1234 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1235 therefore missing the function definition.
1236 - The compiler thus generates a "C" signature for the test function.
1237 - The linker fails to find the "C" function.
1238 - PCRE fails to configure if asked to do so against libbz2.
1239
1240 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1241 messages were output:
1242
1243 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1244 rerunning libtoolize, to keep the correct libtool macros in-tree.
1245 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1246
1247 I have done both of these things.
1248
1249 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1250 most of the time, it *can* run out if it is given a pattern that contains a
1251 runaway infinite recursion. I updated the discussion in the pcrestack man
1252 page.
1253
1254 13. Now that we have gone to the x.xx style of version numbers, the minor
1255 version may start with zero. Using 08 or 09 is a bad idea because users
1256 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1257 interpreted as invalid octal numbers. I've updated the previous comment in
1258 configure.ac, and also added a check that gives an error if 08 or 09 are
1259 used.
1260
1261 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1262 causing partial matching to fail when the end of the subject matched \W
1263 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1264
1265 15. There were some discrepancies between the declarations in pcre_internal.h
1266 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1267 their definitions. The declarations used "const uschar *" and the
1268 definitions used USPTR. Even though USPTR is normally defined as "const
1269 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1270 reported that: "This difference in casting confuses some C++ compilers, for
1271 example, SunCC recognizes above declarations as different functions and
1272 generates broken code for hbpcre." I have changed the declarations to use
1273 USPTR.
1274
1275 16. GNU libtool is named differently on some systems. The autogen.sh script now
1276 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1277 (FreeBSD).
1278
1279 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1280 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1281 comment: "Figure out how to create a longlong from a string: strtoll and
1282 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1283 instance, but it only takes 2 args instead of 3!"
1284
1285 18. A subtle bug concerned with back references has been fixed by a change of
1286 specification, with a corresponding code fix. A pattern such as
1287 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1288 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1289 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1290 same bug. Such groups have to be quantified to be useful, or contained
1291 inside another quantified group. (If there's no repetition, the reference
1292 can never match.) The problem arises because, having left the group and
1293 moved on to the rest of the pattern, a later failure that backtracks into
1294 the group uses the captured value from the final iteration of the group
1295 rather than the correct earlier one. I have fixed this in PCRE by forcing
1296 any group that contains a reference to itself to be an atomic group; that
1297 is, there cannot be any backtracking into it once it has completed. This is
1298 similar to recursive and subroutine calls.
1299
1300
1301 Version 8.00 19-Oct-09
1302 ----------------------
1303
1304 1. The table for translating pcre_compile() error codes into POSIX error codes
1305 was out-of-date, and there was no check on the pcre_compile() error code
1306 being within the table. This could lead to an OK return being given in
1307 error.
1308
1309 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1310 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1311 in a Windows environment.
1312
1313 3. The pcregrep --count option prints the count for each file even when it is
1314 zero, as does GNU grep. However, pcregrep was also printing all files when
1315 --files-with-matches was added. Now, when both options are given, it prints
1316 counts only for those files that have at least one match. (GNU grep just
1317 prints the file name in this circumstance, but including the count seems
1318 more useful - otherwise, why use --count?) Also ensured that the
1319 combination -clh just lists non-zero counts, with no names.
1320
1321 4. The long form of the pcregrep -F option was incorrectly implemented as
1322 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1323 but it seems right to fix it, and I didn't think it was worth preserving
1324 the old behaviour.
1325
1326 5. The command line items --regex=pattern and --regexp=pattern were not
1327 recognized by pcregrep, which required --regex pattern or --regexp pattern
1328 (with a space rather than an '='). The man page documented the '=' forms,
1329 which are compatible with GNU grep; these now work.
1330
1331 6. No libpcreposix.pc file was created for pkg-config; there was just
1332 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1333
1334 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1335 when UCP support is not needed, by modifying the Python script that
1336 generates it from Unicode data files. This should not matter if the module
1337 is correctly used as a library, but I received one complaint about 50K of
1338 unwanted data. My guess is that the person linked everything into his
1339 program rather than using a library. Anyway, it does no harm.
1340
1341 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1342 was a minimum greater than 1 for a wide character in a possessive
1343 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1344 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1345 character. Chaos in the form of incorrect output or a compiling loop could
1346 result.
1347
1348 9. The restrictions on what a pattern can contain when partial matching is
1349 requested for pcre_exec() have been removed. All patterns can now be
1350 partially matched by this function. In addition, if there are at least two
1351 slots in the offset vector, the offset of the earliest inspected character
1352 for the match and the offset of the end of the subject are set in them when
1353 PCRE_ERROR_PARTIAL is returned.
1354
1355 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1356 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1357 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1358 and may be more useful for multi-segment matching.
1359
1360 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1361 used to be given if ever the end of the subject was reached; now it is
1362 given only if matching could not proceed because another character was
1363 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1364 string "Z", which now yields "no match" instead of "partial match". In the
1365 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1366 final character ended with (*FAIL).
1367
1368 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1369 if the pattern had a "must contain" character that was already found in the
1370 earlier partial match, unless partial matching was again requested. For
1371 example, with the pattern /dog.(body)?/, the "must contain" character is
1372 "g". If the first part-match was for the string "dog", restarting with
1373 "sbody" failed. This bug has been fixed.
1374
1375 13. The string returned by pcre_dfa_exec() after a partial match has been
1376 changed so that it starts at the first inspected character rather than the
1377 first character of the match. This makes a difference only if the pattern
1378 starts with a lookbehind assertion or \b or \B (\K is not supported by
1379 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1380 matching functions compatible, and I think it's the right thing to do.
1381
1382 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1383 so that the demonstration program is easily available in environments where
1384 PCRE has not been installed from source.
1385
1386 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1387 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1388 library.
1389
1390 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1391 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1392 is not the first non-POSIX option to be added. Clearly some people find
1393 these options useful.
1394
1395 17. If a caller to the POSIX matching function regexec() passes a non-zero
1396 value for nmatch with a NULL value for pmatch, the value of
1397 nmatch is forced to zero.
1398
1399 18. RunGrepTest did not have a test for the availability of the -u option of
1400 the diff command, as RunTest does. It now checks in the same way as
1401 RunTest, and also checks for the -b option.
1402
1403 19. If an odd number of negated classes containing just a single character
1404 interposed, within parentheses, between a forward reference to a named
1405 subpattern and the definition of the subpattern, compilation crashed with
1406 an internal error, complaining that it could not find the referenced
1407 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1408 [The bug was that it was starting one character too far in when skipping
1409 over the character class, thus treating the ] as data rather than
1410 terminating the class. This meant it could skip too much.]
1411
1412 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1413 /g option in pcretest when the pattern contains \K, which makes it possible
1414 to have an empty string match not at the start, even when the pattern is
1415 anchored. Updated pcretest and pcredemo to use this option.
1416
1417 21. If the maximum number of capturing subpatterns in a recursion was greater
1418 than the maximum at the outer level, the higher number was returned, but
1419 with unset values at the outer level. The correct (outer level) value is
1420 now given.
1421
1422 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1423 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1424 make it do so. The string so far is captured, making this feature
1425 compatible with Perl.
1426
1427 23. The tests have been re-organized, adding tests 11 and 12, to make it
1428 possible to check the Perl 5.10 features against Perl 5.10.
1429
1430 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1431 pattern matches a fixed length string. PCRE did not allow this; now it
1432 does. Neither allows recursion.
1433
1434 25. I finally figured out how to implement a request to provide the minimum
1435 length of subject string that was needed in order to match a given pattern.
1436 (It was back references and recursion that I had previously got hung up
1437 on.) This code has now been added to pcre_study(); it finds a lower bound
1438 to the length of subject needed. It is not necessarily the greatest lower
1439 bound, but using it to avoid searching strings that are too short does give
1440 some useful speed-ups. The value is available to calling programs via
1441 pcre_fullinfo().
1442
1443 26. While implementing 25, I discovered to my embarrassment that pcretest had
1444 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1445 study optimizations had never been tested with that matching function.
1446 Oops. What is worse, even when it was passed study data, there was a bug in
1447 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1448 were also very few tests of studied patterns with pcre_dfa_exec().
1449
1450 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1451 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1452 on the other side of the coin, they are no longer allowed to have different
1453 names, because these cannot be distinguished in PCRE, and this has caused
1454 confusion. (This is a difference from Perl.)
1455
1456 28. When duplicate subpattern names are present (necessarily with different
1457 numbers, as required by 27 above), and a test is made by name in a
1458 conditional pattern, either for a subpattern having been matched, or for
1459 recursion in such a pattern, all the associated numbered subpatterns are
1460 tested, and the overall condition is true if the condition is true for any
1461 one of them. This is the way Perl works, and is also more like the way
1462 testing by number works.
1463
1464
1465 Version 7.9 11-Apr-09
1466 ---------------------
1467
1468 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1469 (pcretest), all targets were linked against these libraries. This included
1470 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1471 libraries. This caused unwanted dependencies to be created. This problem
1472 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1473 pcretest is linked with readline.
1474
1475 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1476 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1477 moved outside it again, because FALSE and TRUE are already defined in AIX,
1478 but BOOL is not.
1479
1480 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1481 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1482
1483 4. The pcregrep documentation said spaces were inserted as well as colons (or
1484 hyphens) following file names and line numbers when outputting matching
1485 lines. This is not true; no spaces are inserted. I have also clarified the
1486 wording for the --colour (or --color) option.
1487
1488 5. In pcregrep, when --colour was used with -o, the list of matching strings
1489 was not coloured; this is different to GNU grep, so I have changed it to be
1490 the same.
1491
1492 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1493 each matching line was coloured. Now it goes on to look for further matches
1494 of any of the test patterns, which is the same behaviour as GNU grep.
1495
1496 7. A pattern that could match an empty string could cause pcregrep to loop; it
1497 doesn't make sense to accept an empty string match in pcregrep, so I have
1498 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1499 seems to be how GNU grep behaves.
1500
1501 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1502 start or after a newline", because the conditional assertion was not being
1503 correctly handled. The rule now is that both the assertion and what follows
1504 in the first alternative must satisfy the test.
1505
1506 9. If auto-callout was enabled in a pattern with a conditional group whose
1507 condition was an assertion, PCRE could crash during matching, both with
1508 pcre_exec() and pcre_dfa_exec().
1509
1510 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1511 used for matching.
1512
1513 11. Unicode property support in character classes was not working for
1514 characters (bytes) greater than 127 when not in UTF-8 mode.
1515
1516 12. Added the -M command line option to pcretest.
1517
1518 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1519
1520 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1521
1522 16. Added comments and documentation about mis-use of no_arg in the C++
1523 wrapper.
1524
1525 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1526 from Martin Jerabek that uses macro names for all relevant character and
1527 string constants.
1528
1529 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1530 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1531 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1532 these, but not everybody uses configure.
1533
1534 19. A conditional group that had only one branch was not being correctly
1535 recognized as an item that could match an empty string. This meant that an
1536 enclosing group might also not be so recognized, causing infinite looping
1537 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1538 with the subject "ab", where knowledge that the repeated group can match
1539 nothing is needed in order to break the loop.
1540
1541 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1542 exec(), but without supplying a callout function, matching went wrong.
1543
1544 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1545 leak if the size of the offset vector was greater than 30. When the vector
1546 is smaller, the saved offsets during recursion go onto a local stack
1547 vector, but for larger vectors malloc() is used. It was failing to free
1548 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1549 error, in fact).
1550
1551 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1552 heapframe that is used only when UTF-8 support is enabled. This caused no
1553 problem, but was untidy.
1554
1555 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1556 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1557 included within another project.
1558
1559 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1560 slightly modified by me:
1561
1562 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1563 not building pcregrep.
1564
1565 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1566 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1567
1568 25. Forward references, both numeric and by name, in patterns that made use of
1569 duplicate group numbers, could behave incorrectly or give incorrect errors,
1570 because when scanning forward to find the reference group, PCRE was not
1571 taking into account the duplicate group numbers. A pattern such as
1572 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1573
1574 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1575 the feature of a custom pointer more persuasive (as requested by a user).
1576
1577 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1578 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1579 pre-defined.
1580
1581 28. Added support for (*UTF8) at the start of a pattern.
1582
1583 29. Arrange for flags added by the "release type" setting in CMake to be shown
1584 in the configuration summary.
1585
1586
1587 Version 7.8 05-Sep-08
1588 ---------------------
1589
1590 1. Replaced UCP searching code with optimized version as implemented for Ad
1591 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1592 stage table and inline lookup instead of a function, giving speed ups of 2
1593 to 5 times on some simple patterns that I tested. Permission was given to
1594 distribute the MultiStage2.py script that generates the tables (it's not in
1595 the tarball, but is in the Subversion repository).
1596
1597 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1598 scripts.
1599
1600 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1601 a group with a zero qualifier. The result of the study could be incorrect,
1602 or the function might crash, depending on the pattern.
1603
1604 4. Caseless matching was not working for non-ASCII characters in back
1605 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1606 It now works when Unicode Property Support is available.
1607
1608 5. In pcretest, an escape such as \x{de} in the data was always generating
1609 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1610 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1611 truncation.
1612
1613 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1614
1615 7. Added two (int) casts to pcregrep when printing the difference of two
1616 pointers, in case they are 64-bit values.
1617
1618 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1619 test 2 if it fails.
1620
1621 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1622 and a #define of that name to empty if it is not externally set. This is to
1623 allow users of MSVC to set it if necessary.
1624
1625 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1626 the convenience functions in the pcre_get.c source file.
1627
1628 11. An option change at the start of a pattern that had top-level alternatives
1629 could cause overwriting and/or a crash. This command provoked a crash in
1630 some environments:
1631
1632 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1633
1634 This potential security problem was recorded as CVE-2008-2371.
1635
1636 12. For a pattern where the match had to start at the beginning or immediately
1637 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1638 pcre_dfa_exec() could read past the end of the passed subject if there was
1639 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1640 pcretest so that it places the subject at the end of its malloc-ed buffer.
1641
1642 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1643 exec() might read past the end of the data buffer in UTF-8 mode.
1644
1645 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1646 the data contained the byte 0x85 as part of a UTF-8 character within its
1647 first line. This applied both to normal and DFA matching.
1648
1649 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1650 /^[^d]*?$/8 failed to match "abc".
1651
1652 16. Added a missing copyright notice to pcrecpp_internal.h.
1653
1654 17. Make it more clear in the documentation that values returned from
1655 pcre_exec() in ovector are byte offsets, not character counts.
1656
1657 18. Tidied a few places to stop certain compilers from issuing warnings.
1658
1659 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1660 supplied by Stefan Weber. I made a further small update for 7.8 because
1661 there is a change of source arrangements: the pcre_searchfuncs.c module is
1662 replaced by pcre_ucd.c.
1663
1664
1665 Version 7.7 07-May-08
1666 ---------------------
1667
1668 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1669 a string to a long long, pretend we don't even have a long long." This is
1670 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1671
1672 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1673 pre-7.6 versions, which defined a global no_arg variable instead of putting
1674 it in the RE class. (See also #8 below.)
1675
1676 3. Remove a line of dead code, identified by coverity and reported by Nuno
1677 Lopes.
1678
1679 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1680
1681 (1) The include/exclude patterns were being applied to the whole pathnames
1682 of files, instead of just to the final components.
1683
1684 (2) If there was more than one level of directory, the subdirectories were
1685 skipped unless they satisfied the include/exclude conditions. This is
1686 inconsistent with GNU grep (and could even be seen as contrary to the
1687 pcregrep specification - which I improved to make it absolutely clear).
1688 The action now is always to scan all levels of directory, and just
1689 apply the include/exclude patterns to regular files.
1690
1691 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1692 --exclude_dir in the tests to avoid scanning .svn directories.
1693
1694 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1695 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1696 doesn't support NULs in patterns.
1697
1698 7. Added some missing "const"s to declarations of static tables in
1699 pcre_compile.c and pcre_dfa_exec.c.
