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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1025 - (show annotations)
Wed Aug 29 11:24:23 2012 UTC (7 years, 3 months ago) by ph10
File size: 235155 byte(s)
Apply user patch to allow PCRE_NO_UTF8_CHECK in pcrecpp.cc.
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
69
70 Version 8.31 06-July-2012
71 -------------------------
72
73 1. Fixing a wrong JIT test case and some compiler warnings.
74
75 2. Removed a bashism from the RunTest script.
76
77 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
78 to unsigned type, result still unsigned" that was given by an MS compiler
79 on encountering the code "-sizeof(xxx)".
80
81 4. Partial matching support is added to the JIT compiler.
82
83 5. Fixed several bugs concerned with partial matching of items that consist
84 of more than one character:
85
86 (a) /^(..)\1/ did not partially match "aba" because checking references was
87 done on an "all or nothing" basis. This also applied to repeated
88 references.
89
90 (b) \R did not give a hard partial match if \r was found at the end of the
91 subject.
92
93 (c) \X did not give a hard partial match after matching one or more
94 characters at the end of the subject.
95
96 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
97 a partial match for the string "\r".
98
99 (e) When newline was set to CRLF, the metacharacter "." did not recognize
100 a partial match for a CR character at the end of the subject string.
101
102 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
103 running pcretest, the text "(JIT)" added to the output whenever JIT is
104 actually used to run the match.
105
106 7. Individual JIT compile options can be set in pcretest by following -s+[+]
107 or /S+[+] with a digit between 1 and 7.
108
109 8. OP_NOT now supports any UTF character not just single-byte ones.
110
111 9. (*MARK) control verb is now supported by the JIT compiler.
112
113 10. The command "./RunTest list" lists the available tests without actually
114 running any of them. (Because I keep forgetting what they all are.)
115
116 11. Add PCRE_INFO_MAXLOOKBEHIND.
117
118 12. Applied a (slightly modified) user-supplied patch that improves performance
119 when the heap is used for recursion (compiled with --disable-stack-for-
120 recursion). Instead of malloc and free for each heap frame each time a
121 logical recursion happens, frames are retained on a chain and re-used where
122 possible. This sometimes gives as much as 30% improvement.
123
124 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
125 call.
126
127 14. As documented, (*COMMIT) is now confined to within a positive assertion.
128
129 15. It is now possible to link pcretest with libedit as an alternative to
130 libreadline.
131
132 16. (*COMMIT) control verb is now supported by the JIT compiler.
133
134 17. The Unicode data tables have been updated to Unicode 6.1.0.
135
136 18. Added --file-list option to pcregrep.
137
138 19. Added binary file support to pcregrep, including the -a, --binary-files,
139 -I, and --text options.
140
141 20. The madvise function is renamed for posix_madvise for QNX compatibility
142 reasons. Fixed by Giuseppe D'Angelo.
143
144 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
145
146 22. Changed the option for creating symbolic links for 16-bit man pages from
147 -s to -sf so that re-installing does not cause issues.
148
149 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
150
151 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
152 matches in certain environments (the workspace was not being correctly
153 retained). Also added to pcre_dfa_exec() a simple plausibility check on
154 some of the workspace data at the beginning of a restart.
155
156 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
157 was not doing so when it should - probably a typo introduced by SVN 528
158 (change 8.10/14).
159
160 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
161 \w+ when the character tables indicated that \x{c4} was a word character.
162 There were several related cases, all because the tests for doing a table
163 lookup were testing for characters less than 127 instead of 255.
164
165 27. If a pattern contains capturing parentheses that are not used in a match,
166 their slots in the ovector are set to -1. For those that are higher than
167 any matched groups, this happens at the end of processing. In the case when
168 there were back references that the ovector was too small to contain
169 (causing temporary malloc'd memory to be used during matching), and the
170 highest capturing number was not used, memory off the end of the ovector
171 was incorrectly being set to -1. (It was using the size of the temporary
172 memory instead of the true size.)
173
174 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
175 ovector size, it uses memory at the end of the block that it has got.
176
177 29. Check for an overlong MARK name and give an error at compile time. The
178 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
179
180 30. JIT compiler update.
181
182 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
183 Rill for the patch.
184
185 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
186
187 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
188
189 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
190 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
191 was enabled.
192
193 35. Improve JIT code generation for greedy plus quantifier.
194
195 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
196 1 to "aa" instead of to an empty string. The bug affected repeated groups
197 that could potentially match an empty string.
198
199 37. Optimizing single character iterators in JIT.
200
201 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
202 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
203 codepoints that are too big for the mode are faulted, and in a UTF mode,
204 disallowed codepoints are also faulted.
205
206 39. If PCRE was compiled with UTF support, in three places in the DFA
207 matcher there was code that should only have been obeyed in UTF mode, but
208 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
209 processing when bytes with values greater than 127 were present. In 16-bit
210 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
211 both cases the values are those that cannot be the first data item in a UTF
212 character. The three items that might have provoked this were recursions,
213 possessively repeated groups, and atomic groups.
214
215 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
216 and pcregrep, because some OS require shared objects to be explicitly
217 passed to ld, causing the link step to fail if they are not.
218
219 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
220 mode, patterns that started with \h* or \R* might be incorrectly matched.
221
222
223 Version 8.30 04-February-2012
224 -----------------------------
225
226 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
227 name is defined in ctype.h.
228
229 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
230 only in quite long subpatterns.
231
232 3. Removed the function pcre_info(), which has been obsolete and deprecated
233 since it was replaced by pcre_fullinfo() in February 2000.
234
235 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
236 match a (*MARK), and the match failed at the start of the subject, a
237 reference to memory before the start of the subject could occur. This bug
238 was introduced by fix 17 of release 8.21.
239
240 5. A reference to an unset group with zero minimum repetition was giving
241 totally wrong answers (in non-JavaScript-compatibility mode). For example,
242 /(another)?(\1?)test/ matched against "hello world test". This bug was
243 introduced in release 8.13.
244
245 6. Add support for 16-bit character strings (a large amount of work involving
246 many changes and refactorings).
247
248 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
249 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
250 from a file.
251
252 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
253 rounding is not applied in this particular case).
254
255 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
256 if they appear, or are escaped, in patterns.
257
258 10. Get rid of a number of -Wunused-but-set-variable warnings.
259
260 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
261 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
262 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
263 also returns the mark "x". This bug applied to capturing parentheses,
264 non-capturing parentheses, and atomic parentheses. It also applied to some
265 assertions.
266
267 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
268 information out of configure.ac instead of relying on pcre.h.generic, which
269 is not stored in the repository.
270
271 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
272 -lreadline.
273
274 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
275
276 15. Applied Graycode's patch to put the top-level frame on the stack rather
277 than the heap when not using the stack for recursion. This gives a
278 performance improvement in many cases when recursion is not deep.
279
280 16. Experimental code added to "pcretest -C" to output the stack frame size.
281
282
283 Version 8.21 12-Dec-2011
284 ------------------------
285
286 1. Updating the JIT compiler.
287
288 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
289 are added as well.
290
291 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
292 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
293 calling _pcre_jit_exec. Some extra comments are added.
294
295 4. (*MARK) settings inside atomic groups that do not contain any capturing
296 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
297 was introduced by change 18 for 8.20.
298
299 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
300 ECMA-262 standard.
301
302 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
303 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
304 This bug was probably introduced by change 9 of 8.13.
305
306 7. While fixing 6 above, I noticed that a number of other items were being
307 incorrectly rejected as "not fixed length". This arose partly because newer
308 opcodes had not been added to the fixed-length checking code. I have (a)
309 corrected the bug and added tests for these items, and (b) arranged for an
310 error to occur if an unknown opcode is encountered while checking for fixed
311 length instead of just assuming "not fixed length". The items that were
312 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
313 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
314 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
315
316 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
317 being incorrectly compiled and would have given unpredicatble results.
318
319 9. A possessively repeated subpattern with minimum repeat count greater than
320 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
321 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
322 the first (A) could occur when it should not.
323
324 10. Add a cast and remove a redundant test from the code.
325
326 11. JIT should use pcre_malloc/pcre_free for allocation.
327
328 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
329 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
330 is anything other than /usr, -L is still shown).
331
332 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
333
334 14. Perl does not support \N without a following name in a [] class; PCRE now
335 also gives an error.
336
337 15. If a forward reference was repeated with an upper limit of around 2000,
338 it caused the error "internal error: overran compiling workspace". The
339 maximum number of forward references (including repeats) was limited by the
340 internal workspace, and dependent on the LINK_SIZE. The code has been
341 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
342 the default depends on LINK_SIZE. There is a new upper limit (for safety)
343 of around 200,000 forward references. While doing this, I also speeded up
344 the filling in of repeated forward references.
345
346 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
347 incorrectly expecting the subject to contain another "a" after the start.
348
349 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
350 in the match, the SKIP should be ignored. This was not happening; instead
351 the SKIP was being treated as NOMATCH. For patterns such as
352 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
353 tested.
354
355 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
356 now much more compatible with Perl, in particular in cases where the result
357 is a non-match for a non-anchored pattern. For example, if
358 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
359 "m", where previously it did not return a name. A side effect of this
360 change is that for partial matches, the last encountered mark name is
361 returned, as for non matches. A number of tests that were previously not
362 Perl-compatible have been moved into the Perl-compatible test files. The
363 refactoring has had the pleasing side effect of removing one argument from
364 the match() function, thus reducing its stack requirements.
365
366 19. If the /S+ option was used in pcretest to study a pattern using JIT,
367 subsequent uses of /S (without +) incorrectly behaved like /S+.
368
369 21. Retrieve executable code size support for the JIT compiler and fixing
370 some warnings.
371
372 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
373 not work when the shorter character appeared right at the end of the
374 subject string.
375
376 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
377 systems.
378
379 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
380 output it when the /M option is used in pcretest.
381
382 25. The CheckMan script was not being included in the distribution. Also, added
383 an explicit "perl" to run Perl scripts from the PrepareRelease script
384 because this is reportedly needed in Windows.
385
386 26. If study data was being save in a file and studying had not found a set of
387 "starts with" bytes for the pattern, the data written to the file (though
388 never used) was taken from uninitialized memory and so caused valgrind to
389 complain.
390
391 27. Updated RunTest.bat as provided by Sheri Pierce.
392
393 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
394
395 29. Computation of memory usage for the table of capturing group names was
396 giving an unnecessarily large value.
397
398
399 Version 8.20 21-Oct-2011
400 ------------------------
401
402 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
403 a POSIX class. After further experiments with Perl, which convinced me that
404 Perl has bugs and confusions, a closing square bracket is no longer allowed
405 in a POSIX name. This bug also affected patterns with classes that started
406 with full stops.
407
408 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
409 captured substring, but while checking the failing first alternative,
410 substring 1 is temporarily captured. If the output vector supplied to
411 pcre_exec() was not big enough for this capture, the yield of the function
412 was still zero ("insufficient space for captured substrings"). This cannot
413 be totally fixed without adding another stack variable, which seems a lot
414 of expense for a edge case. However, I have improved the situation in cases
415 such as /(a)(b)x|abc/ matched against "abc", where the return code
416 indicates that fewer than the maximum number of slots in the ovector have
417 been set.
418
419 3. Related to (2) above: when there are more back references in a pattern than
420 slots in the output vector, pcre_exec() uses temporary memory during
421 matching, and copies in the captures as far as possible afterwards. It was
422 using the entire output vector, but this conflicts with the specification
423 that only 2/3 is used for passing back captured substrings. Now it uses
424 only the first 2/3, for compatibility. This is, of course, another edge
425 case.
426
427 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
428 main code base, and can be used by building with --enable-jit. When this is
429 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
430 runtime --no-jit option is given.
431
432 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
433 ovector, the return from the function was zero, implying that there were
434 other matches that did not fit. The correct "exactly full" value is now
435 returned.
436
437 6. If a subpattern that was called recursively or as a subroutine contained
438 (*PRUNE) or any other control that caused it to give a non-standard return,
439 invalid errors such as "Error -26 (nested recursion at the same subject
440 position)" or even infinite loops could occur.
441
442 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
443 computing the minimum length on reaching *ACCEPT, and so ended up with the
444 wrong value of 1 rather than 0. Further investigation indicates that
445 computing a minimum subject length in the presence of *ACCEPT is difficult
446 (think back references, subroutine calls), and so I have changed the code
447 so that no minimum is registered for a pattern that contains *ACCEPT.
448
449 8. If (*THEN) was present in the first (true) branch of a conditional group,
450 it was not handled as intended. [But see 16 below.]
451
452 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
453 Sheri Pierce.
454
455 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
456 the first byte in a match must be "a".
457
458 11. Change 17 for 8.13 increased the recursion depth for patterns like
459 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
460 pattern contains any instances of (*THEN). If it does not, the old
461 optimizations are restored. It would be nice to do this on a per-group
462 basis, but at the moment that is not feasible.
463
464 12. In some environments, the output of pcretest -C is CRLF terminated. This
465 broke RunTest's code that checks for the link size. A single white space
466 character after the value is now allowed for.
467
468 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
469 For "fr", it uses the Windows-specific input and output files.
470
471 14. If (*THEN) appeared in a group that was called recursively or as a
472 subroutine, it did not work as intended. [But see next item.]
473
474 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
475 pattern fragments (but not containing any | characters). If A and B are
476 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
477 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
478 D. In other words, Perl considers parentheses that do not contain any |
479 characters to be part of a surrounding alternative, whereas PCRE was
480 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
481 differently. PCRE now behaves in the same way as Perl, except in the case
482 of subroutine/recursion calls such as (?1) which have in any case always
483 been different (but PCRE had them first :-).
484
485 16. Related to 15 above: Perl does not treat the | in a conditional group as
486 creating alternatives. Such a group is treated in the same way as an
487 ordinary group without any | characters when processing (*THEN). PCRE has
488 been changed to match Perl's behaviour.
489
490 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
491 RunGrepTest script failed.
492
493 18. Change 22 for version 13 caused atomic groups to use more stack. This is
494 inevitable for groups that contain captures, but it can lead to a lot of
495 stack use in large patterns. The old behaviour has been restored for atomic
496 groups that do not contain any capturing parentheses.
497
498 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
499 suppress the check for a minimum subject length at run time. (If it was
500 given to pcre_exec() or pcre_dfa_exec() it did work.)
501
502 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
503 fail to work when decoding hex characters in data strings in EBCDIC
504 environments.
505
506 21. It appears that in at least one Mac OS environment, the isxdigit() function
507 is implemented as a macro that evaluates to its argument more than once,
508 contravening the C 90 Standard (I haven't checked a later standard). There
509 was an instance in pcretest which caused it to go wrong when processing
510 \x{...} escapes in subject strings. The has been rewritten to avoid using
511 things like p++ in the argument of isxdigit().
512
513
514 Version 8.13 16-Aug-2011
515 ------------------------
516
517 1. The Unicode data tables have been updated to Unicode 6.0.0.
518
519 2. Two minor typos in pcre_internal.h have been fixed.
520
521 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
522 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
523 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
524
525 4. There were a number of related bugs in the code for matching backrefences
526 caselessly in UTF-8 mode when codes for the characters concerned were
527 different numbers of bytes. For example, U+023A and U+2C65 are an upper
528 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
529 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
530 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
531 2-byte code at the end of the subject (it thought there wasn't enough data
532 left).
533
534 5. Comprehensive information about what went wrong is now returned by
535 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
536 as the output vector has at least 2 elements. The offset of the start of
537 the failing character and a reason code are placed in the vector.
538
539 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
540 now returned is for the first byte of the failing character, instead of the
541 last byte inspected. This is an incompatible change, but I hope it is small
542 enough not to be a problem. It makes the returned offset consistent with
543 pcre_exec() and pcre_dfa_exec().
544
545 7. pcretest now gives a text phrase as well as the error number when
546 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
547 failure, the offset and reason code are output.
548
549 8. When \R was used with a maximizing quantifier it failed to skip backwards
550 over a \r\n pair if the subsequent match failed. Instead, it just skipped
551 back over a single character (\n). This seems wrong (because it treated the
552 two characters as a single entity when going forwards), conflicts with the
553 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
554 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
555 has been changed.
556
557 9. Some internal refactoring has changed the processing so that the handling
558 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
559 time (the PCRE_DOTALL option was changed this way some time ago: version
560 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
561 which was always a bit of a fudge. It also means that there is one less
562 argument for the match() function, which reduces its stack requirements
563 slightly. This change also fixes an incompatibility with Perl: the pattern
564 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
565
566 10. More internal refactoring has drastically reduced the number of recursive
567 calls to match() for possessively repeated groups such as (abc)++ when
568 using pcre_exec().
569
570 11. While implementing 10, a number of bugs in the handling of groups were
571 discovered and fixed:
572
573 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
574 (a|)*(?1) gave a compile-time internal error.
575 ((a|)+)+ did not notice that the outer group could match an empty string.
576 (^a|^)+ was not marked as anchored.
577 (.*a|.*)+ was not marked as matching at start or after a newline.
578
579 12. Yet more internal refactoring has removed another argument from the match()
580 function. Special calls to this function are now indicated by setting a
581 value in a variable in the "match data" data block.
582
583 13. Be more explicit in pcre_study() instead of relying on "default" for
584 opcodes that mean there is no starting character; this means that when new
585 ones are added and accidentally left out of pcre_study(), testing should
586 pick them up.
587
588 14. The -s option of pcretest has been documented for ages as being an old
589 synonym of -m (show memory usage). I have changed it to mean "force study
590 for every regex", that is, assume /S for every regex. This is similar to -i
591 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
592 using it. It makes it easier to run collections of tests with and without
593 study enabled, and thereby test pcre_study() more easily. All the standard
594 tests are now run with and without -s (but some patterns can be marked as
595 "never study" - see 20 below).
596
597 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
598 restoration of the capturing data to the outer values was not happening
599 correctly.
600
601 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
602 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
603 pattern had matched an empty string, and so incorrectly returned a no
604 match.
605
606 17. There was optimizing code for the last branch of non-capturing parentheses,
607 and also for the obeyed branch of a conditional subexpression, which used
608 tail recursion to cut down on stack usage. Unfortunately, now that there is
609 the possibility of (*THEN) occurring in these branches, tail recursion is
610 no longer possible because the return has to be checked for (*THEN). These
611 two optimizations have therefore been removed. [But see 8.20/11 above.]
612
613 18. If a pattern containing \R was studied, it was assumed that \R always
614 matched two bytes, thus causing the minimum subject length to be
615 incorrectly computed because \R can also match just one byte.