1700
1701 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1702 caused by fix #2 above. (Subsequently also a second patch to fix the
1703 first patch. And a third patch - this was a messy problem.)
1704
1705 9. Applied Craig's patch to remove the use of push_back().
1706
1707 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1708 matching function regexec().
1709
1710 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1711 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1712 references. PCRE supports relative numbers with this syntax (I don't think
1713 Oniguruma does).
1714
1715 12. Previously, a group with a zero repeat such as (...){0} was completely
1716 omitted from the compiled regex. However, this means that if the group
1717 was called as a subroutine from elsewhere in the pattern, things went wrong
1718 (an internal error was given). Such groups are now left in the compiled
1719 pattern, with a new opcode that causes them to be skipped at execution
1720 time.
1721
1722 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1723 to the way PCRE behaves:
1724
1725 (a) A lone ] character is dis-allowed (Perl treats it as data).
1726
1727 (b) A back reference to an unmatched subpattern matches an empty string
1728 (Perl fails the current match path).
1729
1730 (c) A data ] in a character class must be notated as \] because if the
1731 first data character in a class is ], it defines an empty class. (In
1732 Perl it is not possible to have an empty class.) The empty class []
1733 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1734 The negative empty class [^] matches any one character, independently
1735 of the DOTALL setting.
1736
1737 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1738 non-existent subpattern following a character class starting with ']' and
1739 containing () gave an internal compiling error instead of "reference to
1740 non-existent subpattern". Fortunately, when the pattern did exist, the
1741 compiled code was correct. (When scanning forwards to check for the
1742 existencd of the subpattern, it was treating the data ']' as terminating
1743 the class, so got the count wrong. When actually compiling, the reference
1744 was subsequently set up correctly.)
1745
1746 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1747 it was being rejected as not supported by pcre_dfa_exec(), even though
1748 other assertions are supported. I have made pcre_dfa_exec() support
1749 (*FAIL).
1750
1751 16. The implementation of 13c above involved the invention of a new opcode,
1752 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1753 cannot be changed at match time, I realized I could make a small
1754 improvement to matching performance by compiling OP_ALLANY instead of
1755 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1756 on the OP_ANY path.
1757
1758 17. Compiling pcretest on Windows with readline support failed without the
1759 following two fixes: (1) Make the unistd.h include conditional on
1760 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1761
1762 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1763 ncurses library to be included for pcretest when ReadLine support is
1764 requested, but also to allow for it to be overridden. This patch came from
1765 Daniel Bergström.
1766
1767 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1768 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1769 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1770 spotting this.
1771
1772
1773 Version 7.6 28-Jan-08
1774 ---------------------
1775
1776 1. A character class containing a very large number of characters with
1777 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1778 overflow.
1779
1780 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1781 HAVE_LONG_LONG is not defined.
1782
1783 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1784 bring them up to date and include new features. This patch includes:
1785
1786 - Fixed PH's badly added libz and libbz2 support.
1787 - Fixed a problem with static linking.
1788 - Added pcredemo. [But later removed - see 7 below.]
1789 - Fixed dftables problem and added an option.
1790 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1791 HAVE_LONG_LONG.
1792 - Added readline support for pcretest.
1793 - Added an listing of the option settings after cmake has run.
1794
1795 4. A user submitted a patch to Makefile that makes it easy to create
1796 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1797 Makefile.am that cause it to include this special target, without
1798 affecting anything else. Note that the same mingw target plus all
1799 the other distribution libraries and programs are now supported
1800 when configuring with CMake (see 6 below) instead of with
1801 Configure/Make.
1802
1803 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1804 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1805 exported in the Windows port". It has not yet been confirmed that the patch
1806 solves the problem, but it does no harm.
1807
1808 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1809 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1810 with CMake, and also correct the comment about stack recursion.
1811
1812 7. Remove the automatic building of pcredemo from the ./configure system and
1813 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1814 of a program that users should build themselves after PCRE is installed, so
1815 building it automatically is not really right. What is more, it gave
1816 trouble in some build environments.
1817
1818 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1819
1820
1821 Version 7.5 10-Jan-08
1822 ---------------------
1823
1824 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1825 values in parens when parsing an RE using the C++ wrapper."
1826
1827 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1828 Characters greater than 255 were excluded from the class instead of being
1829 included.
1830
1831 3. The same bug as (2) above applied to negated POSIX classes such as
1832 [:^space:].
1833
1834 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1835 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1836 I have changed it.
1837
1838 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1839 first named subpattern) and a construct such as (?&a) would reference the
1840 first named subpattern whose name started with "a" (in other words, the
1841 length check was missing). Both these problems are fixed. "Subpattern name
1842 expected" is now given for (?&) (a zero-length name), and this patch also
1843 makes it give the same error for \k'' (previously it complained that that
1844 was a reference to a non-existent subpattern).
1845
1846 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1847 this is right because (?- can be followed by option settings as well as by
1848 digits. I have, however, made the messages clearer.
1849
1850 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1851 than the number used in the conditional) now cause a compile-time error.
1852 This is actually not compatible with Perl, which accepts such patterns, but
1853 treats the conditional as always being FALSE (as PCRE used to), but it
1854 seems to me that giving a diagnostic is better.
1855
1856 8. Change "alphameric" to the more common word "alphanumeric" in comments
1857 and messages.
1858
1859 9. Fix two occurrences of "backslash" in comments that should have been
1860 "backspace".
1861
1862 10. Remove two redundant lines of code that can never be obeyed (their function
1863 was moved elsewhere).
1864
1865 11. The program that makes PCRE's Unicode character property table had a bug
1866 which caused it to generate incorrect table entries for sequences of
1867 characters that have the same character type, but are in different scripts.
1868 It amalgamated them into a single range, with the script of the first of
1869 them. In other words, some characters were in the wrong script. There were
1870 thirteen such cases, affecting characters in the following ranges:
1871
1872 U+002b0 - U+002c1
1873 U+0060c - U+0060d
1874 U+0061e - U+00612
1875 U+0064b - U+0065e
1876 U+0074d - U+0076d
1877 U+01800 - U+01805
1878 U+01d00 - U+01d77
1879 U+01d9b - U+01dbf
1880 U+0200b - U+0200f
1881 U+030fc - U+030fe
1882 U+03260 - U+0327f
1883 U+0fb46 - U+0fbb1
1884 U+10450 - U+1049d
1885
1886 12. The -o option (show only the matching part of a line) for pcregrep was not
1887 compatible with GNU grep in that, if there was more than one match in a
1888 line, it showed only the first of them. It now behaves in the same way as
1889 GNU grep.
1890
1891 13. If the -o and -v options were combined for pcregrep, it printed a blank
1892 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1893 does the same. The return code can be used to tell if there were any
1894 non-matching lines.
1895
1896 14. Added --file-offsets and --line-offsets to pcregrep.
1897
1898 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1899 infinitely looping recursion. The bug was that positive lookaheads were not
1900 being skipped when checking for a possible empty match (negative lookaheads
1901 and both kinds of lookbehind were skipped).
1902
1903 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1904 inclusion of <windows.h> to before rather than after the definition of
1905 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1906
1907 17. Specifying a possessive quantifier with a specific limit for a Unicode
1908 character property caused pcre_compile() to compile bad code, which led at
1909 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1910 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1911 caused the error; without that there was no problem.
1912
1913 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1914
1915 19. Added --enable-pcretest-libreadline.
1916
1917 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1918 RE::GlobalReplace(). As a result, the number of replacements returned was
1919 double what it should be. I removed one of the increments, but Craig sent a
1920 later patch that removed the other one (the right fix) and added unit tests
1921 that check the return values (which was not done before).
1922
1923 21. Several CMake things:
1924
1925 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1926 the names libpcre and libpcreposix, not just pcre and pcreposix.
1927
1928 (2) The above change means that pcretest and pcregrep are now correctly
1929 linked with the newly-built libraries, not previously installed ones.
1930
1931 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1932
1933 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1934 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1935 UTF-8 newline character). The key issue is that the pattern starts .*;
1936 this means that the match must be either at the beginning, or after a
1937 newline. The bug was in the code for advancing after a failed match and
1938 checking that the new position followed a newline. It was not taking
1939 account of UTF-8 characters correctly.
1940
1941 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1942 character classes. PCRE was not treating the sequence [:...:] as a
1943 character class unless the ... were all letters. Perl, however, seems to
1944 allow any characters between [: and :], though of course it rejects as
1945 unknown any "names" that contain non-letters, because all the known class
1946 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1947 for example, whereas PCRE did not - it did not recognize a POSIX character
1948 class. This seemed a bit dangerous, so the code has been changed to be
1949 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1950 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1951 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1952 Perl does, and where it didn't before.
1953
1954 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1955 Windows environments %n is disabled by default.
1956
1957
1958 Version 7.4 21-Sep-07
1959 ---------------------
1960
1961 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1962 means that a class such as [\s] counted as "explicit reference to CR or
1963 LF". That isn't really right - the whole point of the change was to try to
1964 help when there was an actual mention of one of the two characters. So now
1965 the change happens only if \r or \n (or a literal CR or LF) character is
1966 encountered.
1967
1968 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1969 of both had grown to the point where there were only 3 bits left.
1970 Fortunately, there was spare space in the data structure, and so I have
1971 moved the internal flags into a new 16-bit field to free up more option
1972 bits.
1973
1974 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1975 but did not set the internal JCHANGED flag - either of these is enough to
1976 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1977 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1978 start sets both bits.
1979
1980 4. Added options (at build time, compile time, exec time) to change \R from
1981 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1982
1983 5. doc/pcresyntax.html was missing from the distribution.
1984
1985 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1986 compatibility, even though it is no longer used.
1987
1988 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1989 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1990 windows.h file is present (where different names are used). [This was
1991 reversed later after testing - see 16 below.]
1992
1993 8. Changed all #include <config.h> to #include "config.h". There were also
1994 some further <pcre.h> cases that I changed to "pcre.h".
1995
1996 9. When pcregrep was used with the --colour option, it missed the line ending
1997 sequence off the lines that it output.
1998
1999 10. It was pointed out to me that arrays of string pointers cause lots of
2000 relocations when a shared library is dynamically loaded. A technique of
2001 using a single long string with a table of offsets can drastically reduce
2002 these. I have refactored PCRE in four places to do this. The result is
2003 dramatic:
2004
2005 Originally: 290
2006 After changing UCP table: 187
2007 After changing error message table: 43
2008 After changing table of "verbs" 36
2009 After changing table of Posix names 22
2010
2011 Thanks to the folks working on Gregex for glib for this insight.
2012
2013 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2014 unicode-properties was also set.
2015
2016 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2017
2018 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2019 checked only for CRLF.
2020
2021 14. Added casts to pcretest.c to avoid compiler warnings.
2022
2023 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2024
2025 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2026 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2027 entirely. This removes changes made in 7 above.
2028
2029 17. The CMake files have been updated, and there is now more information about
2030 building with CMake in the NON-UNIX-USE document.
2031
2032
2033 Version 7.3 28-Aug-07
2034 ---------------------
2035
2036 1. In the rejigging of the build system that eventually resulted in 7.1, the
2037 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2038 brackets there is not right, since it causes compilers to look for an
2039 installed pcre.h, not the version that is in the source that is being
2040 compiled (which of course may be different). I have changed it back to:
2041
2042 #include "pcre.h"
2043
2044 I have a vague recollection that the change was concerned with compiling in
2045 different directories, but in the new build system, that is taken care of
2046 by the VPATH setting the Makefile.
2047
2048 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2049 when the subject happened to end in the byte 0x85 (e.g. if the last
2050 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2051 characters but of course it shouldn't be taken as a newline when it is part
2052 of another character. The bug was that, for an unlimited repeat of . in
2053 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2054 characters when looking for a newline.
2055
2056 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2057
2058 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2059 in debug output.
2060
2061 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2062 long printing in the pcrecpp unittest when running under MinGW.
2063
2064 6. ESC_K was left out of the EBCDIC table.
2065
2066 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2067 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2068 limit also applies to "virtual nesting" when a pattern is recursive, and in
2069 this case 1000 isn't so big. I have been able to remove this limit at the
2070 expense of backing off one optimization in certain circumstances. Normally,
2071 when pcre_exec() would call its internal match() function recursively and
2072 immediately return the result unconditionally, it uses a "tail recursion"
2073 feature to save stack. However, when a subpattern that can match an empty
2074 string has an unlimited repetition quantifier, it no longer makes this
2075 optimization. That gives it a stack frame in which to save the data for
2076 checking that an empty string has been matched. Previously this was taken
2077 from the 1000-entry workspace that had been reserved. So now there is no
2078 explicit limit, but more stack is used.
2079
2080 8. Applied Daniel's patches to solve problems with the import/export magic
2081 syntax that is required for Windows, and which was going wrong for the
2082 pcreposix and pcrecpp parts of the library. These were overlooked when this
2083 problem was solved for the main library.
2084
2085 9. There were some crude static tests to avoid integer overflow when computing
2086 the size of patterns that contain repeated groups with explicit upper
2087 limits. As the maximum quantifier is 65535, the maximum group length was
2088 set at 30,000 so that the product of these two numbers did not overflow a
2089 32-bit integer. However, it turns out that people want to use groups that
2090 are longer than 30,000 bytes (though not repeat them that many times).
2091 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2092 made it possible to implement the integer overflow checks in a much more
2093 dynamic way, which I have now done. The artificial limitation on group
2094 length has been removed - we now have only the limit on the total length of
2095 the compiled pattern, which depends on the LINK_SIZE setting.
2096
2097 10. Fixed a bug in the documentation for get/copy named substring when
2098 duplicate names are permitted. If none of the named substrings are set, the
2099 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2100 empty string.
2101
2102 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2103 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2104 because the ] is interpreted as the first data character and the
2105 terminating ] is not found. PCRE has been made compatible with Perl in this
2106 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2107 cause memory overwriting.
2108
2109 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2110 string has been matched (to stop an infinite loop). It was not recognizing
2111 a conditional subpattern that could match an empty string if that
2112 subpattern was within another subpattern. For example, it looped when
2113 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2114 condition was not nested. This bug has been fixed.
2115
2116 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2117 past the start of the subject in the presence of bytes with the top bit
2118 set, for example "\x8aBCD".
2119
2120 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2121 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2122
2123 14. Optimized (?!) to (*FAIL).
2124
2125 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2126 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2127 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2128 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2129 does: it's just the validity check that is more restrictive.
2130
2131 16. Inserted checks for integer overflows during escape sequence (backslash)
2132 processing, and also fixed erroneous offset values for syntax errors during
2133 backslash processing.
2134
2135 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2136 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2137
2138 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2139 caused an overrun.
2140
2141 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2142 something other than just ASCII characters) inside a group that had an
2143 unlimited repeat caused a loop at compile time (while checking to see
2144 whether the group could match an empty string).
2145
2146 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2147 [\P{Any}] did so. (Error in the code for printing property names.)
2148
2149 21. An orphan \E inside a character class could cause a crash.
2150
2151 22. A repeated capturing bracket such as (A)? could cause a wild memory
2152 reference during compilation.
2153
2154 23. There are several functions in pcre_compile() that scan along a compiled
2155 expression for various reasons (e.g. to see if it's fixed length for look
2156 behind). There were bugs in these functions when a repeated \p or \P was
2157 present in the pattern. These operators have additional parameters compared
2158 with \d, etc, and these were not being taken into account when moving along
2159 the compiled data. Specifically:
2160
2161 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2162 length.
2163
2164 (b) An item such as \pL+ within a repeated group could cause crashes or
2165 loops.
2166
2167 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2168 "reference to non-existent subpattern" error.
2169
2170 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2171
2172 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2173 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2174
2175 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2176
2177 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2178 character were causing crashes (broken optimization).
2179
2180 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2181 \p or \P) caused a compile-time loop.