616
617 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
618 was incorrectly computed.
619
620 20. If /S is present twice on a test pattern in pcretest input, it now
621 *disables* studying, thereby overriding the use of -s on the command line
622 (see 14 above). This is necessary for one or two tests to keep the output
623 identical in both cases.
624
625 21. When (*ACCEPT) was used in an assertion that matched an empty string and
626 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
627
628 22. When an atomic group that contained a capturing parenthesis was
629 successfully matched, but the branch in which it appeared failed, the
630 capturing was not being forgotten if a higher numbered group was later
631 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
632 group 1 to "a", when in fact it should be unset. This applied to multi-
633 branched capturing and non-capturing groups, repeated or not, and also to
634 positive assertions (capturing in negative assertions does not happen
635 in PCRE) and also to nested atomic groups.
636
637 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
638 subject after a captured substring, to make it easier to tell which of a
639 number of identical substrings has been captured.
640
641 24. The way atomic groups are processed by pcre_exec() has been changed so that
642 if they are repeated, backtracking one repetition now resets captured
643 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
644 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
645 "aaa". Previously, it would have been "a". As part of this code
646 refactoring, the way recursive calls are handled has also been changed.
647
648 25. If an assertion condition captured any substrings, they were not passed
649 back unless some other capturing happened later. For example, if
650 (?(?=(a))a) was matched against "a", no capturing was returned.
651
652 26. When studying a pattern that contained subroutine calls or assertions,
653 the code for finding the minimum length of a possible match was handling
654 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
655 group 1 called group 2 while simultaneously a separate group 2 called group
656 1). A stack overflow occurred in this case. I have fixed this by limiting
657 the recursion depth to 10.
658
659 27. Updated RunTest.bat in the distribution to the version supplied by Tom
660 Fortmann. This supports explicit test numbers on the command line, and has
661 argument validation and error reporting.
662
663 28. An instance of \X with an unlimited repeat could fail if at any point the
664 first character it looked at was a mark character.
665
666 29. Some minor code refactoring concerning Unicode properties and scripts
667 should reduce the stack requirement of match() slightly.
668
669 30. Added the '=' option to pcretest to check the setting of unused capturing
670 slots at the end of the pattern, which are documented as being -1, but are
671 not included in the return count.
672
673 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
674 compiled something random. Now it gives a compile-time error (as does
675 Perl).
676
677 32. A *MARK encountered during the processing of a positive assertion is now
678 recorded and passed back (compatible with Perl).
679
680 33. If --only-matching or --colour was set on a pcregrep call whose pattern
681 had alternative anchored branches, the search for a second match in a line
682 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
683 matched the line "0102" twice. The same bug affected patterns that started
684 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
685 twice.
686
687 34. Previously, PCRE did not allow quantification of assertions. However, Perl
688 does, and because of capturing effects, quantifying parenthesized
689 assertions may at times be useful. Quantifiers are now allowed for
690 parenthesized assertions.
691
692 35. A minor code tidy in pcre_compile() when checking options for \R usage.
693
694 36. \g was being checked for fancy things in a character class, when it should
695 just be a literal "g".
696
697 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
698 appearance of a nested POSIX class supersedes an apparent external class.
699 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
700 unescaped square brackets may also appear as part of class names. For
701 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
702 more like Perl. (But see 8.20/1 above.)
703
704 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
705 was because it thought it was \N{name}, which is not supported).
706
707 39. Add minix to OS list not supporting the -S option in pcretest.
708
709 40. PCRE tries to detect cases of infinite recursion at compile time, but it
710 cannot analyze patterns in sufficient detail to catch mutual recursions
711 such as ((?1))((?2)). There is now a runtime test that gives an error if a
712 subgroup is called recursively as a subpattern for a second time at the
713 same position in the subject string. In previous releases this might have
714 been caught by the recursion limit, or it might have run out of stack.
715
716 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
717 happen only once. PCRE was, however incorrectly giving a compile time error
718 "recursive call could loop indefinitely" because it cannot analyze the
719 pattern in sufficient detail. The compile time test no longer happens when
720 PCRE is compiling a conditional subpattern, but actual runaway loops are
721 now caught at runtime (see 40 above).
722
723 42. It seems that Perl allows any characters other than a closing parenthesis
724 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
725 has been changed to be the same.
726
727 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
728 as not to get warnings when autogen.sh is called. Also changed
729 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
730
731 44. To help people who use pcregrep to scan files containing exceedingly long
732 lines, the following changes have been made:
733
734 (a) The default value of the buffer size parameter has been increased from
735 8K to 20K. (The actual buffer used is three times this size.)
736
737 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
738 PCRE is built.
739
740 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
741 to be set at run time.
742
743 (d) Numerical values in pcregrep options can be followed by K or M, for
744 example --buffer-size=50K.
745
746 (e) If a line being scanned overflows pcregrep's buffer, an error is now
747 given and the return code is set to 2.
748
749 45. Add a pointer to the latest mark to the callout data block.
750
751 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
752 partial match of an empty string instead of no match. This was specific to
753 the use of ".".
754
755 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
756 complete match instead of a partial match. This bug was dependent on both
757 the PCRE_UTF8 and PCRE_DOTALL options being set.
758
759 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
760 starting byte set, because \b was not being ignored.
761
762
763 Version 8.12 15-Jan-2011
764 ------------------------
765
766 1. Fixed some typos in the markup of the man pages, and wrote a script that
767 checks for such things as part of the documentation building process.
768
769 2. On a big-endian 64-bit system, pcregrep did not correctly process the
770 --match-limit and --recursion-limit options (added for 8.11). In
771 particular, this made one of the standard tests fail. (The integer value
772 went into the wrong half of a long int.)
773
774 3. If the --colour option was given to pcregrep with -v (invert match), it
775 did strange things, either producing crazy output, or crashing. It should,
776 of course, ignore a request for colour when reporting lines that do not
777 match.
778
779 4. Another pcregrep bug caused similar problems if --colour was specified with
780 -M (multiline) and the pattern match finished with a line ending.
781
782 5. In pcregrep, when a pattern that ended with a literal newline sequence was
783 matched in multiline mode, the following line was shown as part of the
784 match. This seems wrong, so I have changed it.
785
786 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
787 the check for further matches in the same line (so they could be coloured)
788 to overrun the end of the current line. If another match was found, it was
789 incorrectly shown (and then shown again when found in the next line).
790
791 7. If pcregrep was compiled under Windows, there was a reference to the
792 function pcregrep_exit() before it was defined. I am assuming this was
793 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
794 reported by a user. I've moved the definition above the reference.
795
796
797 Version 8.11 10-Dec-2010
798 ------------------------
799
800 1. (*THEN) was not working properly if there were untried alternatives prior
801 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
802 backtracked to try for "b" instead of moving to the next alternative branch
803 at the same level (in this case, to look for "c"). The Perl documentation
804 is clear that when (*THEN) is backtracked onto, it goes to the "next
805 alternative in the innermost enclosing group".
806
807 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
808 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
809 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
810 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
811 (*THEN).
812
813 3. If \s appeared in a character class, it removed the VT character from
814 the class, even if it had been included by some previous item, for example
815 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
816 of \s, but is part of the POSIX "space" class.)
817
818 4. A partial match never returns an empty string (because you can always
819 match an empty string at the end of the subject); however the checking for
820 an empty string was starting at the "start of match" point. This has been
821 changed to the "earliest inspected character" point, because the returned
822 data for a partial match starts at this character. This means that, for
823 example, /(?<=abc)def/ gives a partial match for the subject "abc"
824 (previously it gave "no match").
825
826 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
827 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
828 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
829 has an implication that the given string is incomplete (because a partial
830 match is preferred over a full match). For this reason, these items now
831 give a partial match in this situation. [Aside: previously, the one case
832 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
833 match rather than a full match, which was wrong by the old rules, but is
834 now correct.]
835
836 6. There was a bug in the handling of #-introduced comments, recognized when
837 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
838 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
839 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
840 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
841 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
842 places in pcre_compile().
843
844 7. Related to (6) above, when pcre_compile() was skipping #-introduced
845 comments when looking ahead for named forward references to subpatterns,
846 the only newline sequence it recognized was NL. It now handles newlines
847 according to the set newline convention.
848
849 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
850 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
851 cater for a lack of strerror(). These oversights have been fixed.
852
853 9. Added --match-limit and --recursion-limit to pcregrep.
854
855 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
856
857 11. When the -o option was used, pcregrep was setting a return code of 1, even
858 when matches were found, and --line-buffered was not being honoured.
859
860 12. Added an optional parentheses number to the -o and --only-matching options
861 of pcregrep.
862
863 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
864 can match an empty string. The code to do it in pcretest and pcredemo
865 needed fixing:
866
867 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
868 only one byte after an empty string match just before CRLF (this case
869 just got forgotten; "any" and "anycrlf" were OK).
870
871 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
872 mode when an empty string match preceded an ASCII character followed by
873 a non-ASCII character. (The code for advancing by one character rather
874 than one byte was nonsense.)
875
876 (c) The pcredemo.c sample program did not have any code at all to handle
877 the cases when CRLF is a valid newline sequence.
878
879 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
880 as a starting offset was within the subject string. There is now a new
881 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
882 negative or greater than the length of the string. In order to test this,
883 pcretest is extended to allow the setting of negative starting offsets.
884
885 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
886 starting offset points to the beginning of a UTF-8 character was
887 unnecessarily clumsy. I tidied it up.
888
889 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
890 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
891
892 17. Nobody had reported that the --include_dir option, which was added in
893 release 7.7 should have been called --include-dir (hyphen, not underscore)
894 for compatibility with GNU grep. I have changed it to --include-dir, but
895 left --include_dir as an undocumented synonym, and the same for
896 --exclude-dir, though that is not available in GNU grep, at least as of
897 release 2.5.4.
898
899 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
900 characters from a string of bytes) have been redefined so as not to use
901 loops, in order to improve performance in some environments. At the same
902 time, I abstracted some of the common code into auxiliary macros to save
903 repetition (this should not affect the compiled code).
904
905 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
906 compile-time error is now given if \c is not followed by an ASCII
907 character, that is, a byte less than 128. (In EBCDIC mode, the code is
908 different, and any byte value is allowed.)
909
910 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
911 START_OPTIMIZE option, which is now allowed at compile time - but just
912 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
913 to pcregrep and other applications that have no direct access to PCRE
914 options. The new /Y option in pcretest sets this option when calling
915 pcre_compile().
916
917 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
918 back references. Groups containing recursive back references were forced to
919 be atomic by that change, but in the case of named groups, the amount of
920 memory required was incorrectly computed, leading to "Failed: internal
921 error: code overflow". This has been fixed.
922
923 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
924 pcretest.c, to avoid build problems in some Borland environments.
925
926
927 Version 8.10 25-Jun-2010
928 ------------------------
929
930 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
931 THEN.
932
933 2. (*ACCEPT) was not working when inside an atomic group.
934
935 3. Inside a character class, \B is treated as a literal by default, but
936 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
937 causes the error). The code is unchanged, but I tidied the documentation.
938
939 4. Inside a character class, PCRE always treated \R and \X as literals,
940 whereas Perl faults them if its -w option is set. I have changed PCRE so
941 that it faults them when PCRE_EXTRA is set.
942
943 5. Added support for \N, which always matches any character other than
944 newline. (It is the same as "." when PCRE_DOTALL is not set.)
945
946 6. When compiling pcregrep with newer versions of gcc which may have
947 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
948 declared with attribute warn_unused_result" were given. Just casting the
949 result to (void) does not stop the warnings; a more elaborate fudge is
950 needed. I've used a macro to implement this.
951
952 7. Minor change to pcretest.c to avoid a compiler warning.
953
954 8. Added four artifical Unicode properties to help with an option to make
955 \s etc use properties (see next item). The new properties are: Xan
956 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
957
958 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
959 use Unicode properties. (*UCP) at the start of a pattern can be used to set
960 this option. Modified pcretest to add /W to test this facility. Added
961 REG_UCP to make it available via the POSIX interface.
962
963 10. Added --line-buffered to pcregrep.
964
965 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
966 studied, and the match started with a letter with a code point greater than
967 127 whose first byte was different to the first byte of the other case of
968 the letter, the other case of this starting letter was not recognized
969 (#976).
970
971 12. If a pattern that was studied started with a repeated Unicode property
972 test, for example, \p{Nd}+, there was the theoretical possibility of
973 setting up an incorrect bitmap of starting bytes, but fortunately it could
974 not have actually happened in practice until change 8 above was made (it
975 added property types that matched character-matching opcodes).
976
977 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
978 possible starting bytes for non-anchored patterns.
979
980 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
981 \R, and also a number of cases that involve Unicode properties, both
982 explicit and implicit when PCRE_UCP is set.
983
984 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
985 input, it could crash or give wrong results if characters with values
986 greater than 0xc0 were present in the subject string. (Detail: it assumed
987 UTF-8 input when processing these items.)
988
989 16. Added a lot of (int) casts to avoid compiler warnings in systems where
990 size_t is 64-bit (#991).
991
992 17. Added a check for running out of memory when PCRE is compiled with
993 --disable-stack-for-recursion (#990).
994
995 18. If the last data line in a file for pcretest does not have a newline on
996 the end, a newline was missing in the output.
997
998 19. The default pcre_chartables.c file recognizes only ASCII characters (values
999 less than 128) in its various bitmaps. However, there is a facility for
1000 generating tables according to the current locale when PCRE is compiled. It
1001 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1002 characters, are recognized by isspace() and therefore were getting set in
1003 these tables, and indeed these tables seem to approximate to ISO 8859. This
1004 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1005 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1006 which of course cannot start UTF-8 characters. I have changed the code so
1007 that only real ASCII characters (less than 128) and the correct starting
1008 bytes for UTF-8 encodings are set for characters greater than 127 when in
1009 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1010 altogether.)
1011
1012 20. Added the /T option to pcretest so as to be able to run tests with non-
1013 standard character tables, thus making it possible to include the tests
1014 used for 19 above in the standard set of tests.
1015
1016 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1017 reference to a subpattern the other side of a comment that contains an
1018 opening parenthesis caused either an internal compiling error, or a
1019 reference to the wrong subpattern.
1020
1021
1022 Version 8.02 19-Mar-2010
1023 ------------------------
1024
1025 1. The Unicode data tables have been updated to Unicode 5.2.0.
1026
1027 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1028 configured.
1029
1030 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1031 original author of that file, following a query about its status.
1032
1033 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1034 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1035
1036 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1037 quantifier applied to a forward-referencing subroutine call, could compile
1038 incorrect code or give the error "internal error: previously-checked
1039 referenced subpattern not found".
1040
1041 6. Both MS Visual Studio and Symbian OS have problems with initializing
1042 variables to point to external functions. For these systems, therefore,
1043 pcre_malloc etc. are now initialized to local functions that call the
1044 relevant global functions.
1045
1046 7. There were two entries missing in the vectors called coptable and poptable
1047 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1048 I've fixed the data, and added a kludgy way of testing at compile time that
1049 the lengths are correct (equal to the number of opcodes).
1050
1051 8. Following on from 7, I added a similar kludge to check the length of the
1052 eint vector in pcreposix.c.
1053
1054 9. Error texts for pcre_compile() are held as one long string to avoid too
1055 much relocation at load time. To find a text, the string is searched,
1056 counting zeros. There was no check for running off the end of the string,
1057 which could happen if a new error number was added without updating the
1058 string.
1059
1060 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1061
1062 11. \K was not working if it appeared in an atomic group or in a group that
1063 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1064 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1065 the assertion is positive, but not if it is negative.
1066
1067 12. Change 11 fortuitously reduced the size of the stack frame used in the
1068 "match()" function of pcre_exec.c by one pointer. Forthcoming
1069 implementation of support for (*MARK) will need an extra pointer on the
1070 stack; I have reserved it now, so that the stack frame size does not
1071 decrease.
1072
1073 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1074 item in branch that calls a recursion is a subroutine call - as in the
1075 second branch in the above example - was incorrectly given the compile-
1076 time error "recursive call could loop indefinitely" because pcre_compile()
1077 was not correctly checking the subroutine for matching a non-empty string.
1078
1079 14. The checks for overrunning compiling workspace could trigger after an
1080 overrun had occurred. This is a "should never occur" error, but it can be
1081 triggered by pathological patterns such as hundreds of nested parentheses.
1082 The checks now trigger 100 bytes before the end of the workspace.
1083
1084 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1085
1086
1087 Version 8.01 19-Jan-2010
1088 ------------------------
1089
1090 1. If a pattern contained a conditional subpattern with only one branch (in
1091 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1092 computed the wrong minimum data length (which is of course zero for such
1093 subpatterns). This could cause incorrect "no match" results.
1094
1095 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1096 the pattern is reset in the first branch, pcre_compile() failed with
1097 "internal error: code overflow at offset...". This happened only when
1098 the reset was to the original external option setting. (An optimization
1099 abstracts leading options settings into an external setting, which was the
1100 cause of this.)
1101
1102 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1103 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1104 assertion pattern did not match (meaning that the assertion was true), it
1105 was incorrectly treated as false if the SKIP had been reached during the
1106 matching. This also applied to assertions used as conditions.
1107
1108 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1109 assertion subpattern, including such a pattern used as a condition,
1110 unpredictable results occurred, instead of the error return
1111 PCRE_ERROR_DFA_UITEM.
1112
1113 5. The C++ GlobalReplace function was not working like Perl for the special
1114 situation when an empty string is matched. It now does the fancy magic
1115 stuff that is necessary.
1116
1117 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1118 removed. (These were left over from very, very early versions of PCRE.)
1119
1120 7. Some cosmetic changes to the code to make life easier when compiling it
1121 as part of something else:
1122
1123 (a) Change DEBUG to PCRE_DEBUG.
1124
1125 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1126 called "current" as "current_branch", to prevent a collision with the
1127 Linux macro when compiled as a kernel module.
1128
1129 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1130 prevent a collision with the Linux macro when compiled as a kernel
1131 module.
1132
1133 8. In pcre_compile() there are some checks for integer overflows that used to
1134 cast potentially large values to (double). This has been changed to that
1135 when building, a check for int64_t is made, and if it is found, it is used
1136 instead, thus avoiding the use of floating point arithmetic. (There is no
1137 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1138 double.
1139
1140 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1141 2005 (difference between two addresses compared to an unsigned value).
1142
1143 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1144 custom one, because of the following reported problem in Windows:
1145
1146 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1147 under Win32.
1148 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1149 therefore missing the function definition.
1150 - The compiler thus generates a "C" signature for the test function.
1151 - The linker fails to find the "C" function.
1152 - PCRE fails to configure if asked to do so against libbz2.
1153
1154 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1155 messages were output:
1156
1157 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1158 rerunning libtoolize, to keep the correct libtool macros in-tree.
1159 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1160
1161 I have done both of these things.
1162
1163 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1164 most of the time, it *can* run out if it is given a pattern that contains a
1165 runaway infinite recursion. I updated the discussion in the pcrestack man
1166 page.
1167
1168 13. Now that we have gone to the x.xx style of version numbers, the minor
1169 version may start with zero. Using 08 or 09 is a bad idea because users
1170 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1171 interpreted as invalid octal numbers. I've updated the previous comment in
1172 configure.ac, and also added a check that gives an error if 08 or 09 are
1173 used.
1174
1175 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1176 causing partial matching to fail when the end of the subject matched \W
1177 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1178
1179 15. There were some discrepancies between the declarations in pcre_internal.h
1180 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1181 their definitions. The declarations used "const uschar *" and the
1182 definitions used USPTR. Even though USPTR is normally defined as "const
1183 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1184 reported that: "This difference in casting confuses some C++ compilers, for
1185 example, SunCC recognizes above declarations as different functions and
1186 generates broken code for hbpcre." I have changed the declarations to use
1187 USPTR.
1188
1189 16. GNU libtool is named differently on some systems. The autogen.sh script now
1190 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1191 (FreeBSD).
1192
1193 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1194 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1195 comment: "Figure out how to create a longlong from a string: strtoll and
1196 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1197 instance, but it only takes 2 args instead of 3!"
1198
1199 18. A subtle bug concerned with back references has been fixed by a change of
1200 specification, with a corresponding code fix. A pattern such as
1201 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1202 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1203 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1204 same bug. Such groups have to be quantified to be useful, or contained
1205 inside another quantified group. (If there's no repetition, the reference
1206 can never match.) The problem arises because, having left the group and
1207 moved on to the rest of the pattern, a later failure that backtracks into
1208 the group uses the captured value from the final iteration of the group
1209 rather than the correct earlier one. I have fixed this in PCRE by forcing
1210 any group that contains a reference to itself to be an atomic group; that
1211 is, there cannot be any backtracking into it once it has completed. This is
1212 similar to recursive and subroutine calls.
1213
1214
1215 Version 8.00 19-Oct-09
1216 ----------------------
1217
1218 1. The table for translating pcre_compile() error codes into POSIX error codes
1219 was out-of-date, and there was no check on the pcre_compile() error code
1220 being within the table. This could lead to an OK return being given in
1221 error.
1222
1223 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1224 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1225 in a Windows environment.
1226
1227 3. The pcregrep --count option prints the count for each file even when it is
1228 zero, as does GNU grep. However, pcregrep was also printing all files when
1229 --files-with-matches was added. Now, when both options are given, it prints
1230 counts only for those files that have at least one match. (GNU grep just
1231 prints the file name in this circumstance, but including the count seems
1232 more useful - otherwise, why use --count?) Also ensured that the
1233 combination -clh just lists non-zero counts, with no names.
1234
1235 4. The long form of the pcregrep -F option was incorrectly implemented as
1236 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1237 but it seems right to fix it, and I didn't think it was worth preserving
1238 the old behaviour.
1239
1240 5. The command line items --regex=pattern and --regexp=pattern were not
1241 recognized by pcregrep, which required --regex pattern or --regexp pattern
1242 (with a space rather than an '='). The man page documented the '=' forms,
1243 which are compatible with GNU grep; these now work.
1244
1245 6. No libpcreposix.pc file was created for pkg-config; there was just
1246 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1247
1248 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1249 when UCP support is not needed, by modifying the Python script that
1250 generates it from Unicode data files. This should not matter if the module
1251 is correctly used as a library, but I received one complaint about 50K of
1252 unwanted data. My guess is that the person linked everything into his
1253 program rather than using a library. Anyway, it does no harm.
1254
1255 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1256 was a minimum greater than 1 for a wide character in a possessive
1257 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1258 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1259 character. Chaos in the form of incorrect output or a compiling loop could
1260 result.
1261
1262 9. The restrictions on what a pattern can contain when partial matching is
1263 requested for pcre_exec() have been removed. All patterns can now be
1264 partially matched by this function. In addition, if there are at least two
1265 slots in the offset vector, the offset of the earliest inspected character
1266 for the match and the offset of the end of the subject are set in them when
1267 PCRE_ERROR_PARTIAL is returned.
1268
1269 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1270 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1271 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1272 and may be more useful for multi-segment matching.
1273
1274 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1275 used to be given if ever the end of the subject was reached; now it is
1276 given only if matching could not proceed because another character was
1277 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1278 string "Z", which now yields "no match" instead of "partial match". In the
1279 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1280 final character ended with (*FAIL).
1281
1282 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1283 if the pattern had a "must contain" character that was already found in the
1284 earlier partial match, unless partial matching was again requested. For
1285 example, with the pattern /dog.(body)?/, the "must contain" character is
1286 "g". If the first part-match was for the string "dog", restarting with
1287 "sbody" failed. This bug has been fixed.
1288
1289 13. The string returned by pcre_dfa_exec() after a partial match has been
1290 changed so that it starts at the first inspected character rather than the
1291 first character of the match. This makes a difference only if the pattern
1292 starts with a lookbehind assertion or \b or \B (\K is not supported by
1293 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1294 matching functions compatible, and I think it's the right thing to do.
1295
1296 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1297 so that the demonstration program is easily available in environments where
1298 PCRE has not been installed from source.
1299
1300 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1301 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1302 library.
1303
1304 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1305 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1306 is not the first non-POSIX option to be added. Clearly some people find
1307 these options useful.
1308
1309 17. If a caller to the POSIX matching function regexec() passes a non-zero
1310 value for nmatch with a NULL value for pmatch, the value of
1311 nmatch is forced to zero.
1312
1313 18. RunGrepTest did not have a test for the availability of the -u option of
1314 the diff command, as RunTest does. It now checks in the same way as
1315 RunTest, and also checks for the -b option.
1316
1317 19. If an odd number of negated classes containing just a single character
1318 interposed, within parentheses, between a forward reference to a named
1319 subpattern and the definition of the subpattern, compilation crashed with
1320 an internal error, complaining that it could not find the referenced
1321 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1322 [The bug was that it was starting one character too far in when skipping
1323 over the character class, thus treating the ] as data rather than
1324 terminating the class. This meant it could skip too much.]
1325
1326 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1327 /g option in pcretest when the pattern contains \K, which makes it possible
1328 to have an empty string match not at the start, even when the pattern is
1329 anchored. Updated pcretest and pcredemo to use this option.
1330
1331 21. If the maximum number of capturing subpatterns in a recursion was greater
1332 than the maximum at the outer level, the higher number was returned, but
1333 with unset values at the outer level. The correct (outer level) value is
1334 now given.
1335
1336 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1337 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1338 make it do so. The string so far is captured, making this feature
1339 compatible with Perl.
1340
1341 23. The tests have been re-organized, adding tests 11 and 12, to make it
1342 possible to check the Perl 5.10 features against Perl 5.10.
1343
1344 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1345 pattern matches a fixed length string. PCRE did not allow this; now it
1346 does. Neither allows recursion.
1347
1348 25. I finally figured out how to implement a request to provide the minimum
1349 length of subject string that was needed in order to match a given pattern.
1350 (It was back references and recursion that I had previously got hung up
1351 on.) This code has now been added to pcre_study(); it finds a lower bound
1352 to the length of subject needed. It is not necessarily the greatest lower
1353 bound, but using it to avoid searching strings that are too short does give
1354 some useful speed-ups. The value is available to calling programs via
1355 pcre_fullinfo().
1356
1357 26. While implementing 25, I discovered to my embarrassment that pcretest had
1358 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1359 study optimizations had never been tested with that matching function.
1360 Oops. What is worse, even when it was passed study data, there was a bug in
1361 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1362 were also very few tests of studied patterns with pcre_dfa_exec().
1363
1364 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1365 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1366 on the other side of the coin, they are no longer allowed to have different
1367 names, because these cannot be distinguished in PCRE, and this has caused
1368 confusion. (This is a difference from Perl.)
1369
1370 28. When duplicate subpattern names are present (necessarily with different
1371 numbers, as required by 27 above), and a test is made by name in a
1372 conditional pattern, either for a subpattern having been matched, or for
1373 recursion in such a pattern, all the associated numbered subpatterns are
1374 tested, and the overall condition is true if the condition is true for any
1375 one of them. This is the way Perl works, and is also more like the way
1376 testing by number works.
1377
1378
1379 Version 7.9 11-Apr-09
1380 ---------------------
1381
1382 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1383 (pcretest), all targets were linked against these libraries. This included
1384 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1385 libraries. This caused unwanted dependencies to be created. This problem
1386 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1387 pcretest is linked with readline.
1388
1389 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1390 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1391 moved outside it again, because FALSE and TRUE are already defined in AIX,
1392 but BOOL is not.
1393
1394 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1395 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1396
1397 4. The pcregrep documentation said spaces were inserted as well as colons (or
1398 hyphens) following file names and line numbers when outputting matching
1399 lines. This is not true; no spaces are inserted. I have also clarified the
1400 wording for the --colour (or --color) option.
1401
1402 5. In pcregrep, when --colour was used with -o, the list of matching strings
1403 was not coloured; this is different to GNU grep, so I have changed it to be
1404 the same.
1405
1406 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1407 each matching line was coloured. Now it goes on to look for further matches
1408 of any of the test patterns, which is the same behaviour as GNU grep.
1409
1410 7. A pattern that could match an empty string could cause pcregrep to loop; it
1411 doesn't make sense to accept an empty string match in pcregrep, so I have
1412 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1413 seems to be how GNU grep behaves.
1414
1415 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1416 start or after a newline", because the conditional assertion was not being
1417 correctly handled. The rule now is that both the assertion and what follows
1418 in the first alternative must satisfy the test.
1419
1420 9. If auto-callout was enabled in a pattern with a conditional group whose
1421 condition was an assertion, PCRE could crash during matching, both with
1422 pcre_exec() and pcre_dfa_exec().
1423
1424 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1425 used for matching.
1426
1427 11. Unicode property support in character classes was not working for
1428 characters (bytes) greater than 127 when not in UTF-8 mode.
1429
1430 12. Added the -M command line option to pcretest.
1431
1432 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1433
1434 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1435
1436 16. Added comments and documentation about mis-use of no_arg in the C++
1437 wrapper.
1438
1439 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1440 from Martin Jerabek that uses macro names for all relevant character and
1441 string constants.
1442
1443 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1444 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1445 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1446 these, but not everybody uses configure.
1447
1448 19. A conditional group that had only one branch was not being correctly
1449 recognized as an item that could match an empty string. This meant that an
1450 enclosing group might also not be so recognized, causing infinite looping
1451 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1452 with the subject "ab", where knowledge that the repeated group can match
1453 nothing is needed in order to break the loop.
1454
1455 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1456 exec(), but without supplying a callout function, matching went wrong.
1457
1458 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1459 leak if the size of the offset vector was greater than 30. When the vector
1460 is smaller, the saved offsets during recursion go onto a local stack
1461 vector, but for larger vectors malloc() is used. It was failing to free
1462 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1463 error, in fact).
1464
1465 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1466 heapframe that is used only when UTF-8 support is enabled. This caused no
1467 problem, but was untidy.
1468
1469 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1470 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1471 included within another project.
1472
1473 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1474 slightly modified by me:
1475
1476 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1477 not building pcregrep.
1478
1479 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1480 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1481
1482 25. Forward references, both numeric and by name, in patterns that made use of
1483 duplicate group numbers, could behave incorrectly or give incorrect errors,
1484 because when scanning forward to find the reference group, PCRE was not
1485 taking into account the duplicate group numbers. A pattern such as
1486 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1487
1488 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1489 the feature of a custom pointer more persuasive (as requested by a user).
1490
1491 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1492 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1493 pre-defined.
1494
1495 28. Added support for (*UTF8) at the start of a pattern.
1496
1497 29. Arrange for flags added by the "release type" setting in CMake to be shown
1498 in the configuration summary.
1499
1500
1501 Version 7.8 05-Sep-08
1502 ---------------------
1503
1504 1. Replaced UCP searching code with optimized version as implemented for Ad
1505 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1506 stage table and inline lookup instead of a function, giving speed ups of 2
1507 to 5 times on some simple patterns that I tested. Permission was given to
1508 distribute the MultiStage2.py script that generates the tables (it's not in
1509 the tarball, but is in the Subversion repository).
1510
1511 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1512 scripts.
1513
1514 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1515 a group with a zero qualifier. The result of the study could be incorrect,
1516 or the function might crash, depending on the pattern.
1517
1518 4. Caseless matching was not working for non-ASCII characters in back
1519 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1520 It now works when Unicode Property Support is available.
1521
1522 5. In pcretest, an escape such as \x{de} in the data was always generating
1523 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1524 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1525 truncation.
1526
1527 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1528
1529 7. Added two (int) casts to pcregrep when printing the difference of two
1530 pointers, in case they are 64-bit values.
1531
1532 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1533 test 2 if it fails.
1534
1535 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1536 and a #define of that name to empty if it is not externally set. This is to
1537 allow users of MSVC to set it if necessary.
1538
1539 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1540 the convenience functions in the pcre_get.c source file.
1541
1542 11. An option change at the start of a pattern that had top-level alternatives
1543 could cause overwriting and/or a crash. This command provoked a crash in
1544 some environments:
1545
1546 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1547
1548 This potential security problem was recorded as CVE-2008-2371.
1549
1550 12. For a pattern where the match had to start at the beginning or immediately
1551 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1552 pcre_dfa_exec() could read past the end of the passed subject if there was
1553 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1554 pcretest so that it places the subject at the end of its malloc-ed buffer.
1555
1556 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1557 exec() might read past the end of the data buffer in UTF-8 mode.
1558
1559 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1560 the data contained the byte 0x85 as part of a UTF-8 character within its
1561 first line. This applied both to normal and DFA matching.
1562
1563 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1564 /^[^d]*?$/8 failed to match "abc".
1565
1566 16. Added a missing copyright notice to pcrecpp_internal.h.
1567
1568 17. Make it more clear in the documentation that values returned from
1569 pcre_exec() in ovector are byte offsets, not character counts.
1570
1571 18. Tidied a few places to stop certain compilers from issuing warnings.
1572
1573 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1574 supplied by Stefan Weber. I made a further small update for 7.8 because
1575 there is a change of source arrangements: the pcre_searchfuncs.c module is
1576 replaced by pcre_ucd.c.
1577
1578
1579 Version 7.7 07-May-08
1580 ---------------------
1581
1582 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1583 a string to a long long, pretend we don't even have a long long." This is
1584 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1585
1586 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1587 pre-7.6 versions, which defined a global no_arg variable instead of putting
1588 it in the RE class. (See also #8 below.)
1589
1590 3. Remove a line of dead code, identified by coverity and reported by Nuno
1591 Lopes.
1592
1593 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1594
1595 (1) The include/exclude patterns were being applied to the whole pathnames
1596 of files, instead of just to the final components.
1597
1598 (2) If there was more than one level of directory, the subdirectories were
1599 skipped unless they satisfied the include/exclude conditions. This is
1600 inconsistent with GNU grep (and could even be seen as contrary to the
1601 pcregrep specification - which I improved to make it absolutely clear).
1602 The action now is always to scan all levels of directory, and just
1603 apply the include/exclude patterns to regular files.
1604
1605 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1606 --exclude_dir in the tests to avoid scanning .svn directories.
1607
1608 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1609 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1610 doesn't support NULs in patterns.
1611
1612 7. Added some missing "const"s to declarations of static tables in
1613 pcre_compile.c and pcre_dfa_exec.c.
1614
1615 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1616 caused by fix #2 above. (Subsequently also a second patch to fix the
1617 first patch. And a third patch - this was a messy problem.)
1618
1619 9. Applied Craig's patch to remove the use of push_back().
1620
1621 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1622 matching function regexec().
1623
1624 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1625 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1626 references. PCRE supports relative numbers with this syntax (I don't think
1627 Oniguruma does).
1628
1629 12. Previously, a group with a zero repeat such as (...){0} was completely
1630 omitted from the compiled regex. However, this means that if the group
1631 was called as a subroutine from elsewhere in the pattern, things went wrong
1632 (an internal error was given). Such groups are now left in the compiled
1633 pattern, with a new opcode that causes them to be skipped at execution
1634 time.
1635
1636 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1637 to the way PCRE behaves:
1638
1639 (a) A lone ] character is dis-allowed (Perl treats it as data).
1640
1641 (b) A back reference to an unmatched subpattern matches an empty string
1642 (Perl fails the current match path).
1643
1644 (c) A data ] in a character class must be notated as \] because if the
1645 first data character in a class is ], it defines an empty class. (In
1646 Perl it is not possible to have an empty class.) The empty class []
1647 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1648 The negative empty class [^] matches any one character, independently
1649 of the DOTALL setting.
1650
1651 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1652 non-existent subpattern following a character class starting with ']' and
1653 containing () gave an internal compiling error instead of "reference to
1654 non-existent subpattern". Fortunately, when the pattern did exist, the
1655 compiled code was correct. (When scanning forwards to check for the
1656 existencd of the subpattern, it was treating the data ']' as terminating
1657 the class, so got the count wrong. When actually compiling, the reference
1658 was subsequently set up correctly.)
1659
1660 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1661 it was being rejected as not supported by pcre_dfa_exec(), even though
1662 other assertions are supported. I have made pcre_dfa_exec() support
1663 (*FAIL).
1664
1665 16. The implementation of 13c above involved the invention of a new opcode,
1666 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1667 cannot be changed at match time, I realized I could make a small
1668 improvement to matching performance by compiling OP_ALLANY instead of
1669 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1670 on the OP_ANY path.
1671
1672 17. Compiling pcretest on Windows with readline support failed without the
1673 following two fixes: (1) Make the unistd.h include conditional on
1674 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1675
1676 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1677 ncurses library to be included for pcretest when ReadLine support is
1678 requested, but also to allow for it to be overridden. This patch came from
1679 Daniel Bergström.
1680
1681 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1682 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1683 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1684 spotting this.