2182
2183 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2184 break. For example, the unstudied pattern [\r\n]A does not match the string
2185 "\r\nA" because change 7.0/46 below moves the current point on by two
2186 characters after failing to match at the start. However, the pattern \nA
2187 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2188 the same is true. There doesn't seem any very clean way out of this, but
2189 what I have chosen to do makes the common cases work: PCRE now takes note
2190 of whether there can be an explicit match for \r or \n anywhere in the
2191 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2192 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2193 pattern has explicit CR or LF references.
2194
2195 29. Added (*CR) etc for changing newline setting at start of pattern.
2196
2197
2198 Version 7.2 19-Jun-07
2199 ---------------------
2200
2201 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2202 which is apparently normally available under Windows.
2203
2204 2. Re-jig the pcregrep tests with different newline settings in an attempt
2205 to make them independent of the local environment's newline setting.
2206
2207 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2208
2209 4. Some of the "internals" tests were previously cut out when the link size
2210 was not 2, because the output contained actual offsets. The recent new
2211 "Z" feature of pcretest means that these can be cut out, making the tests
2212 usable with all link sizes.
2213
2214 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2215 stack recursion. This gives a massive performance boost under BSD, but just
2216 a small improvement under Linux. However, it saves one field in the frame
2217 in all cases.
2218
2219 6. Added more features from the forthcoming Perl 5.10:
2220
2221 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2222 recursion call. It refers to the nth most recently opened parentheses.
2223
2224 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2225 to be opened parentheses.
2226
2227 (c) Conditions that refer to capturing parentheses can be specified
2228 relatively, for example, (?(-2)... or (?(+3)...
2229
2230 (d) \K resets the start of the current match so that everything before
2231 is not part of it.
2232
2233 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2234
2235 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2236 reference syntax.
2237
2238 (g) (?| introduces a group in which the numbering of parentheses in each
2239 alternative starts with the same number.
2240
2241 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2242
2243 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2244 PCRE_INFO_JCHANGED.
2245
2246 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2247 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2248 for detecting groups that can match an empty string.
2249
2250 9. A pattern with a very large number of alternatives (more than several
2251 hundred) was running out of internal workspace during the pre-compile
2252 phase, where pcre_compile() figures out how much memory will be needed. A
2253 bit of new cunning has reduced the workspace needed for groups with
2254 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2255 workspace instead of running out of the 4096 that are available.
2256
2257 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2258
2259 11. Applied patch from Google to remove an optimization that didn't quite work.
2260 The report of the bug said:
2261
2262 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2263 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2264 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2265
2266 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2267 it matched the wrong number of bytes.
2268
2269
2270 Version 7.1 24-Apr-07
2271 ---------------------
2272
2273 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2274 that is more "standard", making use of automake and other Autotools. There
2275 is some re-arrangement of the files and adjustment of comments consequent
2276 on this.
2277
2278 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2279 for recursive directory scanning broke on some systems because the files
2280 are not scanned in any specific order and on different systems the order
2281 was different. A call to "sort" has been inserted into RunGrepTest for the
2282 approprate test as a short-term fix. In the longer term there may be an
2283 alternative.
2284
2285 3. I had an email from Eric Raymond about problems translating some of PCRE's
2286 man pages to HTML (despite the fact that I distribute HTML pages, some
2287 people do their own conversions for various reasons). The problems
2288 concerned the use of low-level troff macros .br and .in. I have therefore
2289 removed all such uses from the man pages (some were redundant, some could
2290 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2291 HTML has been updated to handle .nf/.fi and to complain if it encounters
2292 .br or .in.
2293
2294 4. Updated comments in configure.ac that get placed in config.h.in and also
2295 arranged for config.h to be included in the distribution, with the name
2296 config.h.generic, for the benefit of those who have to compile without
2297 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2298
2299 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2300 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2301 makevp.bat for the current PCRE, using the additional files
2302 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2303
2304 6. A Windows user reported a minor discrepancy with test 2, which turned out
2305 to be caused by a trailing space on an input line that had got lost in his
2306 copy. The trailing space was an accident, so I've just removed it.
2307
2308 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2309 that is needed.
2310
2311 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2312 as "const" (a) because they are and (b) because it helps the PHP
2313 maintainers who have recently made a script to detect big data structures
2314 in the php code that should be moved to the .rodata section. I remembered
2315 to update Builducptable as well, so it won't revert if ucptable.h is ever
2316 re-created.
2317
2318 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2319 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2320 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2321 support is not required. This saves 1.5-2K of code, which is important in
2322 some applications.
2323
2324 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2325 so as not to refer to the tables, even though these functions will never be
2326 called when UTF-8 support is disabled. Otherwise there are problems with a
2327 shared library.
2328
2329 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2330
2331 (a) It was defining its arguments as char * instead of void *.
2332
2333 (b) It was assuming that all moves were upwards in memory; this was true
2334 a long time ago when I wrote it, but is no longer the case.
2335
2336 The emulated memove() is provided for those environments that have neither
2337 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2338 is clearly not the case, as these two bugs were recently reported.
2339
2340 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2341 and Detrail to create the HTML documentation, the .txt form of the man
2342 pages, and it removes trailing spaces from listed files. It also creates
2343 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2344 case, it wraps all the #defines with #ifndefs. This script should be run
2345 before "make dist".
2346
2347 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2348 with Unicode property support.
2349
2350 (a) For a maximizing quantifier, if the two different cases of the
2351 character were of different lengths in their UTF-8 codings (there are
2352 some cases like this - I found 11), and the matching function had to
2353 back up over a mixture of the two cases, it incorrectly assumed they
2354 were both the same length.
2355
2356 (b) When PCRE was configured to use the heap rather than the stack for
2357 recursion during matching, it was not correctly preserving the data for
2358 the other case of a UTF-8 character when checking ahead for a match
2359 while processing a minimizing repeat. If the check also involved
2360 matching a wide character, but failed, corruption could cause an
2361 erroneous result when trying to check for a repeat of the original
2362 character.
2363
2364 13. Some tidying changes to the testing mechanism:
2365
2366 (a) The RunTest script now detects the internal link size and whether there
2367 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2368 values substituted by "configure". (The RunGrepTest script already did
2369 this for UTF-8.) The configure.ac script no longer substitutes the
2370 relevant variables.
2371
2372 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2373 with length and offset values. This means that the output is different
2374 for different internal link sizes. Test 2 is skipped for link sizes
2375 other than 2 because of this, bypassing the problem. Unfortunately,
2376 there was also a test in test 3 (the locale tests) that used /B and
2377 failed for link sizes other than 2. Rather than cut the whole test out,
2378 I have added a new /Z option to pcretest that replaces the length and
2379 offset values with spaces. This is now used to make test 3 independent
2380 of link size. (Test 2 will be tidied up later.)
2381
2382 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2383 segmentation fault instead of returning the appropriate error message.
2384
2385 15. In multiline mode when the newline sequence was set to "any", the pattern
2386 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2387 This doesn't seem right; it now treats the CRLF combination as the line
2388 ending, and so does not match in that case. It's only a pattern such as ^$
2389 that would hit this one: something like ^ABC$ would have failed after \r
2390 and then tried again after \r\n.
2391
2392 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2393 in an attempt to make files that differ only in their line terminators
2394 compare equal. This works on Linux.
2395
2396 17. Under certain error circumstances pcregrep might try to free random memory
2397 as it exited. This is now fixed, thanks to valgrind.
2398
2399 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2400 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2401 was because its rules for how to advance for /g after matching an empty
2402 string at the end of a line did not allow for this case. They now check for
2403 it specially.
2404
2405 20. pcretest is supposed to handle patterns and data of any length, by
2406 extending its buffers when necessary. It was getting this wrong when the
2407 buffer for a data line had to be extended.
2408
2409 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2410 CRLF as a newline sequence.
2411
2412 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2413 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2414 I have nevertheless tidied it up.
2415
2416 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2417
2418 24. Added a man page for pcre-config.
2419
2420
2421 Version 7.0 19-Dec-06
2422 ---------------------
2423
2424 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2425 moving to gcc 4.1.1.
2426
2427 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2428 sys/time.h, which is documented as needed for this function. It doesn't
2429 seem to matter on Linux, but it showed up on some releases of OS X.
2430
2431 3. It seems that there are systems where bytes whose values are greater than
2432 127 match isprint() in the "C" locale. The "C" locale should be the
2433 default when a C program starts up. In most systems, only ASCII printing
2434 characters match isprint(). This difference caused the output from pcretest
2435 to vary, making some of the tests fail. I have changed pcretest so that:
2436
2437 (a) When it is outputting text in the compiled version of a pattern, bytes
2438 other than 32-126 are always shown as hex escapes.
2439
2440 (b) When it is outputting text that is a matched part of a subject string,
2441 it does the same, unless a different locale has been set for the match
2442 (using the /L modifier). In this case, it uses isprint() to decide.
2443
2444 4. Fixed a major bug that caused incorrect computation of the amount of memory
2445 required for a compiled pattern when options that changed within the
2446 pattern affected the logic of the preliminary scan that determines the
2447 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2448 that the computed length was too small. The symptoms of this bug were
2449 either the PCRE error "internal error: code overflow" from pcre_compile(),
2450 or a glibc crash with a message such as "pcretest: free(): invalid next
2451 size (fast)". Examples of patterns that provoked this bug (shown in
2452 pcretest format) are:
2453
2454 /(?-x: )/x
2455 /(?x)(?-x: \s*#\s*)/
2456 /((?i)[\x{c0}])/8
2457 /(?i:[\x{c0}])/8
2458
2459 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2460 is now done differently.
2461
2462 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2463 wrapper classes; (b) implement a new function in the C++ scanner that is
2464 more efficient than the old way of doing things because it avoids levels of
2465 recursion in the regex matching; (c) add a paragraph to the documentation
2466 for the FullMatch() function.
2467
2468 6. The escape sequence \n was being treated as whatever was defined as
2469 "newline". Not only was this contrary to the documentation, which states
2470 that \n is character 10 (hex 0A), but it also went horribly wrong when
2471 "newline" was defined as CRLF. This has been fixed.
2472
2473 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2474 was being set to -1 for the "end of line" case (supposedly a value that no
2475 character can have). Though this value is never used (the check for end of
2476 line is "zero bytes in current character"), it caused compiler complaints.
2477 I've changed it to 0xffffffff.
2478
2479 8. In pcre_version.c, the version string was being built by a sequence of
2480 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2481 string (as it is for production releases) called a macro with an empty
2482 argument. The C standard says the result of this is undefined. The gcc
2483 compiler treats it as an empty string (which was what was wanted) but it is
2484 reported that Visual C gives an error. The source has been hacked around to
2485 avoid this problem.
2486
2487 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2488 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2489 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2490 of them did).
2491
2492 10. Originally, pcretest opened its input and output without "b"; then I was
2493 told that "b" was needed in some environments, so it was added for release
2494 5.0 to both the input and output. (It makes no difference on Unix-like
2495 systems.) Later I was told that it is wrong for the input on Windows. I've
2496 now abstracted the modes into two macros, to make it easier to fiddle with
2497 them, and removed "b" from the input mode under Windows.
2498
2499 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2500
2501 12. Added -help and --help to pcretest as an official way of being reminded
2502 of the options.
2503
2504 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2505 and pcrecpp.cc because they annoy compilers at high warning levels.
2506
2507 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2508
2509 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2510 scripts are not C programs :-) and which was not noticed because it works
2511 on Linux.
2512
2513 16. pcretest is supposed to handle any length of pattern and data line (as one
2514 line or as a continued sequence of lines) by extending its input buffer if
2515 necessary. This feature was broken for very long pattern lines, leading to
2516 a string of junk being passed to pcre_compile() if the pattern was longer
2517 than about 50K.
2518
2519 17. I have done a major re-factoring of the way pcre_compile() computes the
2520 amount of memory needed for a compiled pattern. Previously, there was code
2521 that made a preliminary scan of the pattern in order to do this. That was
2522 OK when PCRE was new, but as the facilities have expanded, it has become
2523 harder and harder to keep it in step with the real compile phase, and there
2524 have been a number of bugs (see for example, 4 above). I have now found a
2525 cunning way of running the real compile function in a "fake" mode that
2526 enables it to compute how much memory it would need, while actually only
2527 ever using a few hundred bytes of working memory and without too many
2528 tests of the mode. This should make future maintenance and development
2529 easier. A side effect of this work is that the limit of 200 on the nesting
2530 depth of parentheses has been removed (though this was never a serious
2531 limitation, I suspect). However, there is a downside: pcre_compile() now
2532 runs more slowly than before (30% or more, depending on the pattern). I
2533 hope this isn't a big issue. There is no effect on runtime performance.
2534
2535 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2536 newline (only possible for the last line of a file) and it was a
2537 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2538
2539 19. Added additional timing features to pcretest. (1) The -tm option now times
2540 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2541 separate command line item, by a number that specifies the number of
2542 repeats to use when timing. The default is 50000; this gives better
2543 precision, but takes uncomfortably long for very large patterns.
2544
2545 20. Extended pcre_study() to be more clever in cases where a branch of a
2546 subpattern has no definite first character. For example, (a*|b*)[cd] would
2547 previously give no result from pcre_study(). Now it recognizes that the
2548 first character must be a, b, c, or d.
2549
2550 21. There was an incorrect error "recursive call could loop indefinitely" if
2551 a subpattern (or the entire pattern) that was being tested for matching an
2552 empty string contained only one non-empty item after a nested subpattern.
2553 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2554 incorrectly, because the \d was being skipped in the check.
2555
2556 22. The pcretest program now has a new pattern option /B and a command line
2557 option -b, which is equivalent to adding /B to every pattern. This causes
2558 it to show the compiled bytecode, without the additional information that
2559 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2560 is the same as /B/I).
2561
2562 23. A new optimization is now able automatically to treat some sequences such
2563 as a*b as a*+b. More specifically, if something simple (such as a character
2564 or a simple class like \d) has an unlimited quantifier, and is followed by
2565 something that cannot possibly match the quantified thing, the quantifier
2566 is automatically "possessified".
2567
2568 24. A recursive reference to a subpattern whose number was greater than 39
2569 went wrong under certain circumstances in UTF-8 mode. This bug could also
2570 have affected the operation of pcre_study().
2571
2572 25. Realized that a little bit of performance could be had by replacing
2573 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2574
2575 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2576
2577 27. Possessive quantifiers such as a++ were previously implemented by turning
2578 them into atomic groups such as ($>a+). Now they have their own opcodes,
2579 which improves performance. This includes the automatically created ones
2580 from 23 above.
2581
2582 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2583 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2584 the first matched character to be a colon. This applied both to named and
2585 numbered groups.
2586
2587 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2588
2589 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2590 building PCRE on Windows easier, so I have included it in the distribution.
2591
2592 31. There is now a check in pcretest against a ridiculously large number being
2593 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2594 loop, the loop is abandoned.
2595
2596 32. Forward references to subpatterns in conditions such as (?(2)...) where
2597 subpattern 2 is defined later cause pcre_compile() to search forwards in
2598 the pattern for the relevant set of parentheses. This search went wrong
2599 when there were unescaped parentheses in a character class, parentheses
2600 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2601
2602 33. "Subroutine" calls and backreferences were previously restricted to
2603 referencing subpatterns earlier in the regex. This restriction has now
2604 been removed.
2605
2606 34. Added a number of extra features that are going to be in Perl 5.10. On the
2607 whole, these are just syntactic alternatives for features that PCRE had
2608 previously implemented using the Python syntax or my own invention. The
2609 other formats are all retained for compatibility.
2610
2611 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2612 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2613 also .NET compatible.
2614
2615 (b) A recursion or subroutine call to a named group can now be defined as
2616 (?&name) as well as (?P>name).
2617
2618 (c) A backreference to a named group can now be defined as \k<name> or
2619 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2620 5.10, are also .NET compatible.
2621
2622 (d) A conditional reference to a named group can now use the syntax
2623 (?(<name>) or (?('name') as well as (?(name).
2624
2625 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2626 groups (named and numbered) that are never evaluated inline, but can be
2627 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2628 is always false. There may be only one alternative in such a group.
2629
2630 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2631 as the simple (?(R). The condition is true only if the most recent
2632 recursion is that of the given number or name. It does not search out
2633 through the entire recursion stack.
2634
2635 (g) The escape \gN or \g{N} has been added, where N is a positive or
2636 negative number, specifying an absolute or relative reference.
2637
2638 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2639 some "unreachable code" warnings.
2640
2641 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2642 things, this adds five new scripts.
2643
2644 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2645 There were also incompatibilities regarding the handling of \Q..\E inside
2646 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2647 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2648
2649 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2650 matches an empty string, and forcibly breaks the loop. There were bugs in
2651 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2652 against aaaa the result was just "a" rather than "aaaa", for example. Two
2653 separate and independent bugs (that affected different cases) have been
2654 fixed.
2655
2656 39. Refactored the code to abolish the use of different opcodes for small
2657 capturing bracket numbers. This is a tidy that I avoided doing when I
2658 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2659 The new approach is not only tidier, it makes it possible to reduce the
2660 memory needed to fix the previous bug (38).
2661
2662 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2663 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2664 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2665 mode.
2666
2667 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2668 report.
2669
2670 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2671 copy construction and assignment in the C++ wrapper.
2672
2673 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2674 couple of bugs that could have given wrong results in the "--newline=crlf"
2675 case.
2676
2677 44. Added a number of casts and did some reorganization of signed/unsigned int
2678 variables following suggestions from Dair Grant. Also renamed the variable
2679 "this" as "item" because it is a C++ keyword.
2680
2681 45. Arranged for dftables to add
2682
2683 #include "pcre_internal.h"
2684
2685 to pcre_chartables.c because without it, gcc 4.x may remove the array
2686 definition from the final binary if PCRE is built into a static library and
2687 dead code stripping is activated.
2688
2689 46. For an unanchored pattern, if a match attempt fails at the start of a
2690 newline sequence, and the newline setting is CRLF or ANY, and the next two
2691 characters are CRLF, advance by two characters instead of one.
2692
2693
2694 Version 6.7 04-Jul-06
2695 ---------------------
2696
2697 1. In order to handle tests when input lines are enormously long, pcretest has
2698 been re-factored so that it automatically extends its buffers when
2699 necessary. The code is crude, but this _is_ just a test program. The
2700 default size has been increased from 32K to 50K.
2701
2702 2. The code in pcre_study() was using the value of the re argument before
2703 testing it for NULL. (Of course, in any sensible call of the function, it
2704 won't be NULL.)
2705
2706 3. The memmove() emulation function in pcre_internal.h, which is used on
2707 systems that lack both memmove() and bcopy() - that is, hardly ever -
2708 was missing a "static" storage class specifier.
2709
2710 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2711 containing an extended class (one that cannot be represented by a bitmap
2712 because it contains high-valued characters or Unicode property items, e.g.
2713 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2714 pattern, but PCRE should not loop if you do not (it no longer does).
2715 [Detail: two cases were found: (a) a repeated subpattern containing an
2716 extended class; (b) a recursive reference to a subpattern that followed a
2717 previous extended class. It wasn't skipping over the extended class
2718 correctly when UTF-8 mode was not set.]
2719
2720 5. A negated single-character class was not being recognized as fixed-length
2721 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2722 compile error "lookbehind assertion is not fixed length".
2723
2724 6. The RunPerlTest auxiliary script was showing an unexpected difference
2725 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2726 write a Perl script that can interpret lines of an input file either as
2727 byte characters or as UTF-8, which is what "perltest" was being required to
2728 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2729 can't do is switch easily at run time between having the "use utf8;" pragma
2730 or not. In the end, I fudged it by using the RunPerlTest script to insert
2731 "use utf8;" explicitly for the UTF-8 tests.
2732
2733 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2734 the end of the subject string, contrary to the documentation and to what
2735 Perl does. This was true of both matching functions. Now it matches only at
2736 the start of the subject and immediately after *internal* newlines.
2737
2738 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2739 a pointer to an int instead of a pointer to an unsigned long int. This
2740 caused problems on 64-bit systems.
2741
2742 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2743 instance of the 'standard' template library not being so standard".
2744
2745 10. There was no check on the number of named subpatterns nor the maximum
2746 length of a subpattern name. The product of these values is used to compute
2747 the size of the memory block for a compiled pattern. By supplying a very
2748 long subpattern name and a large number of named subpatterns, the size
2749 computation could be caused to overflow. This is now prevented by limiting
2750 the length of names to 32 characters, and the number of named subpatterns
2751 to 10,000.
2752
2753 11. Subpatterns that are repeated with specific counts have to be replicated in
2754 the compiled pattern. The size of memory for this was computed from the
2755 length of the subpattern and the repeat count. The latter is limited to
2756 65535, but there was no limit on the former, meaning that integer overflow
2757 could in principle occur. The compiled length of a repeated subpattern is
2758 now limited to 30,000 bytes in order to prevent this.
2759
2760 12. Added the optional facility to have named substrings with the same name.
2761
2762 13. Added the ability to use a named substring as a condition, using the
2763 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2764 are numbers (not recommended). Forward references are permitted.
2765
2766 14. Added forward references in named backreferences (if you see what I mean).
2767
2768 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2769 pattern could run off the end of the subject. For example, the pattern
2770 "(?s)(.{1,5})"8 did this with the subject "ab".
2771
2772 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2773 PCRE_CASELESS was set when matching characters that were quantified with ?
2774 or *.
2775
2776 17. A character class other than a single negated character that had a minimum
2777 but no maximum quantifier - for example [ab]{6,} - was not handled
2778 correctly by pce_dfa_exec(). It would match only one character.
2779
2780 18. A valid (though odd) pattern that looked like a POSIX character
2781 class but used an invalid character after [ (for example [[,abc,]]) caused
2782 pcre_compile() to give the error "Failed: internal error: code overflow" or
2783 in some cases to crash with a glibc free() error. This could even happen if
2784 the pattern terminated after [[ but there just happened to be a sequence of
2785 letters, a binary zero, and a closing ] in the memory that followed.
2786
2787 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2788 over the years. Originally (before any Unicode support), just the bottom 8
2789 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2790 output from "man perlunicode" includes this:
2791
2792 The regular expression compiler produces polymorphic opcodes. That
2793 is, the pattern adapts to the data and automatically switches to
2794 the Unicode character scheme when presented with Unicode data--or
2795 instead uses a traditional byte scheme when presented with byte
2796 data.
2797
2798 Sadly, a wide octal escape does not cause a switch, and in a string with
2799 no other multibyte characters, these octal escapes are treated as before.
2800 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2801 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2802 Unicode string.
2803
2804 I have not perpetrated such confusion in PCRE. Up till now, it took just
2805 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2806 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2807 translate to the appropriate multibyte character.
2808
2809 29. Applied some refactoring to reduce the number of warnings from Microsoft
2810 and Borland compilers. This has included removing the fudge introduced
2811 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2812 a warning about an unused variable.
2813
2814 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2815 characters since release 4.0, because Perl (from release 5.004) does not.
2816 [Or at least, is documented not to: some releases seem to be in conflict
2817 with the documentation.] However, when a pattern was studied with
2818 pcre_study() and all its branches started with \s, PCRE still included VT
2819 as a possible starting character. Of course, this did no harm; it just
2820 caused an unnecessary match attempt.
2821
2822 22. Removed a now-redundant internal flag bit that recorded the fact that case
2823 dependency changed within the pattern. This was once needed for "required
2824 byte" processing, but is no longer used. This recovers a now-scarce options
2825 bit. Also moved the least significant internal flag bit to the most-
2826 significant bit of the word, which was not previously used (hangover from
2827 the days when it was an int rather than a uint) to free up another bit for
2828 the future.
2829
2830 23. Added support for CRLF line endings as well as CR and LF. As well as the
2831 default being selectable at build time, it can now be changed at runtime
2832 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2833 specify that it is scanning data with non-default line endings.
2834
2835 24. Changed the definition of CXXLINK to make it agree with the definition of
2836 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2837
2838 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2839 recursions. This makes a big different to stack usage for some patterns.
2840
2841 26. If a subpattern containing a named recursion or subroutine reference such
2842 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2843 the space required for the compiled pattern went wrong and gave too small a
2844 value. Depending on the environment, this could lead to "Failed: internal
2845 error: code overflow at offset 49" or "glibc detected double free or
2846 corruption" errors.
2847
2848 27. Applied patches from Google (a) to support the new newline modes and (b) to
2849 advance over multibyte UTF-8 characters in GlobalReplace.
2850
2851 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2852 difference for some implementation of PCRE in some Windows version.
2853
2854 29. Added some extra testing facilities to pcretest:
2855
2856 \q<number> in a data line sets the "match limit" value
2857 \Q<number> in a data line sets the "match recursion limt" value
2858 -S <number> sets the stack size, where <number> is in megabytes
2859
2860 The -S option isn't available for Windows.
2861
2862
2863 Version 6.6 06-Feb-06
2864 ---------------------
2865
2866 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2867 in pcreposix.h. I have copied the definition from pcre.h.
2868
2869 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2870 because pcre.h is no longer a built file.
2871
2872 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2873 not normally included in the compiled code.
2874
2875
2876 Version 6.5 01-Feb-06
2877 ---------------------
2878
2879 1. When using the partial match feature with pcre_dfa_exec(), it was not
2880 anchoring the second and subsequent partial matches at the new starting
2881 point. This could lead to incorrect results. For example, with the pattern
2882 /1234/, partially matching against "123" and then "a4" gave a match.
2883
2884 2. Changes to pcregrep:
2885
2886 (a) All non-match returns from pcre_exec() were being treated as failures
2887 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2888 error message is output. Some extra information is given for the
2889 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2890 probably the only errors that are likely to be caused by users (by
2891 specifying a regex that has nested indefinite repeats, for instance).
2892 If there are more than 20 of these errors, pcregrep is abandoned.
2893
2894 (b) A binary zero was treated as data while matching, but terminated the
2895 output line if it was written out. This has been fixed: binary zeroes
2896 are now no different to any other data bytes.
2897
2898 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2899 used to set a locale for matching. The --locale=xxxx long option has
2900 been added (no short equivalent) to specify a locale explicitly on the
2901 pcregrep command, overriding the environment variables.
2902
2903 (d) When -B was used with -n, some line numbers in the output were one less
2904 than they should have been.
2905
2906 (e) Added the -o (--only-matching) option.
2907
2908 (f) If -A or -C was used with -c (count only), some lines of context were
2909 accidentally printed for the final match.
2910
2911 (g) Added the -H (--with-filename) option.
2912
2913 (h) The combination of options -rh failed to suppress file names for files
2914 that were found from directory arguments.
2915
2916 (i) Added the -D (--devices) and -d (--directories) options.
2917
2918 (j) Added the -F (--fixed-strings) option.
2919
2920 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2921
2922 (l) Added the --colo(u)r option.
2923
2924 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2925 is not present by default.
2926
2927 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2928 items such as (?R) or (?1), when the recursion could match a number of
2929 alternatives. If it matched one of the alternatives, but subsequently,
2930 outside the recursion, there was a failure, the code tried to back up into
2931 the recursion. However, because of the way PCRE is implemented, this is not
2932 possible, and the result was an incorrect result from the match.
2933
2934 In order to prevent this happening, the specification of recursion has
2935 been changed so that all such subpatterns are automatically treated as
2936 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2937
2938 4. I had overlooked the fact that, in some locales, there are characters for
2939 which isalpha() is true but neither isupper() nor islower() are true. In
2940 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2941 and ordfeminine) are like this. This affected the treatment of \w and \W
2942 when they appeared in character classes, but not when they appeared outside
2943 a character class. The bit map for "word" characters is now created
2944 separately from the results of isalnum() instead of just taking it from the
2945 upper, lower, and digit maps. (Plus the underscore character, of course.)
2946
2947 5. The above bug also affected the handling of POSIX character classes such as
2948 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2949 permanent tables. Instead, the bit maps for such a class were previously
2950 created as the appropriate unions of the upper, lower, and digit bitmaps.
2951 Now they are created by subtraction from the [[:word:]] class, which has
2952 its own bitmap.
2953
2954 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2955 It is created by subtracting the vertical space characters (\x09, \x0a,
2956 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2957 subtraction was done in the overall bitmap for a character class, meaning
2958 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2959 be recognized. This bug has been fixed.
2960
2961 7. Patches from the folks at Google:
2962
2963 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2964 real life, but is still worth protecting against".
2965
2966 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2967 regular expressions".
2968
2969 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2970 have it.
2971
2972 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2973 "configure" and the latter not, in order to fix a problem somebody had
2974 with compiling the Arg class on HP-UX.
2975
2976 (e) Improve the error-handling of the C++ wrapper a little bit.
2977
2978 (f) New tests for checking recursion limiting.
2979
2980 8. The pcre_memmove() function, which is used only if the environment does not
2981 have a standard memmove() function (and is therefore rarely compiled),
2982 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2983 returning a result (though PCRE never actually uses the result).
2984
2985 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2986 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2987 returned instead of calling malloc() with an overflowing number that would
2988 most likely cause subsequent chaos.
2989
2990 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2991
2992 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2993 with this option is matched, the nmatch and pmatch options of regexec() are
2994 ignored.
2995
2996 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2997 provided in case anyone wants to the the POSIX interface with UTF-8
2998 strings.
2999
3000 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3001 C++ linking (needed for some HP-UX environments).
3002
3003 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3004 (unused parameter) and in the pcre_printint() function (omitted "default"
3005 switch label when the default is to do nothing).
3006
3007 15. Added some code to make it possible, when PCRE is compiled as a C++
3008 library, to replace subject pointers for pcre_exec() with a smart pointer
3009 class, thus making it possible to process discontinuous strings.
3010
3011 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3012 much the same function. They were added by different people who were trying
3013 to make PCRE easy to compile on non-Unix systems. It has been suggested
3014 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3015 for compiling on Windows systems. I have therefore replaced it with
3016 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3017 defaults to "extern" for C or "extern C" for C++, which works fine on
3018 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3019 SCOPE with something explicit in config.h. In addition:
3020
3021 (a) pcreposix.h still had just "extern" instead of either of these macros;
3022 I have replaced it with PCRE_DATA_SCOPE.
3023
3024 (b) Functions such as _pcre_xclass(), which are internal to the library,
3025 but external in the C sense, all had PCRE_EXPORT in their definitions.
3026 This is apparently wrong for the Windows case, so I have removed it.
3027 (It makes no difference on Unix-like systems.)
3028
3029 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3030 of recursive calls to match(). This is different to MATCH_LIMIT because
3031 that limits the total number of calls to match(), not all of which increase
3032 the depth of recursion. Limiting the recursion depth limits the amount of
3033 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3034 when PCRE is compiled, and changed at run time. A patch from Google adds
3035 this functionality to the C++ interface.
3036
3037 18. Changes to the handling of Unicode character properties:
3038
3039 (a) Updated the table to Unicode 4.1.0.
3040
3041 (b) Recognize characters that are not in the table as "Cn" (undefined).
3042
3043 (c) I revised the way the table is implemented to a much improved format
3044 which includes recognition of ranges. It now supports the ranges that
3045 are defined in UnicodeData.txt, and it also amalgamates other
3046 characters into ranges. This has reduced the number of entries in the
3047 table from around 16,000 to around 3,000, thus reducing its size
3048 considerably. I realized I did not need to use a tree structure after
3049 all - a binary chop search is just as efficient. Having reduced the
3050 number of entries, I extended their size from 6 bytes to 8 bytes to
3051 allow for more data.
3052
3053 (d) Added support for Unicode script names via properties such as \p{Han}.