1685
1686
1687 Version 7.6 28-Jan-08
1688 ---------------------
1689
1690 1. A character class containing a very large number of characters with
1691 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1692 overflow.
1693
1694 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1695 HAVE_LONG_LONG is not defined.
1696
1697 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1698 bring them up to date and include new features. This patch includes:
1699
1700 - Fixed PH's badly added libz and libbz2 support.
1701 - Fixed a problem with static linking.
1702 - Added pcredemo. [But later removed - see 7 below.]
1703 - Fixed dftables problem and added an option.
1704 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1705 HAVE_LONG_LONG.
1706 - Added readline support for pcretest.
1707 - Added an listing of the option settings after cmake has run.
1708
1709 4. A user submitted a patch to Makefile that makes it easy to create
1710 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1711 Makefile.am that cause it to include this special target, without
1712 affecting anything else. Note that the same mingw target plus all
1713 the other distribution libraries and programs are now supported
1714 when configuring with CMake (see 6 below) instead of with
1715 Configure/Make.
1716
1717 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1718 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1719 exported in the Windows port". It has not yet been confirmed that the patch
1720 solves the problem, but it does no harm.
1721
1722 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1723 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1724 with CMake, and also correct the comment about stack recursion.
1725
1726 7. Remove the automatic building of pcredemo from the ./configure system and
1727 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1728 of a program that users should build themselves after PCRE is installed, so
1729 building it automatically is not really right. What is more, it gave
1730 trouble in some build environments.
1731
1732 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1733
1734
1735 Version 7.5 10-Jan-08
1736 ---------------------
1737
1738 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1739 values in parens when parsing an RE using the C++ wrapper."
1740
1741 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1742 Characters greater than 255 were excluded from the class instead of being
1743 included.
1744
1745 3. The same bug as (2) above applied to negated POSIX classes such as
1746 [:^space:].
1747
1748 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1749 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1750 I have changed it.
1751
1752 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1753 first named subpattern) and a construct such as (?&a) would reference the
1754 first named subpattern whose name started with "a" (in other words, the
1755 length check was missing). Both these problems are fixed. "Subpattern name
1756 expected" is now given for (?&) (a zero-length name), and this patch also
1757 makes it give the same error for \k'' (previously it complained that that
1758 was a reference to a non-existent subpattern).
1759
1760 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1761 this is right because (?- can be followed by option settings as well as by
1762 digits. I have, however, made the messages clearer.
1763
1764 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1765 than the number used in the conditional) now cause a compile-time error.
1766 This is actually not compatible with Perl, which accepts such patterns, but
1767 treats the conditional as always being FALSE (as PCRE used to), but it
1768 seems to me that giving a diagnostic is better.
1769
1770 8. Change "alphameric" to the more common word "alphanumeric" in comments
1771 and messages.
1772
1773 9. Fix two occurrences of "backslash" in comments that should have been
1774 "backspace".
1775
1776 10. Remove two redundant lines of code that can never be obeyed (their function
1777 was moved elsewhere).
1778
1779 11. The program that makes PCRE's Unicode character property table had a bug
1780 which caused it to generate incorrect table entries for sequences of
1781 characters that have the same character type, but are in different scripts.
1782 It amalgamated them into a single range, with the script of the first of
1783 them. In other words, some characters were in the wrong script. There were
1784 thirteen such cases, affecting characters in the following ranges:
1785
1786 U+002b0 - U+002c1
1787 U+0060c - U+0060d
1788 U+0061e - U+00612
1789 U+0064b - U+0065e
1790 U+0074d - U+0076d
1791 U+01800 - U+01805
1792 U+01d00 - U+01d77
1793 U+01d9b - U+01dbf
1794 U+0200b - U+0200f
1795 U+030fc - U+030fe
1796 U+03260 - U+0327f
1797 U+0fb46 - U+0fbb1
1798 U+10450 - U+1049d
1799
1800 12. The -o option (show only the matching part of a line) for pcregrep was not
1801 compatible with GNU grep in that, if there was more than one match in a
1802 line, it showed only the first of them. It now behaves in the same way as
1803 GNU grep.
1804
1805 13. If the -o and -v options were combined for pcregrep, it printed a blank
1806 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1807 does the same. The return code can be used to tell if there were any
1808 non-matching lines.
1809
1810 14. Added --file-offsets and --line-offsets to pcregrep.
1811
1812 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1813 infinitely looping recursion. The bug was that positive lookaheads were not
1814 being skipped when checking for a possible empty match (negative lookaheads
1815 and both kinds of lookbehind were skipped).
1816
1817 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1818 inclusion of <windows.h> to before rather than after the definition of
1819 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1820
1821 17. Specifying a possessive quantifier with a specific limit for a Unicode
1822 character property caused pcre_compile() to compile bad code, which led at
1823 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1824 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1825 caused the error; without that there was no problem.
1826
1827 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1828
1829 19. Added --enable-pcretest-libreadline.
1830
1831 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1832 RE::GlobalReplace(). As a result, the number of replacements returned was
1833 double what it should be. I removed one of the increments, but Craig sent a
1834 later patch that removed the other one (the right fix) and added unit tests
1835 that check the return values (which was not done before).
1836
1837 21. Several CMake things:
1838
1839 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1840 the names libpcre and libpcreposix, not just pcre and pcreposix.
1841
1842 (2) The above change means that pcretest and pcregrep are now correctly
1843 linked with the newly-built libraries, not previously installed ones.
1844
1845 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1846
1847 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1848 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1849 UTF-8 newline character). The key issue is that the pattern starts .*;
1850 this means that the match must be either at the beginning, or after a
1851 newline. The bug was in the code for advancing after a failed match and
1852 checking that the new position followed a newline. It was not taking
1853 account of UTF-8 characters correctly.
1854
1855 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1856 character classes. PCRE was not treating the sequence [:...:] as a
1857 character class unless the ... were all letters. Perl, however, seems to
1858 allow any characters between [: and :], though of course it rejects as
1859 unknown any "names" that contain non-letters, because all the known class
1860 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1861 for example, whereas PCRE did not - it did not recognize a POSIX character
1862 class. This seemed a bit dangerous, so the code has been changed to be
1863 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1864 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1865 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1866 Perl does, and where it didn't before.
1867
1868 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1869 Windows environments %n is disabled by default.
1870
1871
1872 Version 7.4 21-Sep-07
1873 ---------------------
1874
1875 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1876 means that a class such as [\s] counted as "explicit reference to CR or
1877 LF". That isn't really right - the whole point of the change was to try to
1878 help when there was an actual mention of one of the two characters. So now
1879 the change happens only if \r or \n (or a literal CR or LF) character is
1880 encountered.
1881
1882 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1883 of both had grown to the point where there were only 3 bits left.
1884 Fortunately, there was spare space in the data structure, and so I have
1885 moved the internal flags into a new 16-bit field to free up more option
1886 bits.
1887
1888 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1889 but did not set the internal JCHANGED flag - either of these is enough to
1890 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1891 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1892 start sets both bits.
1893
1894 4. Added options (at build time, compile time, exec time) to change \R from
1895 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1896
1897 5. doc/pcresyntax.html was missing from the distribution.
1898
1899 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1900 compatibility, even though it is no longer used.
1901
1902 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1903 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1904 windows.h file is present (where different names are used). [This was
1905 reversed later after testing - see 16 below.]
1906
1907 8. Changed all #include <config.h> to #include "config.h". There were also
1908 some further <pcre.h> cases that I changed to "pcre.h".
1909
1910 9. When pcregrep was used with the --colour option, it missed the line ending
1911 sequence off the lines that it output.
1912
1913 10. It was pointed out to me that arrays of string pointers cause lots of
1914 relocations when a shared library is dynamically loaded. A technique of
1915 using a single long string with a table of offsets can drastically reduce
1916 these. I have refactored PCRE in four places to do this. The result is
1917 dramatic:
1918
1919 Originally: 290
1920 After changing UCP table: 187
1921 After changing error message table: 43
1922 After changing table of "verbs" 36
1923 After changing table of Posix names 22
1924
1925 Thanks to the folks working on Gregex for glib for this insight.
1926
1927 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1928 unicode-properties was also set.
1929
1930 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1931
1932 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1933 checked only for CRLF.
1934
1935 14. Added casts to pcretest.c to avoid compiler warnings.
1936
1937 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1938
1939 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1940 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1941 entirely. This removes changes made in 7 above.
1942
1943 17. The CMake files have been updated, and there is now more information about
1944 building with CMake in the NON-UNIX-USE document.
1945
1946
1947 Version 7.3 28-Aug-07
1948 ---------------------
1949
1950 1. In the rejigging of the build system that eventually resulted in 7.1, the
1951 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1952 brackets there is not right, since it causes compilers to look for an
1953 installed pcre.h, not the version that is in the source that is being
1954 compiled (which of course may be different). I have changed it back to:
1955
1956 #include "pcre.h"
1957
1958 I have a vague recollection that the change was concerned with compiling in
1959 different directories, but in the new build system, that is taken care of
1960 by the VPATH setting the Makefile.
1961
1962 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1963 when the subject happened to end in the byte 0x85 (e.g. if the last
1964 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1965 characters but of course it shouldn't be taken as a newline when it is part
1966 of another character. The bug was that, for an unlimited repeat of . in
1967 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1968 characters when looking for a newline.
1969
1970 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1971
1972 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1973 in debug output.
1974
1975 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1976 long printing in the pcrecpp unittest when running under MinGW.
1977
1978 6. ESC_K was left out of the EBCDIC table.
1979
1980 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1981 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1982 limit also applies to "virtual nesting" when a pattern is recursive, and in
1983 this case 1000 isn't so big. I have been able to remove this limit at the
1984 expense of backing off one optimization in certain circumstances. Normally,
1985 when pcre_exec() would call its internal match() function recursively and
1986 immediately return the result unconditionally, it uses a "tail recursion"
1987 feature to save stack. However, when a subpattern that can match an empty
1988 string has an unlimited repetition quantifier, it no longer makes this
1989 optimization. That gives it a stack frame in which to save the data for
1990 checking that an empty string has been matched. Previously this was taken
1991 from the 1000-entry workspace that had been reserved. So now there is no
1992 explicit limit, but more stack is used.
1993
1994 8. Applied Daniel's patches to solve problems with the import/export magic
1995 syntax that is required for Windows, and which was going wrong for the
1996 pcreposix and pcrecpp parts of the library. These were overlooked when this
1997 problem was solved for the main library.
1998
1999 9. There were some crude static tests to avoid integer overflow when computing
2000 the size of patterns that contain repeated groups with explicit upper
2001 limits. As the maximum quantifier is 65535, the maximum group length was
2002 set at 30,000 so that the product of these two numbers did not overflow a
2003 32-bit integer. However, it turns out that people want to use groups that
2004 are longer than 30,000 bytes (though not repeat them that many times).
2005 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2006 made it possible to implement the integer overflow checks in a much more
2007 dynamic way, which I have now done. The artificial limitation on group
2008 length has been removed - we now have only the limit on the total length of
2009 the compiled pattern, which depends on the LINK_SIZE setting.
2010
2011 10. Fixed a bug in the documentation for get/copy named substring when
2012 duplicate names are permitted. If none of the named substrings are set, the
2013 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2014 empty string.
2015
2016 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2017 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2018 because the ] is interpreted as the first data character and the
2019 terminating ] is not found. PCRE has been made compatible with Perl in this
2020 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2021 cause memory overwriting.
2022
2023 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2024 string has been matched (to stop an infinite loop). It was not recognizing
2025 a conditional subpattern that could match an empty string if that
2026 subpattern was within another subpattern. For example, it looped when
2027 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2028 condition was not nested. This bug has been fixed.
2029
2030 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2031 past the start of the subject in the presence of bytes with the top bit
2032 set, for example "\x8aBCD".
2033
2034 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2035 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2036
2037 14. Optimized (?!) to (*FAIL).
2038
2039 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2040 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2041 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2042 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2043 does: it's just the validity check that is more restrictive.
2044
2045 16. Inserted checks for integer overflows during escape sequence (backslash)
2046 processing, and also fixed erroneous offset values for syntax errors during
2047 backslash processing.
2048
2049 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2050 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2051
2052 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2053 caused an overrun.
2054
2055 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2056 something other than just ASCII characters) inside a group that had an
2057 unlimited repeat caused a loop at compile time (while checking to see
2058 whether the group could match an empty string).
2059
2060 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2061 [\P{Any}] did so. (Error in the code for printing property names.)
2062
2063 21. An orphan \E inside a character class could cause a crash.
2064
2065 22. A repeated capturing bracket such as (A)? could cause a wild memory
2066 reference during compilation.
2067
2068 23. There are several functions in pcre_compile() that scan along a compiled
2069 expression for various reasons (e.g. to see if it's fixed length for look
2070 behind). There were bugs in these functions when a repeated \p or \P was
2071 present in the pattern. These operators have additional parameters compared
2072 with \d, etc, and these were not being taken into account when moving along
2073 the compiled data. Specifically:
2074
2075 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2076 length.
2077
2078 (b) An item such as \pL+ within a repeated group could cause crashes or
2079 loops.
2080
2081 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2082 "reference to non-existent subpattern" error.
2083
2084 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2085
2086 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2087 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2088
2089 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2090
2091 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2092 character were causing crashes (broken optimization).
2093
2094 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2095 \p or \P) caused a compile-time loop.
2096
2097 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2098 break. For example, the unstudied pattern [\r\n]A does not match the string
2099 "\r\nA" because change 7.0/46 below moves the current point on by two
2100 characters after failing to match at the start. However, the pattern \nA
2101 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2102 the same is true. There doesn't seem any very clean way out of this, but
2103 what I have chosen to do makes the common cases work: PCRE now takes note
2104 of whether there can be an explicit match for \r or \n anywhere in the
2105 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2106 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2107 pattern has explicit CR or LF references.
2108
2109 29. Added (*CR) etc for changing newline setting at start of pattern.
2110
2111
2112 Version 7.2 19-Jun-07
2113 ---------------------
2114
2115 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2116 which is apparently normally available under Windows.
2117
2118 2. Re-jig the pcregrep tests with different newline settings in an attempt
2119 to make them independent of the local environment's newline setting.
2120
2121 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2122
2123 4. Some of the "internals" tests were previously cut out when the link size
2124 was not 2, because the output contained actual offsets. The recent new
2125 "Z" feature of pcretest means that these can be cut out, making the tests
2126 usable with all link sizes.
2127
2128 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2129 stack recursion. This gives a massive performance boost under BSD, but just
2130 a small improvement under Linux. However, it saves one field in the frame
2131 in all cases.
2132
2133 6. Added more features from the forthcoming Perl 5.10:
2134
2135 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2136 recursion call. It refers to the nth most recently opened parentheses.
2137
2138 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2139 to be opened parentheses.
2140
2141 (c) Conditions that refer to capturing parentheses can be specified
2142 relatively, for example, (?(-2)... or (?(+3)...
2143
2144 (d) \K resets the start of the current match so that everything before
2145 is not part of it.
2146
2147 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2148
2149 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2150 reference syntax.
2151
2152 (g) (?| introduces a group in which the numbering of parentheses in each
2153 alternative starts with the same number.
2154
2155 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2156
2157 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2158 PCRE_INFO_JCHANGED.
2159
2160 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2161 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2162 for detecting groups that can match an empty string.
2163
2164 9. A pattern with a very large number of alternatives (more than several
2165 hundred) was running out of internal workspace during the pre-compile
2166 phase, where pcre_compile() figures out how much memory will be needed. A
2167 bit of new cunning has reduced the workspace needed for groups with
2168 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2169 workspace instead of running out of the 4096 that are available.
2170
2171 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2172
2173 11. Applied patch from Google to remove an optimization that didn't quite work.
2174 The report of the bug said:
2175
2176 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2177 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2178 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2179
2180 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2181 it matched the wrong number of bytes.
2182
2183
2184 Version 7.1 24-Apr-07
2185 ---------------------
2186
2187 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2188 that is more "standard", making use of automake and other Autotools. There
2189 is some re-arrangement of the files and adjustment of comments consequent
2190 on this.
2191
2192 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2193 for recursive directory scanning broke on some systems because the files
2194 are not scanned in any specific order and on different systems the order
2195 was different. A call to "sort" has been inserted into RunGrepTest for the
2196 approprate test as a short-term fix. In the longer term there may be an
2197 alternative.
2198
2199 3. I had an email from Eric Raymond about problems translating some of PCRE's
2200 man pages to HTML (despite the fact that I distribute HTML pages, some
2201 people do their own conversions for various reasons). The problems
2202 concerned the use of low-level troff macros .br and .in. I have therefore
2203 removed all such uses from the man pages (some were redundant, some could
2204 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2205 HTML has been updated to handle .nf/.fi and to complain if it encounters
2206 .br or .in.
2207
2208 4. Updated comments in configure.ac that get placed in config.h.in and also
2209 arranged for config.h to be included in the distribution, with the name
2210 config.h.generic, for the benefit of those who have to compile without
2211 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2212
2213 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2214 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2215 makevp.bat for the current PCRE, using the additional files
2216 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2217
2218 6. A Windows user reported a minor discrepancy with test 2, which turned out
2219 to be caused by a trailing space on an input line that had got lost in his
2220 copy. The trailing space was an accident, so I've just removed it.
2221
2222 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2223 that is needed.
2224
2225 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2226 as "const" (a) because they are and (b) because it helps the PHP
2227 maintainers who have recently made a script to detect big data structures
2228 in the php code that should be moved to the .rodata section. I remembered
2229 to update Builducptable as well, so it won't revert if ucptable.h is ever
2230 re-created.
2231
2232 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2233 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2234 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2235 support is not required. This saves 1.5-2K of code, which is important in
2236 some applications.
2237
2238 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2239 so as not to refer to the tables, even though these functions will never be
2240 called when UTF-8 support is disabled. Otherwise there are problems with a
2241 shared library.
2242
2243 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2244
2245 (a) It was defining its arguments as char * instead of void *.
2246
2247 (b) It was assuming that all moves were upwards in memory; this was true
2248 a long time ago when I wrote it, but is no longer the case.
2249
2250 The emulated memove() is provided for those environments that have neither
2251 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2252 is clearly not the case, as these two bugs were recently reported.
2253
2254 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2255 and Detrail to create the HTML documentation, the .txt form of the man
2256 pages, and it removes trailing spaces from listed files. It also creates
2257 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2258 case, it wraps all the #defines with #ifndefs. This script should be run
2259 before "make dist".
2260
2261 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2262 with Unicode property support.
2263
2264 (a) For a maximizing quantifier, if the two different cases of the
2265 character were of different lengths in their UTF-8 codings (there are
2266 some cases like this - I found 11), and the matching function had to
2267 back up over a mixture of the two cases, it incorrectly assumed they
2268 were both the same length.
2269
2270 (b) When PCRE was configured to use the heap rather than the stack for
2271 recursion during matching, it was not correctly preserving the data for
2272 the other case of a UTF-8 character when checking ahead for a match
2273 while processing a minimizing repeat. If the check also involved
2274 matching a wide character, but failed, corruption could cause an
2275 erroneous result when trying to check for a repeat of the original
2276 character.
2277
2278 13. Some tidying changes to the testing mechanism:
2279
2280 (a) The RunTest script now detects the internal link size and whether there
2281 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2282 values substituted by "configure". (The RunGrepTest script already did
2283 this for UTF-8.) The configure.ac script no longer substitutes the
2284 relevant variables.
2285
2286 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2287 with length and offset values. This means that the output is different
2288 for different internal link sizes. Test 2 is skipped for link sizes
2289 other than 2 because of this, bypassing the problem. Unfortunately,
2290 there was also a test in test 3 (the locale tests) that used /B and
2291 failed for link sizes other than 2. Rather than cut the whole test out,
2292 I have added a new /Z option to pcretest that replaces the length and
2293 offset values with spaces. This is now used to make test 3 independent
2294 of link size. (Test 2 will be tidied up later.)
2295
2296 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2297 segmentation fault instead of returning the appropriate error message.
2298
2299 15. In multiline mode when the newline sequence was set to "any", the pattern
2300 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2301 This doesn't seem right; it now treats the CRLF combination as the line
2302 ending, and so does not match in that case. It's only a pattern such as ^$
2303 that would hit this one: something like ^ABC$ would have failed after \r
2304 and then tried again after \r\n.
2305
2306 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2307 in an attempt to make files that differ only in their line terminators
2308 compare equal. This works on Linux.
2309
2310 17. Under certain error circumstances pcregrep might try to free random memory
2311 as it exited. This is now fixed, thanks to valgrind.
2312
2313 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2314 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2315 was because its rules for how to advance for /g after matching an empty
2316 string at the end of a line did not allow for this case. They now check for
2317 it specially.
2318
2319 20. pcretest is supposed to handle patterns and data of any length, by
2320 extending its buffers when necessary. It was getting this wrong when the
2321 buffer for a data line had to be extended.
2322
2323 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2324 CRLF as a newline sequence.
2325
2326 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2327 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2328 I have nevertheless tidied it up.
2329
2330 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2331
2332 24. Added a man page for pcre-config.
2333
2334
2335 Version 7.0 19-Dec-06
2336 ---------------------
2337
2338 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2339 moving to gcc 4.1.1.
2340
2341 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2342 sys/time.h, which is documented as needed for this function. It doesn't
2343 seem to matter on Linux, but it showed up on some releases of OS X.
2344
2345 3. It seems that there are systems where bytes whose values are greater than
2346 127 match isprint() in the "C" locale. The "C" locale should be the
2347 default when a C program starts up. In most systems, only ASCII printing
2348 characters match isprint(). This difference caused the output from pcretest
2349 to vary, making some of the tests fail. I have changed pcretest so that:
2350
2351 (a) When it is outputting text in the compiled version of a pattern, bytes
2352 other than 32-126 are always shown as hex escapes.
2353
2354 (b) When it is outputting text that is a matched part of a subject string,
2355 it does the same, unless a different locale has been set for the match
2356 (using the /L modifier). In this case, it uses isprint() to decide.
2357
2358 4. Fixed a major bug that caused incorrect computation of the amount of memory
2359 required for a compiled pattern when options that changed within the
2360 pattern affected the logic of the preliminary scan that determines the
2361 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2362 that the computed length was too small. The symptoms of this bug were
2363 either the PCRE error "internal error: code overflow" from pcre_compile(),
2364 or a glibc crash with a message such as "pcretest: free(): invalid next
2365 size (fast)". Examples of patterns that provoked this bug (shown in
2366 pcretest format) are:
2367
2368 /(?-x: )/x
2369 /(?x)(?-x: \s*#\s*)/
2370 /((?i)[\x{c0}])/8
2371 /(?i:[\x{c0}])/8
2372
2373 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2374 is now done differently.
2375
2376 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2377 wrapper classes; (b) implement a new function in the C++ scanner that is
2378 more efficient than the old way of doing things because it avoids levels of
2379 recursion in the regex matching; (c) add a paragraph to the documentation
2380 for the FullMatch() function.
2381
2382 6. The escape sequence \n was being treated as whatever was defined as
2383 "newline". Not only was this contrary to the documentation, which states
2384 that \n is character 10 (hex 0A), but it also went horribly wrong when
2385 "newline" was defined as CRLF. This has been fixed.
2386
2387 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2388 was being set to -1 for the "end of line" case (supposedly a value that no
2389 character can have). Though this value is never used (the check for end of
2390 line is "zero bytes in current character"), it caused compiler complaints.
2391 I've changed it to 0xffffffff.
2392
2393 8. In pcre_version.c, the version string was being built by a sequence of
2394 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2395 string (as it is for production releases) called a macro with an empty
2396 argument. The C standard says the result of this is undefined. The gcc
2397 compiler treats it as an empty string (which was what was wanted) but it is
2398 reported that Visual C gives an error. The source has been hacked around to
2399 avoid this problem.
2400
2401 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2402 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2403 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2404 of them did).
2405
2406 10. Originally, pcretest opened its input and output without "b"; then I was
2407 told that "b" was needed in some environments, so it was added for release
2408 5.0 to both the input and output. (It makes no difference on Unix-like
2409 systems.) Later I was told that it is wrong for the input on Windows. I've
2410 now abstracted the modes into two macros, to make it easier to fiddle with
2411 them, and removed "b" from the input mode under Windows.
2412
2413 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2414
2415 12. Added -help and --help to pcretest as an official way of being reminded
2416 of the options.
2417
2418 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2419 and pcrecpp.cc because they annoy compilers at high warning levels.
2420
2421 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2422
2423 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2424 scripts are not C programs :-) and which was not noticed because it works
2425 on Linux.
2426
2427 16. pcretest is supposed to handle any length of pattern and data line (as one
2428 line or as a continued sequence of lines) by extending its input buffer if
2429 necessary. This feature was broken for very long pattern lines, leading to
2430 a string of junk being passed to pcre_compile() if the pattern was longer
2431 than about 50K.
2432
2433 17. I have done a major re-factoring of the way pcre_compile() computes the
2434 amount of memory needed for a compiled pattern. Previously, there was code
2435 that made a preliminary scan of the pattern in order to do this. That was
2436 OK when PCRE was new, but as the facilities have expanded, it has become
2437 harder and harder to keep it in step with the real compile phase, and there
2438 have been a number of bugs (see for example, 4 above). I have now found a
2439 cunning way of running the real compile function in a "fake" mode that
2440 enables it to compute how much memory it would need, while actually only
2441 ever using a few hundred bytes of working memory and without too many
2442 tests of the mode. This should make future maintenance and development
2443 easier. A side effect of this work is that the limit of 200 on the nesting
2444 depth of parentheses has been removed (though this was never a serious
2445 limitation, I suspect). However, there is a downside: pcre_compile() now
2446 runs more slowly than before (30% or more, depending on the pattern). I
2447 hope this isn't a big issue. There is no effect on runtime performance.
2448
2449 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2450 newline (only possible for the last line of a file) and it was a
2451 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2452
2453 19. Added additional timing features to pcretest. (1) The -tm option now times
2454 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2455 separate command line item, by a number that specifies the number of
2456 repeats to use when timing. The default is 50000; this gives better
2457 precision, but takes uncomfortably long for very large patterns.
2458
2459 20. Extended pcre_study() to be more clever in cases where a branch of a
2460 subpattern has no definite first character. For example, (a*|b*)[cd] would
2461 previously give no result from pcre_study(). Now it recognizes that the
2462 first character must be a, b, c, or d.
2463
2464 21. There was an incorrect error "recursive call could loop indefinitely" if
2465 a subpattern (or the entire pattern) that was being tested for matching an
2466 empty string contained only one non-empty item after a nested subpattern.
2467 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2468 incorrectly, because the \d was being skipped in the check.
2469
2470 22. The pcretest program now has a new pattern option /B and a command line
2471 option -b, which is equivalent to adding /B to every pattern. This causes
2472 it to show the compiled bytecode, without the additional information that
2473 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2474 is the same as /B/I).
2475
2476 23. A new optimization is now able automatically to treat some sequences such
2477 as a*b as a*+b. More specifically, if something simple (such as a character
2478 or a simple class like \d) has an unlimited quantifier, and is followed by
2479 something that cannot possibly match the quantified thing, the quantifier
2480 is automatically "possessified".
2481
2482 24. A recursive reference to a subpattern whose number was greater than 39
2483 went wrong under certain circumstances in UTF-8 mode. This bug could also
2484 have affected the operation of pcre_study().
2485
2486 25. Realized that a little bit of performance could be had by replacing
2487 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2488
2489 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2490
2491 27. Possessive quantifiers such as a++ were previously implemented by turning
2492 them into atomic groups such as ($>a+). Now they have their own opcodes,
2493 which improves performance. This includes the automatically created ones
2494 from 23 above.
2495
2496 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2497 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2498 the first matched character to be a colon. This applied both to named and
2499 numbered groups.
2500
2501 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2502
2503 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2504 building PCRE on Windows easier, so I have included it in the distribution.
2505
2506 31. There is now a check in pcretest against a ridiculously large number being
2507 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2508 loop, the loop is abandoned.
2509
2510 32. Forward references to subpatterns in conditions such as (?(2)...) where
2511 subpattern 2 is defined later cause pcre_compile() to search forwards in
2512 the pattern for the relevant set of parentheses. This search went wrong
2513 when there were unescaped parentheses in a character class, parentheses
2514 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2515
2516 33. "Subroutine" calls and backreferences were previously restricted to
2517 referencing subpatterns earlier in the regex. This restriction has now
2518 been removed.
2519
2520 34. Added a number of extra features that are going to be in Perl 5.10. On the
2521 whole, these are just syntactic alternatives for features that PCRE had
2522 previously implemented using the Python syntax or my own invention. The
2523 other formats are all retained for compatibility.
2524
2525 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2526 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2527 also .NET compatible.
2528
2529 (b) A recursion or subroutine call to a named group can now be defined as
2530 (?&name) as well as (?P>name).
2531
2532 (c) A backreference to a named group can now be defined as \k<name> or
2533 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2534 5.10, are also .NET compatible.
2535
2536 (d) A conditional reference to a named group can now use the syntax
2537 (?(<name>) or (?('name') as well as (?(name).
2538
2539 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2540 groups (named and numbered) that are never evaluated inline, but can be
2541 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2542 is always false. There may be only one alternative in such a group.
2543
2544 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2545 as the simple (?(R). The condition is true only if the most recent
2546 recursion is that of the given number or name. It does not search out
2547 through the entire recursion stack.
2548
2549 (g) The escape \gN or \g{N} has been added, where N is a positive or
2550 negative number, specifying an absolute or relative reference.
2551
2552 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2553 some "unreachable code" warnings.
2554
2555 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2556 things, this adds five new scripts.
2557
2558 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2559 There were also incompatibilities regarding the handling of \Q..\E inside
2560 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2561 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2562
2563 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2564 matches an empty string, and forcibly breaks the loop. There were bugs in
2565 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2566 against aaaa the result was just "a" rather than "aaaa", for example. Two
2567 separate and independent bugs (that affected different cases) have been
2568 fixed.
2569
2570 39. Refactored the code to abolish the use of different opcodes for small
2571 capturing bracket numbers. This is a tidy that I avoided doing when I
2572 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2573 The new approach is not only tidier, it makes it possible to reduce the
2574 memory needed to fix the previous bug (38).
2575
2576 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2577 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2578 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2579 mode.
2580
2581 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2582 report.
2583
2584 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2585 copy construction and assignment in the C++ wrapper.
2586
2587 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2588 couple of bugs that could have given wrong results in the "--newline=crlf"
2589 case.
2590
2591 44. Added a number of casts and did some reorganization of signed/unsigned int
2592 variables following suggestions from Dair Grant. Also renamed the variable
2593 "this" as "item" because it is a C++ keyword.
2594
2595 45. Arranged for dftables to add
2596
2597 #include "pcre_internal.h"
2598
2599 to pcre_chartables.c because without it, gcc 4.x may remove the array
2600 definition from the final binary if PCRE is built into a static library and
2601 dead code stripping is activated.
2602
2603 46. For an unanchored pattern, if a match attempt fails at the start of a
2604 newline sequence, and the newline setting is CRLF or ANY, and the next two
2605 characters are CRLF, advance by two characters instead of one.
2606
2607
2608 Version 6.7 04-Jul-06
2609 ---------------------
2610
2611 1. In order to handle tests when input lines are enormously long, pcretest has
2612 been re-factored so that it automatically extends its buffers when
2613 necessary. The code is crude, but this _is_ just a test program. The
2614 default size has been increased from 32K to 50K.
2615
2616 2. The code in pcre_study() was using the value of the re argument before
2617 testing it for NULL. (Of course, in any sensible call of the function, it
2618 won't be NULL.)
2619
2620 3. The memmove() emulation function in pcre_internal.h, which is used on
2621 systems that lack both memmove() and bcopy() - that is, hardly ever -
2622 was missing a "static" storage class specifier.
2623
2624 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2625 containing an extended class (one that cannot be represented by a bitmap
2626 because it contains high-valued characters or Unicode property items, e.g.
2627 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2628 pattern, but PCRE should not loop if you do not (it no longer does).
2629 [Detail: two cases were found: (a) a repeated subpattern containing an
2630 extended class; (b) a recursive reference to a subpattern that followed a
2631 previous extended class. It wasn't skipping over the extended class
2632 correctly when UTF-8 mode was not set.]
2633
2634 5. A negated single-character class was not being recognized as fixed-length
2635 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2636 compile error "lookbehind assertion is not fixed length".
2637
2638 6. The RunPerlTest auxiliary script was showing an unexpected difference
2639 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2640 write a Perl script that can interpret lines of an input file either as
2641 byte characters or as UTF-8, which is what "perltest" was being required to
2642 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2643 can't do is switch easily at run time between having the "use utf8;" pragma
2644 or not. In the end, I fudged it by using the RunPerlTest script to insert
2645 "use utf8;" explicitly for the UTF-8 tests.
2646
2647 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2648 the end of the subject string, contrary to the documentation and to what
2649 Perl does. This was true of both matching functions. Now it matches only at
2650 the start of the subject and immediately after *internal* newlines.
2651
2652 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2653 a pointer to an int instead of a pointer to an unsigned long int. This
2654 caused problems on 64-bit systems.
2655
2656 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2657 instance of the 'standard' template library not being so standard".
2658
2659 10. There was no check on the number of named subpatterns nor the maximum
2660 length of a subpattern name. The product of these values is used to compute
2661 the size of the memory block for a compiled pattern. By supplying a very
2662 long subpattern name and a large number of named subpatterns, the size
2663 computation could be caused to overflow. This is now prevented by limiting
2664 the length of names to 32 characters, and the number of named subpatterns
2665 to 10,000.
2666
2667 11. Subpatterns that are repeated with specific counts have to be replicated in
2668 the compiled pattern. The size of memory for this was computed from the
2669 length of the subpattern and the repeat count. The latter is limited to
2670 65535, but there was no limit on the former, meaning that integer overflow
2671 could in principle occur. The compiled length of a repeated subpattern is
2672 now limited to 30,000 bytes in order to prevent this.
2673
2674 12. Added the optional facility to have named substrings with the same name.
2675
2676 13. Added the ability to use a named substring as a condition, using the
2677 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2678 are numbers (not recommended). Forward references are permitted.
2679
2680 14. Added forward references in named backreferences (if you see what I mean).
2681
2682 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2683 pattern could run off the end of the subject. For example, the pattern
2684 "(?s)(.{1,5})"8 did this with the subject "ab".
2685
2686 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2687 PCRE_CASELESS was set when matching characters that were quantified with ?
2688 or *.
2689
2690 17. A character class other than a single negated character that had a minimum
2691 but no maximum quantifier - for example [ab]{6,} - was not handled
2692 correctly by pce_dfa_exec(). It would match only one character.
2693
2694 18. A valid (though odd) pattern that looked like a POSIX character
2695 class but used an invalid character after [ (for example [[,abc,]]) caused
2696 pcre_compile() to give the error "Failed: internal error: code overflow" or
2697 in some cases to crash with a glibc free() error. This could even happen if
2698 the pattern terminated after [[ but there just happened to be a sequence of
2699 letters, a binary zero, and a closing ] in the memory that followed.
2700
2701 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2702 over the years. Originally (before any Unicode support), just the bottom 8
2703 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2704 output from "man perlunicode" includes this:
2705
2706 The regular expression compiler produces polymorphic opcodes. That
2707 is, the pattern adapts to the data and automatically switches to
2708 the Unicode character scheme when presented with Unicode data--or
2709 instead uses a traditional byte scheme when presented with byte
2710 data.
2711
2712 Sadly, a wide octal escape does not cause a switch, and in a string with
2713 no other multibyte characters, these octal escapes are treated as before.
2714 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2715 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2716 Unicode string.
2717
2718 I have not perpetrated such confusion in PCRE. Up till now, it took just
2719 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2720 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2721 translate to the appropriate multibyte character.
2722
2723 29. Applied some refactoring to reduce the number of warnings from Microsoft
2724 and Borland compilers. This has included removing the fudge introduced
2725 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2726 a warning about an unused variable.
2727
2728 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2729 characters since release 4.0, because Perl (from release 5.004) does not.
2730 [Or at least, is documented not to: some releases seem to be in conflict
2731 with the documentation.] However, when a pattern was studied with
2732 pcre_study() and all its branches started with \s, PCRE still included VT
2733 as a possible starting character. Of course, this did no harm; it just
2734 caused an unnecessary match attempt.
2735
2736 22. Removed a now-redundant internal flag bit that recorded the fact that case
2737 dependency changed within the pattern. This was once needed for "required
2738 byte" processing, but is no longer used. This recovers a now-scarce options
2739 bit. Also moved the least significant internal flag bit to the most-
2740 significant bit of the word, which was not previously used (hangover from
2741 the days when it was an int rather than a uint) to free up another bit for
2742 the future.