3054
3055 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3056 matching that character.
3057
3058 20. When matching a repeated Unicode property with a minimum greater than zero,
3059 (for example \pL{2,}), PCRE could look past the end of the subject if it
3060 reached it while seeking the minimum number of characters. This could
3061 happen only if some of the characters were more than one byte long, because
3062 there is a check for at least the minimum number of bytes.
3063
3064 21. Refactored the implementation of \p and \P so as to be more general, to
3065 allow for more different types of property in future. This has changed the
3066 compiled form incompatibly. Anybody with saved compiled patterns that use
3067 \p or \P will have to recompile them.
3068
3069 22. Added "Any" and "L&" to the supported property types.
3070
3071 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3072 but give a compile time error if the value is greater than 0xff.
3073
3074 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3075 accidentally not being installed or uninstalled.
3076
3077 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3078 made were to insert the current release number. This seemed silly, because
3079 it made things harder for people building PCRE on systems that don't run
3080 "configure". I have turned pcre.h into a distributed file, no longer built
3081 by "configure", with the version identification directly included. There is
3082 no longer a pcre.h.in file.
3083
3084 However, this change necessitated a change to the pcre-config script as
3085 well. It is built from pcre-config.in, and one of the substitutions was the
3086 release number. I have updated configure.ac so that ./configure now finds
3087 the release number by grepping pcre.h.
3088
3089 26. Added the ability to run the tests under valgrind.
3090
3091
3092 Version 6.4 05-Sep-05
3093 ---------------------
3094
3095 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3096 "--" to be printed when multiple files were scanned, even when none of the
3097 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3098 consider it to be a bug, and have restored the previous behaviour.
3099
3100 2. A couple of code tidies to get rid of compiler warnings.
3101
3102 3. The pcretest program used to cheat by referring to symbols in the library
3103 whose names begin with _pcre_. These are internal symbols that are not
3104 really supposed to be visible externally, and in some environments it is
3105 possible to suppress them. The cheating is now confined to including
3106 certain files from the library's source, which is a bit cleaner.
3107
3108 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3109 file's purpose clearer.
3110
3111 5. Reorganized pcre_ucp_findchar().
3112
3113
3114 Version 6.3 15-Aug-05
3115 ---------------------
3116
3117 1. The file libpcre.pc.in did not have general read permission in the tarball.
3118
3119 2. There were some problems when building without C++ support:
3120
3121 (a) If C++ support was not built, "make install" and "make test" still
3122 tried to test it.
3123
3124 (b) There were problems when the value of CXX was explicitly set. Some
3125 changes have been made to try to fix these, and ...
3126
3127 (c) --disable-cpp can now be used to explicitly disable C++ support.
3128
3129 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3130 backslash in a target when C++ was disabled. This confuses some
3131 versions of "make", apparently. Using an intermediate variable solves
3132 this. (Same for CPP_LOBJ.)
3133
3134 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3135 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3136 necessary on certain architectures.
3137
3138 4. Added a setting of -export-symbols-regex to the link command to remove
3139 those symbols that are exported in the C sense, but actually are local
3140 within the library, and not documented. Their names all begin with
3141 "_pcre_". This is not a perfect job, because (a) we have to except some
3142 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3143 available (and never for static libraries). I have made a note to try to
3144 find a way round (a) in the future.
3145
3146
3147 Version 6.2 01-Aug-05
3148 ---------------------
3149
3150 1. There was no test for integer overflow of quantifier values. A construction
3151 such as {1111111111111111} would give undefined results. What is worse, if
3152 a minimum quantifier for a parenthesized subpattern overflowed and became
3153 negative, the calculation of the memory size went wrong. This could have
3154 led to memory overwriting.
3155
3156 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3157
3158 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3159 operating environments where this matters.
3160
3161 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3162 PCRE options from within the C++ wrapper.
3163
3164 5. Named capturing subpatterns were not being correctly counted when a pattern
3165 was compiled. This caused two problems: (a) If there were more than 100
3166 such subpatterns, the calculation of the memory needed for the whole
3167 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3168 back references of the form \12, where the number was greater than 9, were
3169 not recognized as back references, even though there were sufficient
3170 previous subpatterns.
3171
3172 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3173 versions of gcc, e.g. 2.95.4.
3174
3175
3176 Version 6.1 21-Jun-05
3177 ---------------------
3178
3179 1. There was one reference to the variable "posix" in pcretest.c that was not
3180 surrounded by "#if !defined NOPOSIX".
3181
3182 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3183 the cross-check on the old pcre_info() function, for the benefit of the
3184 cut-down version of PCRE that is currently imported into Exim.
3185
3186 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3187 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3188 patterns that start (?i) but I don't think that matters. The use of (?i) is
3189 just an example; this all applies to the other options as well.
3190
3191 4. Since libtool seems to echo the compile commands it is issuing, the output
3192 from "make" can be reduced a bit by putting "@" in front of each libtool
3193 compile command.
3194
3195 5. Patch from the folks at Google for configure.in to be a bit more thorough
3196 in checking for a suitable C++ installation before trying to compile the
3197 C++ stuff. This should fix a reported problem when a compiler was present,
3198 but no suitable headers.
3199
3200 6. The man pages all had just "PCRE" as their title. I have changed them to
3201 be the relevant file name. I have also arranged that these names are
3202 retained in the file doc/pcre.txt, which is a concatenation in text format
3203 of all the man pages except the little individual ones for each function.
3204
3205 7. The NON-UNIX-USE file had not been updated for the different set of source
3206 files that come with release 6. I also added a few comments about the C++
3207 wrapper.
3208
3209
3210 Version 6.0 07-Jun-05
3211 ---------------------
3212
3213 1. Some minor internal re-organization to help with my DFA experiments.
3214
3215 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3216 didn't matter for the library itself when fully configured, but did matter
3217 when compiling without UCP support, or within Exim, where the ucp files are
3218 not imported.
3219
3220 3. Refactoring of the library code to split up the various functions into
3221 different source modules. The addition of the new DFA matching code (see
3222 below) to a single monolithic source would have made it really too
3223 unwieldy, quite apart from causing all the code to be include in a
3224 statically linked application, when only some functions are used. This is
3225 relevant even without the DFA addition now that patterns can be compiled in
3226 one application and matched in another.
3227
3228 The downside of splitting up is that there have to be some external
3229 functions and data tables that are used internally in different modules of
3230 the library but which are not part of the API. These have all had their
3231 names changed to start with "_pcre_" so that they are unlikely to clash
3232 with other external names.
3233
3234 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3235 a different (DFA) algorithm. Although it is slower than the original
3236 function, it does have some advantages for certain types of matching
3237 problem.
3238
3239 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3240 including restarting after a partial match.
3241
3242 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3243 defined when compiling for Windows was sent to me. I have put it into the
3244 code, though I have no means of testing or verifying it.
3245
3246 7. Added the pcre_refcount() auxiliary function.
3247
3248 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3249 match before or at the first newline in the subject string. In pcretest,
3250 the /f option on a pattern can be used to set this.
3251
3252 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3253 would behave wrongly. This has been present in PCRE since release 4.0.
3254
3255 10. A number of changes to the pcregrep command:
3256
3257 (a) Refactored how -x works; insert ^(...)$ instead of setting
3258 PCRE_ANCHORED and checking the length, in preparation for adding
3259 something similar for -w.
3260
3261 (b) Added the -w (match as a word) option.
3262
3263 (c) Refactored the way lines are read and buffered so as to have more
3264 than one at a time available.
3265
3266 (d) Implemented a pcregrep test script.
3267
3268 (e) Added the -M (multiline match) option. This allows patterns to match
3269 over several lines of the subject. The buffering ensures that at least
3270 8K, or the rest of the document (whichever is the shorter) is available
3271 for matching (and similarly the previous 8K for lookbehind assertions).
3272
3273 (f) Changed the --help output so that it now says
3274
3275 -w, --word-regex(p)
3276
3277 instead of two lines, one with "regex" and the other with "regexp"
3278 because that confused at least one person since the short forms are the
3279 same. (This required a bit of code, as the output is generated
3280 automatically from a table. It wasn't just a text change.)
3281
3282 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3283 option but starts with a hyphen. Could be a pattern or a path name
3284 starting with a hyphen, for instance.
3285
3286 (h) "-" can be given as a file name to represent stdin.
3287
3288 (i) When file names are being printed, "(standard input)" is used for
3289 the standard input, for compatibility with GNU grep. Previously
3290 "<stdin>" was used.
3291
3292 (j) The option --label=xxx can be used to supply a name to be used for
3293 stdin when file names are being printed. There is no short form.
3294
3295 (k) Re-factored the options decoding logic because we are going to add
3296 two more options that take data. Such options can now be given in four
3297 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3298
3299 (l) Added the -A, -B, and -C options for requesting that lines of context
3300 around matches be printed.
3301
3302 (m) Added the -L option to print the names of files that do not contain
3303 any matching lines, that is, the complement of -l.
3304
3305 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3306 continue to scan other files.
3307
3308 (o) The -s option was incorrectly implemented. For compatibility with other
3309 greps, it now suppresses the error message for a non-existent or non-
3310 accessible file (but not the return code). There is a new option called
3311 -q that suppresses the output of matching lines, which was what -s was
3312 previously doing.
3313
3314 (p) Added --include and --exclude options to specify files for inclusion
3315 and exclusion when recursing.
3316
3317 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3318 Hopefully, it now does.
3319
3320 12. Missing cast in pcre_study().
3321
3322 13. Added an "uninstall" target to the makefile.
3323
3324 14. Replaced "extern" in the function prototypes in Makefile.in with
3325 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3326 world, but is set differently for Windows.
3327
3328 15. Added a second compiling function called pcre_compile2(). The only
3329 difference is that it has an extra argument, which is a pointer to an
3330 integer error code. When there is a compile-time failure, this is set
3331 non-zero, in addition to the error test pointer being set to point to an
3332 error message. The new argument may be NULL if no error number is required
3333 (but then you may as well call pcre_compile(), which is now just a
3334 wrapper). This facility is provided because some applications need a
3335 numeric error indication, but it has also enabled me to tidy up the way
3336 compile-time errors are handled in the POSIX wrapper.
3337
3338 16. Added VPATH=.libs to the makefile; this should help when building with one
3339 prefix path and installing with another. (Or so I'm told by someone who
3340 knows more about this stuff than I do.)
3341
3342 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3343 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3344 match everything, including newlines. This is not POSIX-compatible, but
3345 somebody wanted the feature. From pcretest it can be activated by using
3346 both the P and the s flags.
3347
3348 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3349
3350 19. libpcre.pc was being incorrectly installed as executable.
3351
3352 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3353 it now also looks for '\r' so that it will work unmodified on Windows.
3354
3355 21. Added Google's contributed C++ wrapper to the distribution.
3356
3357 22. Added some untidy missing memory free() calls in pcretest, to keep
3358 Electric Fence happy when testing.
3359
3360
3361
3362 Version 5.0 13-Sep-04
3363 ---------------------
3364
3365 1. Internal change: literal characters are no longer packed up into items
3366 containing multiple characters in a single byte-string. Each character
3367 is now matched using a separate opcode. However, there may be more than one
3368 byte in the character in UTF-8 mode.
3369
3370 2. The pcre_callout_block structure has two new fields: pattern_position and
3371 next_item_length. These contain the offset in the pattern to the next match
3372 item, and its length, respectively.
3373
3374 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3375 insertion of callouts before each pattern item. Added the /C option to
3376 pcretest to make use of this.
3377
3378 4. On the advice of a Windows user, the lines
3379
3380 #if defined(_WIN32) || defined(WIN32)
3381 _setmode( _fileno( stdout ), 0x8000 );
3382 #endif /* defined(_WIN32) || defined(WIN32) */
3383
3384 have been added to the source of pcretest. This apparently does useful
3385 magic in relation to line terminators.
3386
3387 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3388 for the benefit of those environments where the "b" makes a difference.
3389
3390 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3391 to know about it. I have put a hack into configure.in that adds in code
3392 to set GCC=yes if CC=icc. This seems to end up at a point in the
3393 generated configure script that is early enough to affect the setting of
3394 compiler options, which is what is needed, but I have no means of testing
3395 whether it really works. (The user who reported this had patched the
3396 generated configure script, which of course I cannot do.)
3397
3398 LATER: After change 22 below (new libtool files), the configure script
3399 seems to know about icc (and also ecc). Therefore, I have commented out
3400 this hack in configure.in.
3401
3402 7. Added support for pkg-config (2 patches were sent in).
3403
3404 8. Negated POSIX character classes that used a combination of internal tables
3405 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3406 [[:^ascii]]. Typically, they would match almost any characters. The other
3407 POSIX classes were not broken in this way.
3408
3409 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3410 to find the match, as PCRE was deluded into thinking that the match had to
3411 start at the start point or following a newline. The same bug applied to
3412 patterns with negative forward assertions or any backward assertions
3413 preceding ".*" at the start, unless the pattern required a fixed first
3414 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3415
3416 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3417 starting at the last subject character, bytes beyond the end of the subject
3418 string were read.
3419
3420 11. Renamed the variable "class" as "classbits" to make life easier for C++
3421 users. (Previously there was a macro definition, but it apparently wasn't
3422 enough.)
3423
3424 12. Added the new field "tables" to the extra data so that tables can be passed
3425 in at exec time, or the internal tables can be re-selected. This allows
3426 a compiled regex to be saved and re-used at a later time by a different
3427 program that might have everything at different addresses.
3428
3429 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3430 -R library as well as a -L library.
3431
3432 14. The debugging options of pcretest (-d on the command line or D on a
3433 pattern) showed incorrect output for anything following an extended class
3434 that contained multibyte characters and which was followed by a quantifier.
3435
3436 15. Added optional support for general category Unicode character properties
3437 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3438 support. It adds about 90K to the size of the library. The meanings of the
3439 inbuilt class escapes such as \d and \s have NOT been changed.
3440
3441 16. Updated pcredemo.c to include calls to free() to release the memory for the
3442 compiled pattern.
3443
3444 17. The generated file chartables.c was being created in the source directory
3445 instead of in the building directory. This caused the build to fail if the
3446 source directory was different from the building directory, and was
3447 read-only.
3448
3449 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3450 file. No doubt somebody will tell me if they don't make sense... Also added
3451 Dan Mooney's comments about building on OpenVMS.
3452
3453 19. Added support for partial matching via the PCRE_PARTIAL option for
3454 pcre_exec() and the \P data escape in pcretest.
3455
3456 20. Extended pcretest with 3 new pattern features:
3457
3458 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3459 write the compiled pattern to the file whose name is "rest-of-line".
3460 This is a straight binary dump of the data, with the saved pointer to
3461 the character tables forced to be NULL. The study data, if any, is
3462 written too. After writing, pcretest reads a new pattern.
3463
3464 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3465 compiled pattern from the given file. There must not be any
3466 occurrences of "<" in the file name (pretty unlikely); if there are,
3467 pcretest will instead treat the initial "<" as a pattern delimiter.
3468 After reading in the pattern, pcretest goes on to read data lines as
3469 usual.
3470
3471 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3472 and 16-bit fields in a compiled pattern, to simulate a pattern that
3473 was compiled on a host of opposite endianness.
3474
3475 21. The pcre-exec() function can now cope with patterns that were compiled on
3476 hosts of opposite endianness, with this restriction:
3477
3478 As for any compiled expression that is saved and used later, the tables
3479 pointer field cannot be preserved; the extra_data field in the arguments
3480 to pcre_exec() should be used to pass in a tables address if a value
3481 other than the default internal tables were used at compile time.
3482
3483 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3484 now diagnosed as an error. Previously, most of the time, a negative number
3485 would have been treated as zero, but if in addition "ovector" was passed as
3486 NULL, a crash could occur.
3487
3488 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3489 new versions from the libtool 1.5 distribution (the last one is a copy of
3490 a file called libtool.m4). This seems to have fixed the need to patch
3491 "configure" to support Darwin 1.3 (which I used to do). However, I still
3492 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3493 workstation).