2743
2744 23. Added support for CRLF line endings as well as CR and LF. As well as the
2745 default being selectable at build time, it can now be changed at runtime
2746 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2747 specify that it is scanning data with non-default line endings.
2748
2749 24. Changed the definition of CXXLINK to make it agree with the definition of
2750 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2751
2752 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2753 recursions. This makes a big different to stack usage for some patterns.
2754
2755 26. If a subpattern containing a named recursion or subroutine reference such
2756 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2757 the space required for the compiled pattern went wrong and gave too small a
2758 value. Depending on the environment, this could lead to "Failed: internal
2759 error: code overflow at offset 49" or "glibc detected double free or
2760 corruption" errors.
2761
2762 27. Applied patches from Google (a) to support the new newline modes and (b) to
2763 advance over multibyte UTF-8 characters in GlobalReplace.
2764
2765 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2766 difference for some implementation of PCRE in some Windows version.
2767
2768 29. Added some extra testing facilities to pcretest:
2769
2770 \q<number> in a data line sets the "match limit" value
2771 \Q<number> in a data line sets the "match recursion limt" value
2772 -S <number> sets the stack size, where <number> is in megabytes
2773
2774 The -S option isn't available for Windows.
2775
2776
2777 Version 6.6 06-Feb-06
2778 ---------------------
2779
2780 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2781 in pcreposix.h. I have copied the definition from pcre.h.
2782
2783 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2784 because pcre.h is no longer a built file.
2785
2786 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2787 not normally included in the compiled code.
2788
2789
2790 Version 6.5 01-Feb-06
2791 ---------------------
2792
2793 1. When using the partial match feature with pcre_dfa_exec(), it was not
2794 anchoring the second and subsequent partial matches at the new starting
2795 point. This could lead to incorrect results. For example, with the pattern
2796 /1234/, partially matching against "123" and then "a4" gave a match.
2797
2798 2. Changes to pcregrep:
2799
2800 (a) All non-match returns from pcre_exec() were being treated as failures
2801 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2802 error message is output. Some extra information is given for the
2803 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2804 probably the only errors that are likely to be caused by users (by
2805 specifying a regex that has nested indefinite repeats, for instance).
2806 If there are more than 20 of these errors, pcregrep is abandoned.
2807
2808 (b) A binary zero was treated as data while matching, but terminated the
2809 output line if it was written out. This has been fixed: binary zeroes
2810 are now no different to any other data bytes.
2811
2812 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2813 used to set a locale for matching. The --locale=xxxx long option has
2814 been added (no short equivalent) to specify a locale explicitly on the
2815 pcregrep command, overriding the environment variables.
2816
2817 (d) When -B was used with -n, some line numbers in the output were one less
2818 than they should have been.
2819
2820 (e) Added the -o (--only-matching) option.
2821
2822 (f) If -A or -C was used with -c (count only), some lines of context were
2823 accidentally printed for the final match.
2824
2825 (g) Added the -H (--with-filename) option.
2826
2827 (h) The combination of options -rh failed to suppress file names for files
2828 that were found from directory arguments.
2829
2830 (i) Added the -D (--devices) and -d (--directories) options.
2831
2832 (j) Added the -F (--fixed-strings) option.
2833
2834 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2835
2836 (l) Added the --colo(u)r option.
2837
2838 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2839 is not present by default.
2840
2841 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2842 items such as (?R) or (?1), when the recursion could match a number of
2843 alternatives. If it matched one of the alternatives, but subsequently,
2844 outside the recursion, there was a failure, the code tried to back up into
2845 the recursion. However, because of the way PCRE is implemented, this is not
2846 possible, and the result was an incorrect result from the match.
2847
2848 In order to prevent this happening, the specification of recursion has
2849 been changed so that all such subpatterns are automatically treated as
2850 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2851
2852 4. I had overlooked the fact that, in some locales, there are characters for
2853 which isalpha() is true but neither isupper() nor islower() are true. In
2854 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2855 and ordfeminine) are like this. This affected the treatment of \w and \W
2856 when they appeared in character classes, but not when they appeared outside
2857 a character class. The bit map for "word" characters is now created
2858 separately from the results of isalnum() instead of just taking it from the
2859 upper, lower, and digit maps. (Plus the underscore character, of course.)
2860
2861 5. The above bug also affected the handling of POSIX character classes such as
2862 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2863 permanent tables. Instead, the bit maps for such a class were previously
2864 created as the appropriate unions of the upper, lower, and digit bitmaps.
2865 Now they are created by subtraction from the [[:word:]] class, which has
2866 its own bitmap.
2867
2868 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2869 It is created by subtracting the vertical space characters (\x09, \x0a,
2870 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2871 subtraction was done in the overall bitmap for a character class, meaning
2872 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2873 be recognized. This bug has been fixed.
2874
2875 7. Patches from the folks at Google:
2876
2877 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2878 real life, but is still worth protecting against".
2879
2880 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2881 regular expressions".
2882
2883 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2884 have it.
2885
2886 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2887 "configure" and the latter not, in order to fix a problem somebody had
2888 with compiling the Arg class on HP-UX.
2889
2890 (e) Improve the error-handling of the C++ wrapper a little bit.
2891
2892 (f) New tests for checking recursion limiting.
2893
2894 8. The pcre_memmove() function, which is used only if the environment does not
2895 have a standard memmove() function (and is therefore rarely compiled),
2896 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2897 returning a result (though PCRE never actually uses the result).
2898
2899 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2900 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2901 returned instead of calling malloc() with an overflowing number that would
2902 most likely cause subsequent chaos.
2903
2904 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2905
2906 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2907 with this option is matched, the nmatch and pmatch options of regexec() are
2908 ignored.
2909
2910 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2911 provided in case anyone wants to the the POSIX interface with UTF-8
2912 strings.
2913
2914 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2915 C++ linking (needed for some HP-UX environments).
2916
2917 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2918 (unused parameter) and in the pcre_printint() function (omitted "default"
2919 switch label when the default is to do nothing).
2920
2921 15. Added some code to make it possible, when PCRE is compiled as a C++
2922 library, to replace subject pointers for pcre_exec() with a smart pointer
2923 class, thus making it possible to process discontinuous strings.
2924
2925 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2926 much the same function. They were added by different people who were trying
2927 to make PCRE easy to compile on non-Unix systems. It has been suggested
2928 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2929 for compiling on Windows systems. I have therefore replaced it with
2930 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2931 defaults to "extern" for C or "extern C" for C++, which works fine on
2932 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2933 SCOPE with something explicit in config.h. In addition:
2934
2935 (a) pcreposix.h still had just "extern" instead of either of these macros;
2936 I have replaced it with PCRE_DATA_SCOPE.
2937
2938 (b) Functions such as _pcre_xclass(), which are internal to the library,
2939 but external in the C sense, all had PCRE_EXPORT in their definitions.
2940 This is apparently wrong for the Windows case, so I have removed it.
2941 (It makes no difference on Unix-like systems.)
2942
2943 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2944 of recursive calls to match(). This is different to MATCH_LIMIT because
2945 that limits the total number of calls to match(), not all of which increase
2946 the depth of recursion. Limiting the recursion depth limits the amount of
2947 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2948 when PCRE is compiled, and changed at run time. A patch from Google adds
2949 this functionality to the C++ interface.
2950
2951 18. Changes to the handling of Unicode character properties:
2952
2953 (a) Updated the table to Unicode 4.1.0.
2954
2955 (b) Recognize characters that are not in the table as "Cn" (undefined).
2956
2957 (c) I revised the way the table is implemented to a much improved format
2958 which includes recognition of ranges. It now supports the ranges that
2959 are defined in UnicodeData.txt, and it also amalgamates other
2960 characters into ranges. This has reduced the number of entries in the
2961 table from around 16,000 to around 3,000, thus reducing its size
2962 considerably. I realized I did not need to use a tree structure after
2963 all - a binary chop search is just as efficient. Having reduced the
2964 number of entries, I extended their size from 6 bytes to 8 bytes to
2965 allow for more data.
2966
2967 (d) Added support for Unicode script names via properties such as \p{Han}.
2968
2969 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2970 matching that character.
2971
2972 20. When matching a repeated Unicode property with a minimum greater than zero,
2973 (for example \pL{2,}), PCRE could look past the end of the subject if it
2974 reached it while seeking the minimum number of characters. This could
2975 happen only if some of the characters were more than one byte long, because
2976 there is a check for at least the minimum number of bytes.
2977
2978 21. Refactored the implementation of \p and \P so as to be more general, to
2979 allow for more different types of property in future. This has changed the
2980 compiled form incompatibly. Anybody with saved compiled patterns that use
2981 \p or \P will have to recompile them.
2982
2983 22. Added "Any" and "L&" to the supported property types.
2984
2985 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2986 but give a compile time error if the value is greater than 0xff.
2987
2988 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2989 accidentally not being installed or uninstalled.
2990
2991 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2992 made were to insert the current release number. This seemed silly, because
2993 it made things harder for people building PCRE on systems that don't run
2994 "configure". I have turned pcre.h into a distributed file, no longer built
2995 by "configure", with the version identification directly included. There is
2996 no longer a pcre.h.in file.
2997
2998 However, this change necessitated a change to the pcre-config script as
2999 well. It is built from pcre-config.in, and one of the substitutions was the
3000 release number. I have updated configure.ac so that ./configure now finds
3001 the release number by grepping pcre.h.
3002
3003 26. Added the ability to run the tests under valgrind.
3004
3005
3006 Version 6.4 05-Sep-05
3007 ---------------------
3008
3009 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3010 "--" to be printed when multiple files were scanned, even when none of the
3011 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3012 consider it to be a bug, and have restored the previous behaviour.
3013
3014 2. A couple of code tidies to get rid of compiler warnings.
3015
3016 3. The pcretest program used to cheat by referring to symbols in the library
3017 whose names begin with _pcre_. These are internal symbols that are not
3018 really supposed to be visible externally, and in some environments it is
3019 possible to suppress them. The cheating is now confined to including
3020 certain files from the library's source, which is a bit cleaner.
3021
3022 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3023 file's purpose clearer.
3024
3025 5. Reorganized pcre_ucp_findchar().
3026
3027
3028 Version 6.3 15-Aug-05
3029 ---------------------
3030
3031 1. The file libpcre.pc.in did not have general read permission in the tarball.
3032
3033 2. There were some problems when building without C++ support:
3034
3035 (a) If C++ support was not built, "make install" and "make test" still
3036 tried to test it.
3037
3038 (b) There were problems when the value of CXX was explicitly set. Some
3039 changes have been made to try to fix these, and ...
3040
3041 (c) --disable-cpp can now be used to explicitly disable C++ support.
3042
3043 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3044 backslash in a target when C++ was disabled. This confuses some
3045 versions of "make", apparently. Using an intermediate variable solves
3046 this. (Same for CPP_LOBJ.)
3047
3048 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3049 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3050 necessary on certain architectures.
3051
3052 4. Added a setting of -export-symbols-regex to the link command to remove
3053 those symbols that are exported in the C sense, but actually are local
3054 within the library, and not documented. Their names all begin with
3055 "_pcre_". This is not a perfect job, because (a) we have to except some
3056 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3057 available (and never for static libraries). I have made a note to try to
3058 find a way round (a) in the future.
3059
3060
3061 Version 6.2 01-Aug-05
3062 ---------------------
3063
3064 1. There was no test for integer overflow of quantifier values. A construction
3065 such as {1111111111111111} would give undefined results. What is worse, if
3066 a minimum quantifier for a parenthesized subpattern overflowed and became
3067 negative, the calculation of the memory size went wrong. This could have
3068 led to memory overwriting.
3069
3070 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3071
3072 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3073 operating environments where this matters.
3074
3075 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3076 PCRE options from within the C++ wrapper.
3077
3078 5. Named capturing subpatterns were not being correctly counted when a pattern
3079 was compiled. This caused two problems: (a) If there were more than 100
3080 such subpatterns, the calculation of the memory needed for the whole
3081 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3082 back references of the form \12, where the number was greater than 9, were
3083 not recognized as back references, even though there were sufficient
3084 previous subpatterns.
3085
3086 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3087 versions of gcc, e.g. 2.95.4.
3088
3089
3090 Version 6.1 21-Jun-05
3091 ---------------------
3092
3093 1. There was one reference to the variable "posix" in pcretest.c that was not
3094 surrounded by "#if !defined NOPOSIX".
3095
3096 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3097 the cross-check on the old pcre_info() function, for the benefit of the
3098 cut-down version of PCRE that is currently imported into Exim.
3099
3100 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3101 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3102 patterns that start (?i) but I don't think that matters. The use of (?i) is
3103 just an example; this all applies to the other options as well.
3104
3105 4. Since libtool seems to echo the compile commands it is issuing, the output
3106 from "make" can be reduced a bit by putting "@" in front of each libtool
3107 compile command.
3108
3109 5. Patch from the folks at Google for configure.in to be a bit more thorough
3110 in checking for a suitable C++ installation before trying to compile the
3111 C++ stuff. This should fix a reported problem when a compiler was present,
3112 but no suitable headers.
3113
3114 6. The man pages all had just "PCRE" as their title. I have changed them to
3115 be the relevant file name. I have also arranged that these names are
3116 retained in the file doc/pcre.txt, which is a concatenation in text format
3117 of all the man pages except the little individual ones for each function.
3118
3119 7. The NON-UNIX-USE file had not been updated for the different set of source
3120 files that come with release 6. I also added a few comments about the C++
3121 wrapper.
3122
3123
3124 Version 6.0 07-Jun-05
3125 ---------------------
3126
3127 1. Some minor internal re-organization to help with my DFA experiments.
3128
3129 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3130 didn't matter for the library itself when fully configured, but did matter
3131 when compiling without UCP support, or within Exim, where the ucp files are
3132 not imported.
3133
3134 3. Refactoring of the library code to split up the various functions into
3135 different source modules. The addition of the new DFA matching code (see
3136 below) to a single monolithic source would have made it really too
3137 unwieldy, quite apart from causing all the code to be include in a
3138 statically linked application, when only some functions are used. This is
3139 relevant even without the DFA addition now that patterns can be compiled in
3140 one application and matched in another.
3141
3142 The downside of splitting up is that there have to be some external
3143 functions and data tables that are used internally in different modules of
3144 the library but which are not part of the API. These have all had their
3145 names changed to start with "_pcre_" so that they are unlikely to clash
3146 with other external names.
3147
3148 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3149 a different (DFA) algorithm. Although it is slower than the original
3150 function, it does have some advantages for certain types of matching
3151 problem.
3152
3153 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3154 including restarting after a partial match.
3155
3156 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3157 defined when compiling for Windows was sent to me. I have put it into the
3158 code, though I have no means of testing or verifying it.
3159
3160 7. Added the pcre_refcount() auxiliary function.
3161
3162 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3163 match before or at the first newline in the subject string. In pcretest,
3164 the /f option on a pattern can be used to set this.
3165
3166 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3167 would behave wrongly. This has been present in PCRE since release 4.0.
3168
3169 10. A number of changes to the pcregrep command:
3170
3171 (a) Refactored how -x works; insert ^(...)$ instead of setting
3172 PCRE_ANCHORED and checking the length, in preparation for adding
3173 something similar for -w.
3174
3175 (b) Added the -w (match as a word) option.
3176
3177 (c) Refactored the way lines are read and buffered so as to have more
3178 than one at a time available.
3179
3180 (d) Implemented a pcregrep test script.
3181
3182 (e) Added the -M (multiline match) option. This allows patterns to match
3183 over several lines of the subject. The buffering ensures that at least
3184 8K, or the rest of the document (whichever is the shorter) is available
3185 for matching (and similarly the previous 8K for lookbehind assertions).
3186
3187 (f) Changed the --help output so that it now says
3188
3189 -w, --word-regex(p)
3190
3191 instead of two lines, one with "regex" and the other with "regexp"
3192 because that confused at least one person since the short forms are the
3193 same. (This required a bit of code, as the output is generated
3194 automatically from a table. It wasn't just a text change.)
3195
3196 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3197 option but starts with a hyphen. Could be a pattern or a path name
3198 starting with a hyphen, for instance.
3199
3200 (h) "-" can be given as a file name to represent stdin.
3201
3202 (i) When file names are being printed, "(standard input)" is used for
3203 the standard input, for compatibility with GNU grep. Previously
3204 "<stdin>" was used.
3205
3206 (j) The option --label=xxx can be used to supply a name to be used for
3207 stdin when file names are being printed. There is no short form.
3208
3209 (k) Re-factored the options decoding logic because we are going to add
3210 two more options that take data. Such options can now be given in four
3211 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3212
3213 (l) Added the -A, -B, and -C options for requesting that lines of context
3214 around matches be printed.
3215
3216 (m) Added the -L option to print the names of files that do not contain
3217 any matching lines, that is, the complement of -l.
3218
3219 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3220 continue to scan other files.
3221
3222 (o) The -s option was incorrectly implemented. For compatibility with other
3223 greps, it now suppresses the error message for a non-existent or non-
3224 accessible file (but not the return code). There is a new option called
3225 -q that suppresses the output of matching lines, which was what -s was
3226 previously doing.
3227
3228 (p) Added --include and --exclude options to specify files for inclusion
3229 and exclusion when recursing.
3230
3231 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3232 Hopefully, it now does.
3233
3234 12. Missing cast in pcre_study().
3235
3236 13. Added an "uninstall" target to the makefile.
3237
3238 14. Replaced "extern" in the function prototypes in Makefile.in with
3239 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3240 world, but is set differently for Windows.
3241
3242 15. Added a second compiling function called pcre_compile2(). The only
3243 difference is that it has an extra argument, which is a pointer to an
3244 integer error code. When there is a compile-time failure, this is set
3245 non-zero, in addition to the error test pointer being set to point to an
3246 error message. The new argument may be NULL if no error number is required
3247 (but then you may as well call pcre_compile(), which is now just a
3248 wrapper). This facility is provided because some applications need a
3249 numeric error indication, but it has also enabled me to tidy up the way
3250 compile-time errors are handled in the POSIX wrapper.
3251
3252 16. Added VPATH=.libs to the makefile; this should help when building with one
3253 prefix path and installing with another. (Or so I'm told by someone who
3254 knows more about this stuff than I do.)
3255
3256 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3257 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3258 match everything, including newlines. This is not POSIX-compatible, but
3259 somebody wanted the feature. From pcretest it can be activated by using
3260 both the P and the s flags.
3261
3262 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3263
3264 19. libpcre.pc was being incorrectly installed as executable.
3265
3266 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3267 it now also looks for '\r' so that it will work unmodified on Windows.
3268
3269 21. Added Google's contributed C++ wrapper to the distribution.
3270
3271 22. Added some untidy missing memory free() calls in pcretest, to keep
3272 Electric Fence happy when testing.
3273
3274
3275
3276 Version 5.0 13-Sep-04
3277 ---------------------
3278
3279 1. Internal change: literal characters are no longer packed up into items
3280 containing multiple characters in a single byte-string. Each character
3281 is now matched using a separate opcode. However, there may be more than one
3282 byte in the character in UTF-8 mode.
3283
3284 2. The pcre_callout_block structure has two new fields: pattern_position and
3285 next_item_length. These contain the offset in the pattern to the next match
3286 item, and its length, respectively.
3287
3288 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3289 insertion of callouts before each pattern item. Added the /C option to
3290 pcretest to make use of this.
3291
3292 4. On the advice of a Windows user, the lines
3293
3294 #if defined(_WIN32) || defined(WIN32)
3295 _setmode( _fileno( stdout ), 0x8000 );
3296 #endif /* defined(_WIN32) || defined(WIN32) */
3297
3298 have been added to the source of pcretest. This apparently does useful
3299 magic in relation to line terminators.
3300
3301 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3302 for the benefit of those environments where the "b" makes a difference.
3303
3304 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3305 to know about it. I have put a hack into configure.in that adds in code
3306 to set GCC=yes if CC=icc. This seems to end up at a point in the
3307 generated configure script that is early enough to affect the setting of
3308 compiler options, which is what is needed, but I have no means of testing
3309 whether it really works. (The user who reported this had patched the
3310 generated configure script, which of course I cannot do.)
3311
3312 LATER: After change 22 below (new libtool files), the configure script
3313 seems to know about icc (and also ecc). Therefore, I have commented out
3314 this hack in configure.in.
3315
3316 7. Added support for pkg-config (2 patches were sent in).
3317
3318 8. Negated POSIX character classes that used a combination of internal tables
3319 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3320 [[:^ascii]]. Typically, they would match almost any characters. The other
3321 POSIX classes were not broken in this way.
3322
3323 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3324 to find the match, as PCRE was deluded into thinking that the match had to
3325 start at the start point or following a newline. The same bug applied to
3326 patterns with negative forward assertions or any backward assertions
3327 preceding ".*" at the start, unless the pattern required a fixed first
3328 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3329
3330 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3331 starting at the last subject character, bytes beyond the end of the subject
3332 string were read.
3333
3334 11. Renamed the variable "class" as "classbits" to make life easier for C++
3335 users. (Previously there was a macro definition, but it apparently wasn't
3336 enough.)
3337
3338 12. Added the new field "tables" to the extra data so that tables can be passed
3339 in at exec time, or the internal tables can be re-selected. This allows
3340 a compiled regex to be saved and re-used at a later time by a different
3341 program that might have everything at different addresses.
3342
3343 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3344 -R library as well as a -L library.
3345
3346 14. The debugging options of pcretest (-d on the command line or D on a
3347 pattern) showed incorrect output for anything following an extended class
3348 that contained multibyte characters and which was followed by a quantifier.
3349
3350 15. Added optional support for general category Unicode character properties
3351 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3352 support. It adds about 90K to the size of the library. The meanings of the
3353 inbuilt class escapes such as \d and \s have NOT been changed.
3354
3355 16. Updated pcredemo.c to include calls to free() to release the memory for the
3356 compiled pattern.
3357
3358 17. The generated file chartables.c was being created in the source directory
3359 instead of in the building directory. This caused the build to fail if the
3360 source directory was different from the building directory, and was
3361 read-only.
3362
3363 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3364 file. No doubt somebody will tell me if they don't make sense... Also added
3365 Dan Mooney's comments about building on OpenVMS.
3366
3367 19. Added support for partial matching via the PCRE_PARTIAL option for
3368 pcre_exec() and the \P data escape in pcretest.
3369
3370 20. Extended pcretest with 3 new pattern features:
3371
3372 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3373 write the compiled pattern to the file whose name is "rest-of-line".
3374 This is a straight binary dump of the data, with the saved pointer to
3375 the character tables forced to be NULL. The study data, if any, is
3376 written too. After writing, pcretest reads a new pattern.
3377
3378 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3379 compiled pattern from the given file. There must not be any
3380 occurrences of "<" in the file name (pretty unlikely); if there are,
3381 pcretest will instead treat the initial "<" as a pattern delimiter.
3382 After reading in the pattern, pcretest goes on to read data lines as
3383 usual.
3384
3385 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3386 and 16-bit fields in a compiled pattern, to simulate a pattern that
3387 was compiled on a host of opposite endianness.
3388
3389 21. The pcre-exec() function can now cope with patterns that were compiled on
3390 hosts of opposite endianness, with this restriction:
3391
3392 As for any compiled expression that is saved and used later, the tables
3393 pointer field cannot be preserved; the extra_data field in the arguments
3394 to pcre_exec() should be used to pass in a tables address if a value
3395 other than the default internal tables were used at compile time.
3396
3397 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3398 now diagnosed as an error. Previously, most of the time, a negative number
3399 would have been treated as zero, but if in addition "ovector" was passed as
3400 NULL, a crash could occur.
3401
3402 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3403 new versions from the libtool 1.5 distribution (the last one is a copy of
3404 a file called libtool.m4). This seems to have fixed the need to patch
3405 "configure" to support Darwin 1.3 (which I used to do). However, I still
3406 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3407 workstation).
3408
3409 24. Changed the PCRE licence to be the more standard "BSD" licence.
3410
3411
3412 Version 4.5 01-Dec-03
3413 ---------------------
3414
3415 1. There has been some re-arrangement of the code for the match() function so
3416 that it can be compiled in a version that does not call itself recursively.
3417 Instead, it keeps those local variables that need separate instances for
3418 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3419 needs to "recurse". Keeping track of where control must go is done by means
3420 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3421 hide most of the details from the main code, and operates only if
3422 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3423 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3424 operating.
3425
3426 To make it easier for callers to provide specially tailored get/free
3427 functions for this usage, two new functions, pcre_stack_malloc, and
3428 pcre_stack_free, are used. They are always called in strict stacking order,
3429 and the size of block requested is always the same.
3430
3431 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3432 PCRE has been compiled to use the stack or the heap for recursion. The
3433 -C option of pcretest uses this to show which version is compiled.
3434
3435 A new data escape \S, is added to pcretest; it causes the amounts of store
3436 obtained and freed by both kinds of malloc/free at match time to be added
3437 to the output.
3438
3439 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3440 what's available on my current Linux desktop machine.
3441
3442 3. When matching a UTF-8 string, the test for a valid string at the start has
3443 been extended. If start_offset is not zero, PCRE now checks that it points
3444 to a byte that is the start of a UTF-8 character. If not, it returns
3445 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3446 this is necessary because there may be backward assertions in the pattern.
3447 When matching the same subject several times, it may save resources to use
3448 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3449
3450 4. The code for checking the validity of UTF-8 strings has been tightened so
3451 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3452 containing "overlong sequences".
3453
3454 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3455 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3456 should have been "&", but it just so happened that all the cases this let
3457 through by mistake were picked up later in the function.
3458
3459 6. I had used a variable called "isblank" - this is a C99 function, causing
3460 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3461
3462 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3463 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3464 so the version is shown; (c) stop "make test" echoing "./RunTest".
3465
3466 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3467
3468 9. The prototype for memmove() for systems that don't have it was using
3469 size_t, but the inclusion of the header that defines size_t was later. I've
3470 moved the #includes for the C headers earlier to avoid this.
3471
3472 10. Added some adjustments to the code to make it easier to compiler on certain
3473 special systems:
3474
3475 (a) Some "const" qualifiers were missing.
3476 (b) Added the macro EXPORT before all exported functions; by default this
3477 is defined to be empty.
3478 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3479 that it reads its output file name as an argument instead of writing
3480 to the standard output and assuming this can be redirected.
3481
3482 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3483 class containing characters with values greater than 255, PCRE compilation
3484 went into a loop.
3485
3486 12. A recursive reference to a subpattern that was within another subpattern
3487 that had a minimum quantifier of zero caused PCRE to crash. For example,
3488 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3489 recursion. If the recursively-called subpattern itself had a zero repeat,
3490 that was OK.
3491
3492 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3493 buffer into which it was copied (for escape processing) was still set at
3494 1024, so long lines caused crashes.
3495
3496 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3497 "internal error: code overflow...". This applied to any character class
3498 that was followed by a possessive quantifier.
3499
3500 15. Modified the Makefile to add libpcre.la as a prerequisite for
3501 libpcreposix.la because I was told this is needed for a parallel build to
3502 work.
3503
3504 16. If a pattern that contained .* following optional items at the start was
3505 studied, the wrong optimizing data was generated, leading to matching
3506 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3507 matching string must start with a or b or c. The correct conclusion for
3508 this pattern is that a match can start with any character.
3509
3510
3511 Version 4.4 13-Aug-03
3512 ---------------------
3513
3514 1. In UTF-8 mode, a character class containing characters with values between
3515 127 and 255 was not handled correctly if the compiled pattern was studied.
3516 In fixing this, I have also improved the studying algorithm for such
3517 classes (slightly).
3518
3519 2. Three internal functions had redundant arguments passed to them. Removal
3520 might give a very teeny performance improvement.
3521
3522 3. Documentation bug: the value of the capture_top field in a callout is *one
3523 more than* the number of the hightest numbered captured substring.
3524
3525 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3526 in incorrectly linking with a previously installed version. They now link
3527 explicitly with libpcre.la.
3528
3529 5. configure.in no longer needs to recognize Cygwin specially.
3530
3531 6. A problem in pcre.in for Windows platforms is fixed.
3532
3533 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3534 pcretest, it used to include the size of the study block as part of its
3535 output. Unfortunately, the structure contains a field that has a different
3536 size on different hardware architectures. This meant that the tests that
3537 showed this size failed. As the block is currently always of a fixed size,
3538 this information isn't actually particularly useful in pcretest output, so
3539 I have just removed it.
3540
3541 8. Three pre-processor statements accidentally did not start in column 1.
3542 Sadly, there are *still* compilers around that complain, even though
3543 standard C has not required this for well over a decade. Sigh.
3544
3545 9. In pcretest, the code for checking callouts passed small integers in the
3546 callout_data field, which is a void * field. However, some picky compilers
3547 complained about the casts involved for this on 64-bit systems. Now
3548 pcretest passes the address of the small integer instead, which should get
3549 rid of the warnings.
3550
3551 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3552 both compile and run time, and gives an error if an invalid UTF-8 sequence
3553 is found. There is a option for disabling this check in cases where the
3554 string is known to be correct and/or the maximum performance is wanted.
3555
3556 11. In response to a bug report, I changed one line in Makefile.in from
3557
3558 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3559 to
3560 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3561
3562 to look similar to other lines, but I have no way of telling whether this
3563 is the right thing to do, as I do not use Windows. No doubt I'll get told
3564 if it's wrong...
3565
3566
3567 Version 4.3 21-May-03
3568 ---------------------
3569
3570 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3571 Makefile.
3572
3573 2. Some refactoring to improve the quality of the code:
3574
3575 (i) The utf8_table... variables are now declared "const".
3576
3577 (ii) The code for \cx, which used the "case flipping" table to upper case
3578 lower case letters, now just substracts 32. This is ASCII-specific,
3579 but the whole concept of \cx is ASCII-specific, so it seems
3580 reasonable.
3581
3582 (iii) PCRE was using its character types table to recognize decimal and
3583 hexadecimal digits in the pattern. This is silly, because it handles
3584 only 0-9, a-f, and A-F, but the character types table is locale-
3585 specific, which means strange things might happen. A private
3586 table is now used for this - though it costs 256 bytes, a table is
3587 much faster than multiple explicit tests. Of course, the standard
3588 character types table is still used for matching digits in subject
3589 strings against \d.
3590
3591 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3592 ending in _t are). So I've renamed it as ESC_tee.
3593
3594 3. The first argument for regexec() in the POSIX wrapper should have been
3595 defined as "const".
3596
3597 4. Changed pcretest to use malloc() for its buffers so that they can be
3598 Electric Fenced for debugging.
3599
3600 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3601 to read one or more bytes before the start of the subject string. Often this
3602 had no effect on PCRE's behaviour, but in some circumstances it could
3603 provoke a segmentation fault.
3604
3605 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3606 to try to read one or more bytes before the start of the subject string.
3607
3608 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3609 UTF-8 support could misbehave in various ways if the subject string
3610 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3611 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3612 back over UTF-8 characters.)
3613
3614
3615 Version 4.2 14-Apr-03
3616 ---------------------
3617
3618 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3619
3620 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3621 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3622 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3623 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3624 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3625 and BUILD_EXEEXT
3626 Note: automatic setting of the BUILD variables is not yet working
3627 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3628 compile-time but not at link-time
3629 [LINK]: use for linking executables only
3630 make different versions for Windows and non-Windows
3631 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3632 libraries
3633 [LINK_FOR_BUILD]: new variable
3634 [OBJEXT]: use throughout
3635 [EXEEXT]: use throughout
3636 <winshared>: new target
3637 <wininstall>: new target
3638 <dftables.o>: use native compiler
3639 <dftables>: use native linker
3640 <install>: handle Windows platform correctly
3641 <clean>: ditto
3642 <check>: ditto
3643 copy DLL to top builddir before testing
3644
3645 As part of these changes, -no-undefined was removed again. This was reported
3646 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3647 in any case.
3648
3649 3. Some tidies to get rid of compiler warnings:
3650
3651 . In the match_data structure, match_limit was an unsigned long int, whereas
3652 match_call_count was an int. I've made them both unsigned long ints.
3653
3654 . In pcretest the fact that a const uschar * doesn't automatically cast to
3655 a void * provoked a warning.
3656
3657 . Turning on some more compiler warnings threw up some "shadow" variables
3658 and a few more missing casts.
3659
3660 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3661 option, a class that contained a single character with a value between 128
3662 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3663
3664 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3665 option, a class that contained several characters, but with at least one
3666 whose value was between 128 and 255 caused PCRE to crash.
3667
3668
3669 Version 4.1 12-Mar-03
3670 ---------------------
3671
3672 1. Compiling with gcc -pedantic found a couple of places where casts were
3673 needed, and a string in dftables.c that was longer than standard compilers are
3674 required to support.
3675
3676 2. Compiling with Sun's compiler found a few more places where the code could
3677 be tidied up in order to avoid warnings.
3678
3679 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3680 first of these names is deprecated in the latest Autoconf in favour of the name
3681 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3682 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3683 analogy I have changed it to CFLAGS_FOR_BUILD.
3684
3685 4. Added -no-undefined to the linking command in the Makefile, because this is
3686 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3687 linking step for the pcreposix library.
3688
3689 5. PCRE was failing to diagnose the case of two named groups with the same
3690 name.
3691
3692 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3693 literal character that is needed in the subject for a match, and scans along to
3694 ensure that it is present before embarking on the full matching process. This
3695 saves time in cases of nested unlimited repeats that are never going to match.
3696 Problem: the scan can take a lot of time if the subject is very long (e.g.
3697 megabytes), thus penalizing straightforward matches. It is now done only if the
3698 amount of subject to be scanned is less than 1000 bytes.
3699
3700 7. A lesser problem with the same optimization is that it was recording the
3701 first character of an anchored pattern as "needed", thus provoking a search
3702 right along the subject, even when the first match of the pattern was going to
3703 fail. The "needed" character is now not set for anchored patterns, unless it
3704 follows something in the pattern that is of non-fixed length. Thus, it still
3705 fulfils its original purpose of finding quick non-matches in cases of nested
3706 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3707
3708
3709 Version 4.0 17-Feb-03
3710 ---------------------
3711
3712 1. If a comment in an extended regex that started immediately after a meta-item
3713 extended to the end of string, PCRE compiled incorrect data. This could lead to
3714 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3715
3716 2. Moved to autoconf 2.53 and libtool 1.4.2.
3717
3718 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3719 the special perltest8 script is no longer needed - all the tests can be run
3720 from a single perltest script.
3721
3722 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3723 by \s. It has now been removed in PCRE. This means it isn't recognized as
3724 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3725 class [:space:] *does* include VT, thereby creating a mess.
3726
3727 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3728 space and tab.
3729
3730 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3731 its new features into the main test script, reducing the number of scripts.
3732
3733 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3734 were backward compatible, and made the (?i) apply to the whole pattern, as if
3735 /i were given. Now it behaves more logically, and applies the option setting
3736 only to what follows. PCRE has been changed to follow suit. However, if it
3737 finds options settings right at the start of the pattern, it extracts them into
3738 the global options, as before. Thus, they show up in the info data.
3739
3740 8. Added support for the \Q...\E escape sequence. Characters in between are
3741 treated as literals. This is slightly different from Perl in that $ and @ are
3742 also handled as literals inside the quotes. In Perl, they will cause variable
3743 interpolation. Note the following examples:
3744
3745 Pattern PCRE matches Perl matches
3746
3747 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3748 \Qabc\$xyz\E abc\$xyz abc\$xyz
3749 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3750
3751 For compatibility with Perl, \Q...\E sequences are recognized inside character
3752 classes as well as outside them.
3753
3754 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3755 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3756 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3757 signed/unsigned warnings.
3758
3759 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3760 option for pcretest, so I've replaced it by a simple function that does just
3761 that job.
3762
3763 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3764 "pcregrep -".
3765
3766 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3767 Java package. This provides some syntactic sugar for simple cases of what my
3768 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3769 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3770 item, you can use this simplified notation. Note that only makes sense with
3771 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3772 greediness, whatever the setting of the PCRE_UNGREEDY option.
3773
3774 13. A change of greediness default within a pattern was not taking effect at
3775 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3776 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3777 was abstracted outside.
3778
3779 14. PCRE now supports the \G assertion. It is true when the current matching
3780 position is at the start point of the match. This differs from \A when the
3781 starting offset is non-zero. Used with the /g option of pcretest (or similar
3782 code), it works in the same way as it does for Perl's /g option. If all
3783 alternatives of a regex begin with \G, the expression is anchored to the start
3784 match position, and the "anchored" flag is set in the compiled expression.