3494
3495 24. Changed the PCRE licence to be the more standard "BSD" licence.
3496
3497
3498 Version 4.5 01-Dec-03
3499 ---------------------
3500
3501 1. There has been some re-arrangement of the code for the match() function so
3502 that it can be compiled in a version that does not call itself recursively.
3503 Instead, it keeps those local variables that need separate instances for
3504 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3505 needs to "recurse". Keeping track of where control must go is done by means
3506 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3507 hide most of the details from the main code, and operates only if
3508 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3509 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3510 operating.
3511
3512 To make it easier for callers to provide specially tailored get/free
3513 functions for this usage, two new functions, pcre_stack_malloc, and
3514 pcre_stack_free, are used. They are always called in strict stacking order,
3515 and the size of block requested is always the same.
3516
3517 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3518 PCRE has been compiled to use the stack or the heap for recursion. The
3519 -C option of pcretest uses this to show which version is compiled.
3520
3521 A new data escape \S, is added to pcretest; it causes the amounts of store
3522 obtained and freed by both kinds of malloc/free at match time to be added
3523 to the output.
3524
3525 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3526 what's available on my current Linux desktop machine.
3527
3528 3. When matching a UTF-8 string, the test for a valid string at the start has
3529 been extended. If start_offset is not zero, PCRE now checks that it points
3530 to a byte that is the start of a UTF-8 character. If not, it returns
3531 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3532 this is necessary because there may be backward assertions in the pattern.
3533 When matching the same subject several times, it may save resources to use
3534 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3535
3536 4. The code for checking the validity of UTF-8 strings has been tightened so
3537 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3538 containing "overlong sequences".
3539
3540 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3541 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3542 should have been "&", but it just so happened that all the cases this let
3543 through by mistake were picked up later in the function.
3544
3545 6. I had used a variable called "isblank" - this is a C99 function, causing
3546 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3547
3548 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3549 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3550 so the version is shown; (c) stop "make test" echoing "./RunTest".
3551
3552 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3553
3554 9. The prototype for memmove() for systems that don't have it was using
3555 size_t, but the inclusion of the header that defines size_t was later. I've
3556 moved the #includes for the C headers earlier to avoid this.
3557
3558 10. Added some adjustments to the code to make it easier to compiler on certain
3559 special systems:
3560
3561 (a) Some "const" qualifiers were missing.
3562 (b) Added the macro EXPORT before all exported functions; by default this
3563 is defined to be empty.
3564 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3565 that it reads its output file name as an argument instead of writing
3566 to the standard output and assuming this can be redirected.
3567
3568 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3569 class containing characters with values greater than 255, PCRE compilation
3570 went into a loop.
3571
3572 12. A recursive reference to a subpattern that was within another subpattern
3573 that had a minimum quantifier of zero caused PCRE to crash. For example,
3574 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3575 recursion. If the recursively-called subpattern itself had a zero repeat,
3576 that was OK.
3577
3578 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3579 buffer into which it was copied (for escape processing) was still set at
3580 1024, so long lines caused crashes.
3581
3582 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3583 "internal error: code overflow...". This applied to any character class
3584 that was followed by a possessive quantifier.
3585
3586 15. Modified the Makefile to add libpcre.la as a prerequisite for
3587 libpcreposix.la because I was told this is needed for a parallel build to
3588 work.
3589
3590 16. If a pattern that contained .* following optional items at the start was
3591 studied, the wrong optimizing data was generated, leading to matching
3592 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3593 matching string must start with a or b or c. The correct conclusion for
3594 this pattern is that a match can start with any character.
3595
3596
3597 Version 4.4 13-Aug-03
3598 ---------------------
3599
3600 1. In UTF-8 mode, a character class containing characters with values between
3601 127 and 255 was not handled correctly if the compiled pattern was studied.
3602 In fixing this, I have also improved the studying algorithm for such
3603 classes (slightly).
3604
3605 2. Three internal functions had redundant arguments passed to them. Removal
3606 might give a very teeny performance improvement.
3607
3608 3. Documentation bug: the value of the capture_top field in a callout is *one
3609 more than* the number of the hightest numbered captured substring.
3610
3611 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3612 in incorrectly linking with a previously installed version. They now link
3613 explicitly with libpcre.la.
3614
3615 5. configure.in no longer needs to recognize Cygwin specially.
3616
3617 6. A problem in pcre.in for Windows platforms is fixed.
3618
3619 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3620 pcretest, it used to include the size of the study block as part of its
3621 output. Unfortunately, the structure contains a field that has a different
3622 size on different hardware architectures. This meant that the tests that
3623 showed this size failed. As the block is currently always of a fixed size,
3624 this information isn't actually particularly useful in pcretest output, so
3625 I have just removed it.
3626
3627 8. Three pre-processor statements accidentally did not start in column 1.
3628 Sadly, there are *still* compilers around that complain, even though
3629 standard C has not required this for well over a decade. Sigh.
3630
3631 9. In pcretest, the code for checking callouts passed small integers in the
3632 callout_data field, which is a void * field. However, some picky compilers
3633 complained about the casts involved for this on 64-bit systems. Now
3634 pcretest passes the address of the small integer instead, which should get
3635 rid of the warnings.
3636
3637 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3638 both compile and run time, and gives an error if an invalid UTF-8 sequence
3639 is found. There is a option for disabling this check in cases where the
3640 string is known to be correct and/or the maximum performance is wanted.
3641
3642 11. In response to a bug report, I changed one line in Makefile.in from
3643
3644 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3645 to
3646 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3647
3648 to look similar to other lines, but I have no way of telling whether this
3649 is the right thing to do, as I do not use Windows. No doubt I'll get told
3650 if it's wrong...
3651
3652
3653 Version 4.3 21-May-03
3654 ---------------------
3655
3656 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3657 Makefile.
3658
3659 2. Some refactoring to improve the quality of the code:
3660
3661 (i) The utf8_table... variables are now declared "const".
3662
3663 (ii) The code for \cx, which used the "case flipping" table to upper case
3664 lower case letters, now just substracts 32. This is ASCII-specific,
3665 but the whole concept of \cx is ASCII-specific, so it seems
3666 reasonable.
3667
3668 (iii) PCRE was using its character types table to recognize decimal and
3669 hexadecimal digits in the pattern. This is silly, because it handles
3670 only 0-9, a-f, and A-F, but the character types table is locale-
3671 specific, which means strange things might happen. A private
3672 table is now used for this - though it costs 256 bytes, a table is
3673 much faster than multiple explicit tests. Of course, the standard
3674 character types table is still used for matching digits in subject
3675 strings against \d.
3676
3677 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3678 ending in _t are). So I've renamed it as ESC_tee.
3679
3680 3. The first argument for regexec() in the POSIX wrapper should have been
3681 defined as "const".
3682
3683 4. Changed pcretest to use malloc() for its buffers so that they can be
3684 Electric Fenced for debugging.
3685
3686 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3687 to read one or more bytes before the start of the subject string. Often this
3688 had no effect on PCRE's behaviour, but in some circumstances it could
3689 provoke a segmentation fault.
3690
3691 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3692 to try to read one or more bytes before the start of the subject string.
3693
3694 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3695 UTF-8 support could misbehave in various ways if the subject string
3696 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3697 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3698 back over UTF-8 characters.)
3699
3700
3701 Version 4.2 14-Apr-03
3702 ---------------------
3703
3704 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3705
3706 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3707 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3708 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3709 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3710 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3711 and BUILD_EXEEXT
3712 Note: automatic setting of the BUILD variables is not yet working
3713 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3714 compile-time but not at link-time
3715 [LINK]: use for linking executables only
3716 make different versions for Windows and non-Windows
3717 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3718 libraries
3719 [LINK_FOR_BUILD]: new variable
3720 [OBJEXT]: use throughout
3721 [EXEEXT]: use throughout
3722 <winshared>: new target
3723 <wininstall>: new target
3724 <dftables.o>: use native compiler
3725 <dftables>: use native linker
3726 <install>: handle Windows platform correctly
3727 <clean>: ditto
3728 <check>: ditto
3729 copy DLL to top builddir before testing
3730
3731 As part of these changes, -no-undefined was removed again. This was reported
3732 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3733 in any case.
3734
3735 3. Some tidies to get rid of compiler warnings:
3736
3737 . In the match_data structure, match_limit was an unsigned long int, whereas
3738 match_call_count was an int. I've made them both unsigned long ints.
3739
3740 . In pcretest the fact that a const uschar * doesn't automatically cast to
3741 a void * provoked a warning.
3742
3743 . Turning on some more compiler warnings threw up some "shadow" variables
3744 and a few more missing casts.
3745
3746 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3747 option, a class that contained a single character with a value between 128
3748 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3749
3750 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3751 option, a class that contained several characters, but with at least one
3752 whose value was between 128 and 255 caused PCRE to crash.
3753
3754
3755 Version 4.1 12-Mar-03
3756 ---------------------
3757
3758 1. Compiling with gcc -pedantic found a couple of places where casts were
3759 needed, and a string in dftables.c that was longer than standard compilers are
3760 required to support.
3761
3762 2. Compiling with Sun's compiler found a few more places where the code could
3763 be tidied up in order to avoid warnings.
3764
3765 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3766 first of these names is deprecated in the latest Autoconf in favour of the name
3767 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3768 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3769 analogy I have changed it to CFLAGS_FOR_BUILD.
3770
3771 4. Added -no-undefined to the linking command in the Makefile, because this is
3772 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3773 linking step for the pcreposix library.
3774
3775 5. PCRE was failing to diagnose the case of two named groups with the same
3776 name.
3777
3778 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3779 literal character that is needed in the subject for a match, and scans along to
3780 ensure that it is present before embarking on the full matching process. This
3781 saves time in cases of nested unlimited repeats that are never going to match.
3782 Problem: the scan can take a lot of time if the subject is very long (e.g.
3783 megabytes), thus penalizing straightforward matches. It is now done only if the
3784 amount of subject to be scanned is less than 1000 bytes.
3785
3786 7. A lesser problem with the same optimization is that it was recording the
3787 first character of an anchored pattern as "needed", thus provoking a search
3788 right along the subject, even when the first match of the pattern was going to
3789 fail. The "needed" character is now not set for anchored patterns, unless it
3790 follows something in the pattern that is of non-fixed length. Thus, it still
3791 fulfils its original purpose of finding quick non-matches in cases of nested
3792 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3793
3794
3795 Version 4.0 17-Feb-03
3796 ---------------------
3797
3798 1. If a comment in an extended regex that started immediately after a meta-item
3799 extended to the end of string, PCRE compiled incorrect data. This could lead to
3800 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3801
3802 2. Moved to autoconf 2.53 and libtool 1.4.2.
3803
3804 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3805 the special perltest8 script is no longer needed - all the tests can be run
3806 from a single perltest script.
3807
3808 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3809 by \s. It has now been removed in PCRE. This means it isn't recognized as
3810 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3811 class [:space:] *does* include VT, thereby creating a mess.
3812
3813 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3814 space and tab.
3815
3816 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3817 its new features into the main test script, reducing the number of scripts.
3818
3819 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3820 were backward compatible, and made the (?i) apply to the whole pattern, as if
3821 /i were given. Now it behaves more logically, and applies the option setting
3822 only to what follows. PCRE has been changed to follow suit. However, if it
3823 finds options settings right at the start of the pattern, it extracts them into
3824 the global options, as before. Thus, they show up in the info data.
3825
3826 8. Added support for the \Q...\E escape sequence. Characters in between are
3827 treated as literals. This is slightly different from Perl in that $ and @ are
3828 also handled as literals inside the quotes. In Perl, they will cause variable
3829 interpolation. Note the following examples:
3830
3831 Pattern PCRE matches Perl matches
3832
3833 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3834 \Qabc\$xyz\E abc\$xyz abc\$xyz
3835 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3836
3837 For compatibility with Perl, \Q...\E sequences are recognized inside character
3838 classes as well as outside them.
3839
3840 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3841 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3842 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3843 signed/unsigned warnings.
3844
3845 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3846 option for pcretest, so I've replaced it by a simple function that does just
3847 that job.
3848
3849 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3850 "pcregrep -".
3851
3852 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3853 Java package. This provides some syntactic sugar for simple cases of what my
3854 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3855 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3856 item, you can use this simplified notation. Note that only makes sense with
3857 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3858 greediness, whatever the setting of the PCRE_UNGREEDY option.
3859
3860 13. A change of greediness default within a pattern was not taking effect at
3861 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3862 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3863 was abstracted outside.
3864
3865 14. PCRE now supports the \G assertion. It is true when the current matching
3866 position is at the start point of the match. This differs from \A when the
3867 starting offset is non-zero. Used with the /g option of pcretest (or similar
3868 code), it works in the same way as it does for Perl's /g option. If all
3869 alternatives of a regex begin with \G, the expression is anchored to the start
3870 match position, and the "anchored" flag is set in the compiled expression.
3871
3872 15. Some bugs concerning the handling of certain option changes within patterns
3873 have been fixed. These applied to options other than (?ims). For example,
3874 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3875 been the other way round. Some of this was related to change 7 above.
3876
3877 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3878 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3879 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3880 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3881
3882 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3883 mode. Unlike ".", it always matches newline, whatever the setting of
3884 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3885 assertions. Perl allows it, but it doesn't (in general) work because it can't
3886 calculate the length of the lookbehind. At least, that's the case for Perl
3887 5.8.0 - I've been told they are going to document that it doesn't work in
3888 future.
3889
3890 18. Added an error diagnosis for escapes that PCRE does not support: these are
3891 \L, \l, \N, \P, \p, \U, \u, and \X.
3892
3893 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3894 reading past the end of the pattern in cases such as /[abcd/.
3895
3896 20. PCRE was getting more memory than necessary for patterns with classes that
3897 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3898
3899 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3900 compiling PCRE for use with Virtual Pascal.
3901
3902 22. Small fix to the Makefile to make it work properly if the build is done
3903 outside the source tree.
3904
3905 23. Added a new extension: a condition to go with recursion. If a conditional
3906 subpattern starts with (?(R) the "true" branch is used if recursion has
3907 happened, whereas the "false" branch is used only at the top level.
3908
3909 24. When there was a very long string of literal characters (over 255 bytes
3910 without UTF support, over 250 bytes with UTF support), the computation of how
3911 much memory was required could be incorrect, leading to segfaults or other
3912 strange effects.
3913
3914 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3915 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3916 there was a subsequent back reference to those brackets. This meant that, for
3917 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3918 possible to check for precisely this case. All we can do is abandon the
3919 optimization if .* occurs inside capturing brackets when there are any back
3920 references whatsoever. (See below for a better fix that came later.)
3921
3922 26. The handling of the optimization for finding the first character of a
3923 non-anchored pattern, and for finding a character that is required later in the
3924 match were failing in some cases. This didn't break the matching; it just
3925 failed to optimize when it could. The way this is done has been re-implemented.
3926
3927 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3928
3929 28. Added a new feature that provides some of the functionality that Perl
3930 provides with (?{...}). The facility is termed a "callout". The way it is done
3931 in PCRE is for the caller to provide an optional function, by setting
3932 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3933 global variable. By default it is unset, which disables all calling out. To get
3934 the function called, the regex must include (?C) at appropriate points. This
3935 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3936 This provides a means of identifying different callout points. When PCRE
3937 reaches such a point in the regex, if pcre_callout has been set, the external
3938 function is called. It is provided with data in a structure called
3939 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3940 matching continues; if it returns a non-zero value, the match at the current
3941 point fails. However, backtracking will occur if possible. [This was changed
3942 later and other features added - see item 49 below.]
3943
3944 29. pcretest is upgraded to test the callout functionality. It provides a
3945 callout function that displays information. By default, it shows the start of
3946 the match and the current position in the text. There are some new data escapes
3947 to vary what happens:
3948
3949 \C+ in addition, show current contents of captured substrings
3950 \C- do not supply a callout function
3951 \C!n return 1 when callout number n is reached
3952 \C!n!m return 1 when callout number n is reached for the mth time
3953
3954 30. If pcregrep was called with the -l option and just a single file name, it
3955 output "<stdin>" if a match was found, instead of the file name.
3956
3957 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3958 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3959 pcre_exec(). This saves a malloc/free per call. The default value of
3960 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3961 when configuring.
3962
3963 32. The default maximum size of a compiled pattern is 64K. There have been a
3964 few cases of people hitting this limit. The code now uses macros to handle the
3965 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3966 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3967 configuring. Tests 2 and 5 work only with 2-byte links because they output
3968 debugging information about compiled patterns.
3969
3970 33. Internal code re-arrangements:
3971
3972 (a) Moved the debugging function for printing out a compiled regex into
3973 its own source file (printint.c) and used #include to pull it into
3974 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3975 separate copies.
3976
3977 (b) Defined the list of op-code names for debugging as a macro in
3978 internal.h so that it is next to the definition of the opcodes.
3979
3980 (c) Defined a table of op-code lengths for simpler skipping along compiled
3981 code. This is again a macro in internal.h so that it is next to the
3982 definition of the opcodes.
3983
3984 34. Added support for recursive calls to individual subpatterns, along the
3985 lines of Robin Houston's patch (but implemented somewhat differently).
3986
3987 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3988 allow it to read and process whole directories in Win32. This code was
3989 contributed by Lionel Fourquaux; it has not been tested by me.
3990
3991 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3992 used to name a group. Names consist of alphanumerics and underscores, and must
3993 be unique. Back references use the syntax (?P=name) and recursive calls use
3994 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3995 numbers. The function pcre_fullinfo() can be used after compilation to extract
3996 a name/number map. There are three relevant calls:
3997
3998 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3999 PCRE_INFO_NAMECOUNT yields the number of entries
4000 PCRE_INFO_NAMETABLE yields a pointer to the map.
4001
4002 The map is a vector of fixed-size entries. The size of each entry depends on
4003 the length of the longest name used. The first two bytes of each entry are the
4004 group number, most significant byte first. There follows the corresponding
4005 name, zero terminated. The names are in alphabetical order.
4006
4007 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4008 case instead of 255. Making it the same both with and without UTF-8 support
4009 means that the same test output works with both.
4010
4011 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4012 calling malloc() with a zero argument.
4013
4014 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4015 optimization. I've improved things by keeping a bitmap of backreferences with
4016 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4017 fact referenced, the optimization can be applied. It is unlikely that a
4018 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4019 the match to follow \n) will appear inside brackets with a number greater than
4020 31, but if it does, any back reference > 31 suppresses the optimization.
4021
4022 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4023 of disabling numbered capturing parentheses. Any opening parenthesis that is
4024 not followed by ? behaves as if it were followed by ?: but named parentheses
4025 can still be used for capturing (and they will acquire numbers in the usual
4026 way).
4027
4028 41. Redesigned the return codes from the match() function into yes/no/error so
4029 that errors can be passed back from deep inside the nested calls. A malloc
4030 failure while inside a recursive subpattern call now causes the
4031 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4032
4033 42. It is now possible to set a limit on the number of times the match()
4034 function is called in a call to pcre_exec(). This facility makes it possible to
4035 limit the amount of recursion and backtracking, though not in a directly
4036 obvious way, because the match() function is used in a number of different
4037 circumstances. The count starts from zero for each position in the subject
4038 string (for non-anchored patterns). The default limit is, for compatibility, a
4039 large number, namely 10 000 000. You can change this in two ways:
4040
4041 (a) When configuring PCRE before making, you can use --with-match-limit=n
4042 to set a default value for the compiled library.
4043
4044 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4045 a different value is set. See 45 below.
4046
4047 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4048
4049 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4050 of things that can be changed at compile time. The first argument specifies
4051 what is wanted and the second points to where the information is to be placed.
4052 The current list of available information is:
4053
4054 PCRE_CONFIG_UTF8
4055
4056 The output is an integer that is set to one if UTF-8 support is available;
4057 otherwise it is set to zero.
4058
4059 PCRE_CONFIG_NEWLINE
4060
4061 The output is an integer that it set to the value of the code that is used for
4062 newline. It is either LF (10) or CR (13).
4063
4064 PCRE_CONFIG_LINK_SIZE
4065
4066 The output is an integer that contains the number of bytes used for internal
4067 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4068
4069 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4070
4071 The output is an integer that contains the threshold above which the POSIX
4072 interface uses malloc() for output vectors. See item 31 above.
4073
4074 PCRE_CONFIG_MATCH_LIMIT
4075
4076 The output is an unsigned integer that contains the default limit of the number
4077 of match() calls in a pcre_exec() execution. See 42 above.
4078
4079 44. pcretest has been upgraded by the addition of the -C option. This causes it
4080 to extract all the available output from the new pcre_config() function, and to
4081 output it. The program then exits immediately.
4082
4083 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4084 order to support additional features. One way would have been to define
4085 pcre_exec2() (for example) with extra arguments, but this would not have been
4086 extensible, and would also have required all calls to the original function to
4087 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4088 is used for passing in "extra" data from pcre_study().
4089
4090 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4091 contains the following fields:
4092
4093 flags a bitmap indicating which of the following fields are set
4094 study_data opaque data from pcre_study()
4095 match_limit a way of specifying a limit on match() calls for a specific
4096 call to pcre_exec()
4097 callout_data data for callouts (see 49 below)
4098
4099 The flag bits are also defined in pcre.h, and are
4100
4101 PCRE_EXTRA_STUDY_DATA
4102 PCRE_EXTRA_MATCH_LIMIT
4103 PCRE_EXTRA_CALLOUT_DATA
4104
4105 The pcre_study() function now returns one of these new pcre_extra blocks, with
4106 the actual study data pointed to by the study_data field, and the
4107 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4108 before. That is, this change is entirely upwards-compatible and requires no
4109 change to existing code.
4110
4111 If you want to pass in additional data to pcre_exec(), you can either place it
4112 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4113 block.
4114
4115 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4116 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4117 times with different match limits, until it finds the minimum value needed for
4118 pcre_exec() to complete. The value is then output. This can be instructive; for
4119 most simple matches the number is quite small, but for pathological cases it
4120 gets very large very quickly.
4121
4122 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4123 returns the size of the data block pointed to by the study_data field in a
4124 pcre_extra block, that is, the value that was passed as the argument to
4125 pcre_malloc() when PCRE was getting memory in which to place the information
4126 created by pcre_study(). The fourth argument should point to a size_t variable.
4127 pcretest has been extended so that this information is shown after a successful
4128 pcre_study() call when information about the compiled regex is being displayed.
4129
4130 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4131 because what follows is always an absolute path. (Later: it turns out that this
4132 is more than cosmetic for MinGW, because it doesn't like empty path
4133 components.)
4134
4135 49. Some changes have been made to the callout feature (see 28 above):
4136
4137 (i) A callout function now has three choices for what it returns:
4138
4139 0 => success, carry on matching
4140 > 0 => failure at this point, but backtrack if possible
4141 < 0 => serious error, return this value from pcre_exec()
4142
4143 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4144 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4145 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4146 use by callout functions. It will never be used by PCRE itself.
4147
4148 (ii) The pcre_extra structure (see 45 above) has a void * field called
4149 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4150 pcre_callout_block structure has a field of the same name. The contents of
4151 the field passed in the pcre_extra structure are passed to the callout
4152 function in the corresponding field in the callout block. This makes it
4153 easier to use the same callout-containing regex from multiple threads. For
4154 testing, the pcretest program has a new data escape
4155
4156 \C*n pass the number n (may be negative) as callout_data
4157
4158 If the callout function in pcretest receives a non-zero value as
4159 callout_data, it returns that value.
4160
4161 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4162 there were some redundant $(CFLAGS) in commands that are now specified as
4163 $(LINK), which already includes $(CFLAGS).
4164
4165 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4166 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4167 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4168 one-byte characters throughout. Note that case-insensitive matching applies
4169 only to characters whose values are less than 256. PCRE doesn't support the
4170 notion of cases for higher-valued characters.
4171
4172 (i) A character class whose characters are all within 0-255 is handled as
4173 a bit map, and the map is inverted for negative classes. Previously, a
4174 character > 255 always failed to match such a class; however it should
4175 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4176
4177 (ii) A negated character class with a single character < 255 is coded as
4178 "not this character" (OP_NOT). This wasn't working properly when the test
4179 character was multibyte, either singly or repeated.
4180
4181 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4182 mode, for example: \x{100}{2,3}.
4183
4184 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4185 singly or repeated) now correctly test multibyte characters. However,
4186 PCRE doesn't recognize any characters with values greater than 255 as
4187 digits, spaces, or word characters. Such characters always match \D, \S,
4188 and \W, and never match \d, \s, or \w.
4189
4190 (v) Classes may now contain characters and character ranges with values
4191 greater than 255. For example: [ab\x{100}-\x{400}].
4192
4193 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4194 PCRE in UTF-8 mode.
4195
4196 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4197 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4198 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4199 value.)
4200
4201 53. The single man page has become too large. I have therefore split it up into
4202 a number of separate man pages. These also give rise to individual HTML pages;
4203 these are now put in a separate directory, and there is an index.html page that
4204 lists them all. Some hyperlinking between the pages has been installed.
4205
4206 54. Added convenience functions for handling named capturing parentheses.
4207
4208 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4209 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4210 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4211 are faulted.
4212
4213 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4214 calling configure. These values are used when compiling the dftables.c program
4215 which is run to generate the source of the default character tables. They
4216 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4217 you will need to set these values.
4218
4219 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4220
4221
4222 Version 3.9 02-Jan-02
4223 ---------------------
4224
4225 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4226
4227 2. If --disable-static was given, the building process failed when trying to
4228 build pcretest and pcregrep. (For some reason it was using libtool to compile
4229 them, which is not right, as they aren't part of the library.)
4230
4231
4232 Version 3.8 18-Dec-01
4233 ---------------------
4234
4235 1. The experimental UTF-8 code was completely screwed up. It was packing the
4236 bytes in the wrong order. How dumb can you get?
4237
4238
4239 Version 3.7 29-Oct-01
4240 ---------------------
4241
4242 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4243 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4244 this didn't happen under Solaris 8, where I normally test things.
4245
4246 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4247 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4248 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4249 this fix an hour or so after the initial 3.7 release.)
4250
4251
4252 Version 3.6 23-Oct-01
4253 ---------------------
4254
4255 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4256 offsets passed as NULL with zero offset count.
4257
4258 2. The config.guess and config.sub files had not been updated when I moved to
4259 the latest autoconf.
4260
4261
4262 Version 3.5 15-Aug-01
4263 ---------------------
4264
4265 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4266 had been forgotten.
4267
4268 2. By using declared but undefined structures, we can avoid using "void"
4269 definitions in pcre.h while keeping the internal definitions of the structures
4270 private.
4271
4272 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4273 user point of view, this means that both static and shared libraries are built
4274 by default, but this can be individually controlled. More of the work of
4275 handling this static/shared cases is now inside libtool instead of PCRE's make
4276 file.
4277
4278 4. The pcretest utility is now installed along with pcregrep because it is
4279 useful for users (to test regexs) and by doing this, it automatically gets
4280 relinked by libtool. The documentation has been turned into a man page, so
4281 there are now .1, .txt, and .html versions in /doc.
4282
4283 5. Upgrades to pcregrep:
4284 (i) Added long-form option names like gnu grep.
4285 (ii) Added --help to list all options with an explanatory phrase.
4286 (iii) Added -r, --recursive to recurse into sub-directories.
4287 (iv) Added -f, --file to read patterns from a file.
4288
4289 6. pcre_exec() was referring to its "code" argument before testing that
4290 argument for NULL (and giving an error if it was NULL).
4291
4292 7. Upgraded Makefile.in to allow for compiling in a different directory from
4293 the source directory.
4294
4295 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4296 options bits, the pointer it was passed was to an int instead of to an unsigned
4297 long int. This mattered only on 64-bit systems.
4298
4299 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4300 generated) instead of pcre.in, which it its source. Also made the same change
4301 in several of the .c files.
4302
4303 10. A new release of gcc defines printf() as a macro, which broke pcretest
4304 because it had an ifdef in the middle of a string argument for printf(). Fixed
4305 by using separate calls to printf().
4306
4307 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4308 script, to force use of CR or LF instead of \n in the source. On non-Unix
4309 systems, the value can be set in config.h.
4310
4311 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4312 absolute limit. Changed the text of the error message to make this clear, and
4313 likewise updated the man page.
4314
4315 13. The limit of 99 on the number of capturing subpatterns has been removed.
4316 The new limit is 65535, which I hope will not be a "real" limit.
4317
4318
4319 Version 3.4 22-Aug-00
4320 ---------------------
4321
4322 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4323
4324 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4325
4326
4327 Version 3.3 01-Aug-00
4328 ---------------------
4329
4330 1. If an octal character was given, but the value was greater than \377, it
4331 was not getting masked to the least significant bits, as documented. This could
4332 lead to crashes in some systems.
4333
4334 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4335 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4336
4337 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4338 These just pass their arguments on to (pcre_free)(), but they are provided
4339 because some uses of PCRE bind it to non-C systems that can call its functions,
4340 but cannot call free() or pcre_free() directly.
4341
4342 4. Add "make test" as a synonym for "make check". Corrected some comments in
4343 the Makefile.
4344
4345 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4346 Makefile.
4347
4348 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4349 command called pgrep for grepping around the active processes.
4350
4351 7. Added the beginnings of support for UTF-8 character strings.
4352
4353 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4354 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4355 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4356 out for the ar command.)
4357
4358
4359 Version 3.2 12-May-00
4360 ---------------------
4361
4362 This is purely a bug fixing release.
4363
4364 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4365 of ZA. This was just one example of several cases that could provoke this bug,
4366 which was introduced by change 9 of version 2.00. The code for breaking
4367 infinite loops after an iteration that matches an empty string was't working
4368 correctly.
4369
4370 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4371 when matched against abbab (for example). After matching an empty string, it
4372 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4373 caused it to match further down the string than it should.
4374
4375 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4376 was there because it doesn't seem to be needed, and it causes trouble on some
4377 systems, as it is not a Standard C header. It has been removed.
4378
4379 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4380 were reported on the Macintosh. The changes were from
4381
4382 while ((c = *(++ptr)) != 0 && c != '\n');
4383 to
4384 while ((c = *(++ptr)) != 0 && c != '\n') ;
4385
4386 Totally extraordinary, but if that's what it takes...
4387
4388 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4389 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4390 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4391 assumes the way PCRE uses memmove() (always moving upwards).
4392
4393 6. PCRE is being used in one environment where strchr() is not available. There
4394 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4395 faster code anyway.
4396
4397
4398 Version 3.1 09-Feb-00
4399 ---------------------
4400
4401 The only change in this release is the fixing of some bugs in Makefile.in for
4402 the "install" target:
4403
4404 (1) It was failing to install pcreposix.h.
4405
4406 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4407
4408
4409 Version 3.0 01-Feb-00
4410 ---------------------
4411
4412 1. Add support for the /+ modifier to perltest (to output $` like it does in
4413 pcretest).
4414
4415 2. Add support for the /g modifier to perltest.
4416
4417 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4418 matches null strings.
4419
4420 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4421 pattern. Perl treats empty patterns specially - it reuses the most recent
4422 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4423 effect.
4424
4425 5. The POSIX interface was broken in that it was just handing over the POSIX
4426 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4427 required a bigger vector, with some working space on the end. This means that
4428 the POSIX wrapper now has to get and free some memory, and copy the results.
4429
4430 6. Added some simple autoconf support, placing the test data and the
4431 documentation in separate directories, re-organizing some of the
4432 information files, and making it build pcre-config (a GNU standard). Also added
4433 libtool support for building PCRE as a shared library, which is now the
4434 default.
4435
4436 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4437 09 are not valid octal constants. Single digits will be used for minor values
4438 less than 10.
4439
4440 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4441 existing programs that set these in the POSIX interface can use PCRE without
4442 modification.
4443
4444 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4445 return all that pcre_info() returns, plus additional data. The pcre_info()
4446 function is retained for compatibility, but is considered to be obsolete.
4447
4448 10. Added experimental recursion feature (?R) to handle one common case that
4449 Perl 5.6 will be able to do with (?p{...}).
4450
4451 11. Added support for POSIX character classes like [:alpha:], which Perl is
4452 adopting.
4453
4454
4455 Version 2.08 31-Aug-99
4456 ----------------------
4457
4458 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4459 trying to match at the startoffset position, but instead was moving forward to
4460 the next newline as if a previous match had failed.
4461
4462 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4463 and could get into a loop if a null string was matched other than at the start
4464 of the subject.
4465
4466 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4467 be distinguished at compile time, and for completeness also added PCRE_DATE.
4468
4469 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4470 in GnuWin32 environments.
4471
4472
4473 Version 2.07 29-Jul-99
4474 ----------------------
4475
4476 1. The documentation is now supplied in plain text form and HTML as well as in
4477 the form of man page sources.
4478
4479 2. C++ compilers don't like assigning (void *) values to other pointer types.
4480 In particular this affects malloc(). Although there is no problem in Standard
4481 C, I've put in casts to keep C++ compilers happy.
4482
4483 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4484 should be (const char *).
4485
4486 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4487 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4488 However, I haven't made this a standard facility. The documentation doesn't
4489 mention it, and the Makefile doesn't support it.
4490
4491 5. The Makefile now contains an "install" target, with editable destinations at
4492 the top of the file. The pcretest program is not installed.
4493
4494 6. pgrep -V now gives the PCRE version number and date.
4495
4496 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4497 causing the entire string to be ignored, instead of just the last character.
4498
4499 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4500 non-matching string, it can take a very, very long time, even for strings of
4501 quite modest length, because of the nested recursion. PCRE now does better in
4502 some of these cases. It does this by remembering the last required literal
4503 character in the pattern, and pre-searching the subject to ensure it is present
4504 before running the real match. In other words, it applies a heuristic to detect
4505 some types of certain failure quickly, and in the above example, if presented
4506 with a string that has no trailing " it gives "no match" very quickly.
4507
4508 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4509 other alternatives are tried instead.
4510
4511
4512 Version 2.06 09-Jun-99
4513 ----------------------
4514
4515 1. Change pcretest's output for amount of store used to show just the code
4516 space, because the remainder (the data block) varies in size between 32-bit and
4517 64-bit systems.
4518
4519 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4520 start matching at. This allows lookbehinds to work when searching for multiple
4521 occurrences in a string.
4522
4523 3. Added additional options to pcretest for testing multiple occurrences:
4524
4525 /+ outputs the rest of the string that follows a match
4526 /g loops for multiple occurrences, using the new startoffset argument
4527 /G loops for multiple occurrences by passing an incremented pointer
4528
4529 4. PCRE wasn't doing the "first character" optimization for patterns starting
4530 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4531 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4532 the letter 'x'. On long subject strings, this gives a significant speed-up.
4533
4534
4535 Version 2.05 21-Apr-99
4536 ----------------------
4537
4538 1. Changed the type of magic_number from int to long int so that it works
4539 properly on 16-bit systems.
4540
4541 2. Fixed a bug which caused patterns starting with .* not to work correctly
4542 when the subject string contained newline characters. PCRE was assuming
4543 anchoring for such patterns in all cases, which is not correct because .* will
4544 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4545 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4546 must be retried after every newline in the subject.
4547
4548
4549 Version 2.04 18-Feb-99
4550 ----------------------
4551
4552 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4553 computation of the store needed to hold the pattern was incorrect (too large).
4554 If such patterns were nested a few deep, this could multiply and become a real
4555 problem.
4556
4557 2. Added /M option to pcretest to show the memory requirement of a specific
4558 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4559
4560 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4561 compiled in such a way that the backtracking after subsequent failure was
4562 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4563 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4564
4565
4566 Version 2.03 02-Feb-99
4567 ----------------------
4568
4569 1. Fixed typo and small mistake in man page.
4570
4571 2. Added 4th condition (GPL supersedes if conflict) and created separate
4572 LICENCE file containing the conditions.
4573
4574 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4575 Perl, that is the internal \ allows the delimiter to be included in the
4576 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4577 the final delimiter, add \ to the end of the pattern (to test the error).
4578
4579 4. Added the convenience functions for extracting substrings after a successful
4580 match. Updated pcretest to make it able to test these functions.
4581
4582
4583 Version 2.02 14-Jan-99
4584 ----------------------
4585
4586 1. Initialized the working variables associated with each extraction so that
4587 their saving and restoring doesn't refer to uninitialized store.
4588
4589 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4590 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4591 fix the problem.
4592
4593 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4594 calls, and wasn't printing the correct value for compiling calls. Increased the
4595 default value of LOOPREPEAT, and the number of significant figures in the
4596 times.
4597
4598 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4599
4600 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4601 a building problem on Windows NT with a FAT file system.
4602
4603
4604 Version 2.01 21-Oct-98
4605 ----------------------
4606
4607 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4608 to character tables built by pcre_maketables() in the current locale. If NULL
4609 is passed, the default tables are used.
4610
4611
4612 Version 2.00 24-Sep-98
4613 ----------------------
4614
4615 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4616 it any more.
4617
4618 2. Allow quantification of (?>) groups, and make it work correctly.
4619
4620 3. The first character computation wasn't working for (?>) groups.
4621
4622 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4623 end of the subject) and add 5.005's \z, which really does match only at the
4624 very end of the subject.
4625
4626 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4627
4628 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4629 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4630 localized options. All options to pcre_study() were also removed.
4631
4632 7. Add other new features from 5.005:
4633
4634 $(?<= positive lookbehind
4635 $(?<! negative lookbehind
4636 (?imsx-imsx) added the unsetting capability
4637 such a setting is global if at outer level; local otherwise
4638 (?imsx-imsx:) non-capturing groups with option setting
4639 (?(cond)re|re) conditional pattern matching
4640
4641 A backreference to itself in a repeated group matches the previous
4642 captured string.
4643
4644 8. General tidying up of studying (both automatic and via "study")
4645 consequential on the addition of new assertions.
4646
4647 9. As in 5.005, unlimited repeated groups that could match an empty substring
4648 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4649 runtime if any iteration does actually match an empty substring.
4650
4651 10. Include the RunTest script in the distribution.
4652
4653 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4654 discrepancies, some of which were old and were also with respect to 5.004. They
4655 have now been fixed.
4656
4657
4658 Version 1.09 28-Apr-98
4659 ----------------------
4660
4661 1. A negated single character class followed by a quantifier with a minimum
4662 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4663 program crashes, or just wrong answers. This did not apply to negated classes
4664 containing more than one character, or to minima other than one.
4665
4666
4667 Version 1.08 27-Mar-98
4668 ----------------------
4669
4670 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4671
4672 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4673 latter must appear before anything that relies on it in the pattern.
4674
4675
4676 Version 1.07 16-Feb-98
4677 ----------------------
4678
4679 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4680 repeat of a potentially empty string).
4681
4682
4683 Version 1.06 23-Jan-98
4684 ----------------------
4685
4686 1. Added Markus Oberhumer's little patches for C++.
4687
4688 2. Literal strings longer than 255 characters were broken.
4689
4690
4691 Version 1.05 23-Dec-97
4692 ----------------------
4693
4694 1. Negated character classes containing more than one character were failing if
4695 PCRE_CASELESS was set at run time.
4696
4697
4698 Version 1.04 19-Dec-97
4699 ----------------------
4700
4701 1. Corrected the man page, where some "const" qualifiers had been omitted.
4702
4703 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4704 input syntax.
4705
4706 3. Fixed memory leak which occurred when a regex with back references was
4707 matched with an offsets vector that wasn't big enough. The temporary memory
4708 that is used in this case wasn't being freed if the match failed.
4709
4710 4. Tidied pcretest to ensure it frees memory that it gets.
4711
4712 5. Temporary memory was being obtained in the case where the passed offsets
4713 vector was exactly big enough.
4714
4715 6. Corrected definition of offsetof() from change 5 below.
4716
4717 7. I had screwed up change 6 below and broken the rules for the use of
4718 setjmp(). Now fixed.
4719
4720
4721 Version 1.03 18-Dec-97
4722 ----------------------
4723
4724 1. A erroneous regex with a missing opening parenthesis was correctly
4725 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4726 on some systems.
4727
4728 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4729 it was reported that one broken compiler failed on the former because "code" is
4730 also an independent variable.
4731
4732 3. The erroneous regex a[]b caused an array overrun reference.
4733
4734 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4735 fail on data ending with that character. (It was going on too far, and checking
4736 the next character, typically a binary zero.) This was specific to the
4737 optimized code for single-character negative classes.
4738
4739 5. Added a contributed patch from the TIN world which does the following:
4740
4741 + Add an undef for memmove, in case the the system defines a macro for it.
4742
4743 + Add a definition of offsetof(), in case there isn't one. (I don't know
4744 the reason behind this - offsetof() is part of the ANSI standard - but
4745 it does no harm).
4746
4747 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4748 most of the places where whitespace preceded '#'. I have given up and
4749 allowed the remaining 2 cases to be at the margin.
4750
4751 + Rename some variables in pcre to eliminate shadowing. This seems very
4752 pedantic, but does no harm, of course.
4753
4754 6. Moved the call to setjmp() into its own function, to get rid of warnings
4755 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4756
4757 7. Constructs such as \d{8,} were compiling into the equivalent of
4758 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4759 outcome, but in this particular case used more store than had been allocated,
4760 which caused the bug to be discovered because it threw up an internal error.
4761
4762 8. The debugging code in both pcre and pcretest for outputting the compiled
4763 form of a regex was going wrong in the case of back references followed by
4764 curly-bracketed repeats.
4765
4766
4767 Version 1.02 12-Dec-97
4768 ----------------------
4769
4770 1. Typos in pcre.3 and comments in the source fixed.
4771
4772 2. Applied a contributed patch to get rid of places where it used to remove
4773 'const' from variables, and fixed some signed/unsigned and uninitialized
4774 variable warnings.
4775
4776 3. Added the "runtest" target to Makefile.
4777
4778 4. Set default compiler flag to -O2 rather than just -O.
4779
4780
4781 Version 1.01 19-Nov-97
4782 ----------------------
4783
4784 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4785 like /([ab]*)*/, that is, for classes with more than one character in them.
4786
4787 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4788 as /((?>a*))*/ (a PCRE_EXTRA facility).
4789
4790
4791 Version 1.00 18-Nov-97
4792 ----------------------
4793
4794 1. Added compile-time macros to support systems such as SunOS4 which don't have
4795 memmove() or strerror() but have other things that can be used instead.
4796
4797 2. Arranged that "make clean" removes the executables.
4798
4799
4800 Version 0.99 27-Oct-97
4801 ----------------------
4802
4803 1. Fixed bug in code for optimizing classes with only one character. It was
4804 initializing a 32-byte map regardless, which could cause it to run off the end
4805 of the memory it had got.
4806
4807 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4808
4809
4810 Version 0.98 22-Oct-97
4811 ----------------------
4812
4813 1. Fixed bug in code for handling temporary memory usage when there are more
4814 back references than supplied space in the ovector. This could cause segfaults.
4815
4816
4817 Version 0.97 21-Oct-97
4818 ----------------------
4819
4820 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4821
4822 2. Optimized negated single characters not to use a bit map.
4823
4824 3. Brought error texts together as macro definitions; clarified some of them;
4825 fixed one that was wrong - it said "range out of order" when it meant "invalid
4826 escape sequence".
4827
4828 4. Changed some char * arguments to const char *.
4829
4830 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4831
4832 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4833 pcretest.
4834
4835
4836 Version 0.96 16-Oct-97
4837 ----------------------
4838
4839 1. Added a simple "pgrep" utility to the distribution.
4840
4841 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4842 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4843 where "ddd" means "one or more decimal digits".
4844
4845 3. Fixed serious bug. If a pattern had a back reference, but the call to
4846 pcre_exec() didn't supply a large enough ovector to record the related
4847 identifying subpattern, the match always failed. PCRE now remembers the number
4848 of the largest back reference, and gets some temporary memory in which to save
4849 the offsets during matching if necessary, in order to ensure that
4850 backreferences always work.
4851
4852 4. Increased the compatibility with Perl in a number of ways:
4853
4854 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4855 to request this handling. The option can be set at compile or exec time.
4856
4857 (b) $ matches before a terminating newline by default; an option
4858 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4859 mode). The option can be set at compile or exec time.
4860
4861 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4862 the same as Perl's. If the decimal number it represents is less than 10
4863 or there aren't that many previous left capturing parentheses, an octal
4864 escape is read. Inside a character class, it's always an octal escape,
4865 even if it is a single digit.
4866
4867 (d) An escaped but undefined alphabetic character is taken as a literal,
4868 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4869 escapes.
4870
4871 (e) {0} is now permitted. (The previous item is removed from the compiled
4872 pattern).
4873
4874 5. Changed all the names of code files so that the basic parts are no longer
4875 than 10 characters, and abolished the teeny "globals.c" file.
4876
4877 6. Changed the handling of character classes; they are now done with a 32-byte
4878 bit map always.
4879
4880 7. Added the -d and /D options to pcretest to make it possible to look at the
4881 internals of compilation without having to recompile pcre.
4882
4883
4884 Version 0.95 23-Sep-97
4885 ----------------------
4886
4887 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4888 \x20 at the start of a run of normal characters. These were being treated as
4889 real characters, instead of the source characters being re-checked.
4890
4891
4892 Version 0.94 18-Sep-97
4893 ----------------------
4894
4895 1. The functions are now thread-safe, with the caveat that the global variables
4896 containing pointers to malloc() and free() or alternative functions are the
4897 same for all threads.
4898
4899 2. Get pcre_study() to generate a bitmap of initial characters for non-
4900 anchored patterns when this is possible, and use it if passed to pcre_exec().
4901
4902
4903 Version 0.93 15-Sep-97
4904 ----------------------
4905
4906 1. /(b)|(:+)/ was computing an incorrect first character.
4907
4908 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4909 but not actually doing anything yet.
4910
4911 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4912 as Perl does (e.g. [-az] or [az-]).
4913
4914 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4915 all possible positions.
4916
4917 5. Split up into different modules to avoid including unneeded functions in a
4918 compiled binary. However, compile and exec are still in one module. The "study"
4919 function is split off.
4920
4921 6. The character tables are now in a separate module whose source is generated
4922 by an auxiliary program - but can then be edited by hand if required. There are
4923 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4924 toupper() in the code.
4925
4926 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4927 make them global. Abolish the function for setting them, as the caller can now
4928 set them directly.
4929
4930
4931 Version 0.92 11-Sep-97
4932 ----------------------
4933
4934 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4935 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4936
4937 2. Caseless matching was not working in character classes if the characters in
4938 the pattern were in upper case.
4939
4940 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4941
4942 4. Make PCRE_ANCHORED public and accept as a compile option.
4943
4944 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4945 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4946 pass them.
4947
4948 6. Give an error if bad option bits passed at compile or run time.
4949
4950 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4951 pcretest to cause it to pass that flag.
4952
4953 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4954 options, and the first character, if set.
4955
4956 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4957
4958
4959 Version 0.91 10-Sep-97
4960 ----------------------
4961
4962 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4963 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4964
4965 2. PCRE was looping on encountering an indefinitely repeated back reference to
4966 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4967 Perl does - treats the match as successful.
4968
4969 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5