3785
3786 15. Some bugs concerning the handling of certain option changes within patterns
3787 have been fixed. These applied to options other than (?ims). For example,
3788 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3789 been the other way round. Some of this was related to change 7 above.
3790
3791 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3792 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3793 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3794 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3795
3796 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3797 mode. Unlike ".", it always matches newline, whatever the setting of
3798 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3799 assertions. Perl allows it, but it doesn't (in general) work because it can't
3800 calculate the length of the lookbehind. At least, that's the case for Perl
3801 5.8.0 - I've been told they are going to document that it doesn't work in
3802 future.
3803
3804 18. Added an error diagnosis for escapes that PCRE does not support: these are
3805 \L, \l, \N, \P, \p, \U, \u, and \X.
3806
3807 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3808 reading past the end of the pattern in cases such as /[abcd/.
3809
3810 20. PCRE was getting more memory than necessary for patterns with classes that
3811 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3812
3813 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3814 compiling PCRE for use with Virtual Pascal.
3815
3816 22. Small fix to the Makefile to make it work properly if the build is done
3817 outside the source tree.
3818
3819 23. Added a new extension: a condition to go with recursion. If a conditional
3820 subpattern starts with (?(R) the "true" branch is used if recursion has
3821 happened, whereas the "false" branch is used only at the top level.
3822
3823 24. When there was a very long string of literal characters (over 255 bytes
3824 without UTF support, over 250 bytes with UTF support), the computation of how
3825 much memory was required could be incorrect, leading to segfaults or other
3826 strange effects.
3827
3828 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3829 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3830 there was a subsequent back reference to those brackets. This meant that, for
3831 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3832 possible to check for precisely this case. All we can do is abandon the
3833 optimization if .* occurs inside capturing brackets when there are any back
3834 references whatsoever. (See below for a better fix that came later.)
3835
3836 26. The handling of the optimization for finding the first character of a
3837 non-anchored pattern, and for finding a character that is required later in the
3838 match were failing in some cases. This didn't break the matching; it just
3839 failed to optimize when it could. The way this is done has been re-implemented.
3840
3841 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3842
3843 28. Added a new feature that provides some of the functionality that Perl
3844 provides with (?{...}). The facility is termed a "callout". The way it is done
3845 in PCRE is for the caller to provide an optional function, by setting
3846 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3847 global variable. By default it is unset, which disables all calling out. To get
3848 the function called, the regex must include (?C) at appropriate points. This
3849 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3850 This provides a means of identifying different callout points. When PCRE
3851 reaches such a point in the regex, if pcre_callout has been set, the external
3852 function is called. It is provided with data in a structure called
3853 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3854 matching continues; if it returns a non-zero value, the match at the current
3855 point fails. However, backtracking will occur if possible. [This was changed
3856 later and other features added - see item 49 below.]
3857
3858 29. pcretest is upgraded to test the callout functionality. It provides a
3859 callout function that displays information. By default, it shows the start of
3860 the match and the current position in the text. There are some new data escapes
3861 to vary what happens:
3862
3863 \C+ in addition, show current contents of captured substrings
3864 \C- do not supply a callout function
3865 \C!n return 1 when callout number n is reached
3866 \C!n!m return 1 when callout number n is reached for the mth time
3867
3868 30. If pcregrep was called with the -l option and just a single file name, it
3869 output "<stdin>" if a match was found, instead of the file name.
3870
3871 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3872 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3873 pcre_exec(). This saves a malloc/free per call. The default value of
3874 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3875 when configuring.
3876
3877 32. The default maximum size of a compiled pattern is 64K. There have been a
3878 few cases of people hitting this limit. The code now uses macros to handle the
3879 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3880 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3881 configuring. Tests 2 and 5 work only with 2-byte links because they output
3882 debugging information about compiled patterns.
3883
3884 33. Internal code re-arrangements:
3885
3886 (a) Moved the debugging function for printing out a compiled regex into
3887 its own source file (printint.c) and used #include to pull it into
3888 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3889 separate copies.
3890
3891 (b) Defined the list of op-code names for debugging as a macro in
3892 internal.h so that it is next to the definition of the opcodes.
3893
3894 (c) Defined a table of op-code lengths for simpler skipping along compiled
3895 code. This is again a macro in internal.h so that it is next to the
3896 definition of the opcodes.
3897
3898 34. Added support for recursive calls to individual subpatterns, along the
3899 lines of Robin Houston's patch (but implemented somewhat differently).
3900
3901 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3902 allow it to read and process whole directories in Win32. This code was
3903 contributed by Lionel Fourquaux; it has not been tested by me.
3904
3905 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3906 used to name a group. Names consist of alphanumerics and underscores, and must
3907 be unique. Back references use the syntax (?P=name) and recursive calls use
3908 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3909 numbers. The function pcre_fullinfo() can be used after compilation to extract
3910 a name/number map. There are three relevant calls:
3911
3912 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3913 PCRE_INFO_NAMECOUNT yields the number of entries
3914 PCRE_INFO_NAMETABLE yields a pointer to the map.
3915
3916 The map is a vector of fixed-size entries. The size of each entry depends on
3917 the length of the longest name used. The first two bytes of each entry are the
3918 group number, most significant byte first. There follows the corresponding
3919 name, zero terminated. The names are in alphabetical order.
3920
3921 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3922 case instead of 255. Making it the same both with and without UTF-8 support
3923 means that the same test output works with both.
3924
3925 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3926 calling malloc() with a zero argument.
3927
3928 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3929 optimization. I've improved things by keeping a bitmap of backreferences with
3930 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3931 fact referenced, the optimization can be applied. It is unlikely that a
3932 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3933 the match to follow \n) will appear inside brackets with a number greater than
3934 31, but if it does, any back reference > 31 suppresses the optimization.
3935
3936 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3937 of disabling numbered capturing parentheses. Any opening parenthesis that is
3938 not followed by ? behaves as if it were followed by ?: but named parentheses
3939 can still be used for capturing (and they will acquire numbers in the usual
3940 way).
3941
3942 41. Redesigned the return codes from the match() function into yes/no/error so
3943 that errors can be passed back from deep inside the nested calls. A malloc
3944 failure while inside a recursive subpattern call now causes the
3945 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3946
3947 42. It is now possible to set a limit on the number of times the match()
3948 function is called in a call to pcre_exec(). This facility makes it possible to
3949 limit the amount of recursion and backtracking, though not in a directly
3950 obvious way, because the match() function is used in a number of different
3951 circumstances. The count starts from zero for each position in the subject
3952 string (for non-anchored patterns). The default limit is, for compatibility, a
3953 large number, namely 10 000 000. You can change this in two ways:
3954
3955 (a) When configuring PCRE before making, you can use --with-match-limit=n
3956 to set a default value for the compiled library.
3957
3958 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3959 a different value is set. See 45 below.
3960
3961 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3962
3963 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3964 of things that can be changed at compile time. The first argument specifies
3965 what is wanted and the second points to where the information is to be placed.
3966 The current list of available information is:
3967
3968 PCRE_CONFIG_UTF8
3969
3970 The output is an integer that is set to one if UTF-8 support is available;
3971 otherwise it is set to zero.
3972
3973 PCRE_CONFIG_NEWLINE
3974
3975 The output is an integer that it set to the value of the code that is used for
3976 newline. It is either LF (10) or CR (13).
3977
3978 PCRE_CONFIG_LINK_SIZE
3979
3980 The output is an integer that contains the number of bytes used for internal
3981 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3982
3983 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3984
3985 The output is an integer that contains the threshold above which the POSIX
3986 interface uses malloc() for output vectors. See item 31 above.
3987
3988 PCRE_CONFIG_MATCH_LIMIT
3989
3990 The output is an unsigned integer that contains the default limit of the number
3991 of match() calls in a pcre_exec() execution. See 42 above.
3992
3993 44. pcretest has been upgraded by the addition of the -C option. This causes it
3994 to extract all the available output from the new pcre_config() function, and to
3995 output it. The program then exits immediately.
3996
3997 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3998 order to support additional features. One way would have been to define
3999 pcre_exec2() (for example) with extra arguments, but this would not have been
4000 extensible, and would also have required all calls to the original function to
4001 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4002 is used for passing in "extra" data from pcre_study().
4003
4004 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4005 contains the following fields:
4006
4007 flags a bitmap indicating which of the following fields are set
4008 study_data opaque data from pcre_study()
4009 match_limit a way of specifying a limit on match() calls for a specific
4010 call to pcre_exec()
4011 callout_data data for callouts (see 49 below)
4012
4013 The flag bits are also defined in pcre.h, and are
4014
4015 PCRE_EXTRA_STUDY_DATA
4016 PCRE_EXTRA_MATCH_LIMIT
4017 PCRE_EXTRA_CALLOUT_DATA
4018
4019 The pcre_study() function now returns one of these new pcre_extra blocks, with
4020 the actual study data pointed to by the study_data field, and the
4021 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4022 before. That is, this change is entirely upwards-compatible and requires no
4023 change to existing code.
4024
4025 If you want to pass in additional data to pcre_exec(), you can either place it
4026 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4027 block.
4028
4029 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4030 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4031 times with different match limits, until it finds the minimum value needed for
4032 pcre_exec() to complete. The value is then output. This can be instructive; for
4033 most simple matches the number is quite small, but for pathological cases it
4034 gets very large very quickly.
4035
4036 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4037 returns the size of the data block pointed to by the study_data field in a
4038 pcre_extra block, that is, the value that was passed as the argument to
4039 pcre_malloc() when PCRE was getting memory in which to place the information
4040 created by pcre_study(). The fourth argument should point to a size_t variable.
4041 pcretest has been extended so that this information is shown after a successful
4042 pcre_study() call when information about the compiled regex is being displayed.
4043
4044 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4045 because what follows is always an absolute path. (Later: it turns out that this
4046 is more than cosmetic for MinGW, because it doesn't like empty path
4047 components.)
4048
4049 49. Some changes have been made to the callout feature (see 28 above):
4050
4051 (i) A callout function now has three choices for what it returns:
4052
4053 0 => success, carry on matching
4054 > 0 => failure at this point, but backtrack if possible
4055 < 0 => serious error, return this value from pcre_exec()
4056
4057 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4058 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4059 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4060 use by callout functions. It will never be used by PCRE itself.
4061
4062 (ii) The pcre_extra structure (see 45 above) has a void * field called
4063 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4064 pcre_callout_block structure has a field of the same name. The contents of
4065 the field passed in the pcre_extra structure are passed to the callout
4066 function in the corresponding field in the callout block. This makes it
4067 easier to use the same callout-containing regex from multiple threads. For
4068 testing, the pcretest program has a new data escape
4069
4070 \C*n pass the number n (may be negative) as callout_data
4071
4072 If the callout function in pcretest receives a non-zero value as
4073 callout_data, it returns that value.
4074
4075 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4076 there were some redundant $(CFLAGS) in commands that are now specified as
4077 $(LINK), which already includes $(CFLAGS).
4078
4079 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4080 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4081 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4082 one-byte characters throughout. Note that case-insensitive matching applies
4083 only to characters whose values are less than 256. PCRE doesn't support the
4084 notion of cases for higher-valued characters.
4085
4086 (i) A character class whose characters are all within 0-255 is handled as
4087 a bit map, and the map is inverted for negative classes. Previously, a
4088 character > 255 always failed to match such a class; however it should
4089 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4090
4091 (ii) A negated character class with a single character < 255 is coded as
4092 "not this character" (OP_NOT). This wasn't working properly when the test
4093 character was multibyte, either singly or repeated.
4094
4095 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4096 mode, for example: \x{100}{2,3}.
4097
4098 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4099 singly or repeated) now correctly test multibyte characters. However,
4100 PCRE doesn't recognize any characters with values greater than 255 as
4101 digits, spaces, or word characters. Such characters always match \D, \S,
4102 and \W, and never match \d, \s, or \w.
4103
4104 (v) Classes may now contain characters and character ranges with values
4105 greater than 255. For example: [ab\x{100}-\x{400}].
4106
4107 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4108 PCRE in UTF-8 mode.
4109
4110 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4111 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4112 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4113 value.)
4114
4115 53. The single man page has become too large. I have therefore split it up into
4116 a number of separate man pages. These also give rise to individual HTML pages;
4117 these are now put in a separate directory, and there is an index.html page that
4118 lists them all. Some hyperlinking between the pages has been installed.
4119
4120 54. Added convenience functions for handling named capturing parentheses.
4121
4122 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4123 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4124 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4125 are faulted.
4126
4127 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4128 calling configure. These values are used when compiling the dftables.c program
4129 which is run to generate the source of the default character tables. They
4130 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4131 you will need to set these values.
4132
4133 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4134
4135
4136 Version 3.9 02-Jan-02
4137 ---------------------
4138
4139 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4140
4141 2. If --disable-static was given, the building process failed when trying to
4142 build pcretest and pcregrep. (For some reason it was using libtool to compile
4143 them, which is not right, as they aren't part of the library.)
4144
4145
4146 Version 3.8 18-Dec-01
4147 ---------------------
4148
4149 1. The experimental UTF-8 code was completely screwed up. It was packing the
4150 bytes in the wrong order. How dumb can you get?
4151
4152
4153 Version 3.7 29-Oct-01
4154 ---------------------
4155
4156 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4157 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4158 this didn't happen under Solaris 8, where I normally test things.
4159
4160 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4161 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4162 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4163 this fix an hour or so after the initial 3.7 release.)
4164
4165
4166 Version 3.6 23-Oct-01
4167 ---------------------
4168
4169 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4170 offsets passed as NULL with zero offset count.
4171
4172 2. The config.guess and config.sub files had not been updated when I moved to
4173 the latest autoconf.
4174
4175
4176 Version 3.5 15-Aug-01
4177 ---------------------
4178
4179 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4180 had been forgotten.
4181
4182 2. By using declared but undefined structures, we can avoid using "void"
4183 definitions in pcre.h while keeping the internal definitions of the structures
4184 private.
4185
4186 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4187 user point of view, this means that both static and shared libraries are built
4188 by default, but this can be individually controlled. More of the work of
4189 handling this static/shared cases is now inside libtool instead of PCRE's make
4190 file.
4191
4192 4. The pcretest utility is now installed along with pcregrep because it is
4193 useful for users (to test regexs) and by doing this, it automatically gets
4194 relinked by libtool. The documentation has been turned into a man page, so
4195 there are now .1, .txt, and .html versions in /doc.
4196
4197 5. Upgrades to pcregrep:
4198 (i) Added long-form option names like gnu grep.
4199 (ii) Added --help to list all options with an explanatory phrase.
4200 (iii) Added -r, --recursive to recurse into sub-directories.
4201 (iv) Added -f, --file to read patterns from a file.
4202
4203 6. pcre_exec() was referring to its "code" argument before testing that
4204 argument for NULL (and giving an error if it was NULL).
4205
4206 7. Upgraded Makefile.in to allow for compiling in a different directory from
4207 the source directory.
4208
4209 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4210 options bits, the pointer it was passed was to an int instead of to an unsigned
4211 long int. This mattered only on 64-bit systems.
4212
4213 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4214 generated) instead of pcre.in, which it its source. Also made the same change
4215 in several of the .c files.
4216
4217 10. A new release of gcc defines printf() as a macro, which broke pcretest
4218 because it had an ifdef in the middle of a string argument for printf(). Fixed
4219 by using separate calls to printf().
4220
4221 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4222 script, to force use of CR or LF instead of \n in the source. On non-Unix
4223 systems, the value can be set in config.h.
4224
4225 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4226 absolute limit. Changed the text of the error message to make this clear, and
4227 likewise updated the man page.
4228
4229 13. The limit of 99 on the number of capturing subpatterns has been removed.
4230 The new limit is 65535, which I hope will not be a "real" limit.
4231
4232
4233 Version 3.4 22-Aug-00
4234 ---------------------
4235
4236 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4237
4238 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4239
4240
4241 Version 3.3 01-Aug-00
4242 ---------------------
4243
4244 1. If an octal character was given, but the value was greater than \377, it
4245 was not getting masked to the least significant bits, as documented. This could
4246 lead to crashes in some systems.
4247
4248 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4249 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4250
4251 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4252 These just pass their arguments on to (pcre_free)(), but they are provided
4253 because some uses of PCRE bind it to non-C systems that can call its functions,
4254 but cannot call free() or pcre_free() directly.
4255
4256 4. Add "make test" as a synonym for "make check". Corrected some comments in
4257 the Makefile.
4258
4259 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4260 Makefile.
4261
4262 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4263 command called pgrep for grepping around the active processes.
4264
4265 7. Added the beginnings of support for UTF-8 character strings.
4266
4267 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4268 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4269 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4270 out for the ar command.)
4271
4272
4273 Version 3.2 12-May-00
4274 ---------------------
4275
4276 This is purely a bug fixing release.
4277
4278 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4279 of ZA. This was just one example of several cases that could provoke this bug,
4280 which was introduced by change 9 of version 2.00. The code for breaking
4281 infinite loops after an iteration that matches an empty string was't working
4282 correctly.
4283
4284 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4285 when matched against abbab (for example). After matching an empty string, it
4286 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4287 caused it to match further down the string than it should.
4288
4289 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4290 was there because it doesn't seem to be needed, and it causes trouble on some
4291 systems, as it is not a Standard C header. It has been removed.
4292
4293 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4294 were reported on the Macintosh. The changes were from
4295
4296 while ((c = *(++ptr)) != 0 && c != '\n');
4297 to
4298 while ((c = *(++ptr)) != 0 && c != '\n') ;
4299
4300 Totally extraordinary, but if that's what it takes...
4301
4302 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4303 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4304 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4305 assumes the way PCRE uses memmove() (always moving upwards).
4306
4307 6. PCRE is being used in one environment where strchr() is not available. There
4308 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4309 faster code anyway.
4310
4311
4312 Version 3.1 09-Feb-00
4313 ---------------------
4314
4315 The only change in this release is the fixing of some bugs in Makefile.in for
4316 the "install" target:
4317
4318 (1) It was failing to install pcreposix.h.
4319
4320 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4321
4322
4323 Version 3.0 01-Feb-00
4324 ---------------------
4325
4326 1. Add support for the /+ modifier to perltest (to output $` like it does in
4327 pcretest).
4328
4329 2. Add support for the /g modifier to perltest.
4330
4331 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4332 matches null strings.
4333
4334 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4335 pattern. Perl treats empty patterns specially - it reuses the most recent
4336 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4337 effect.
4338
4339 5. The POSIX interface was broken in that it was just handing over the POSIX
4340 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4341 required a bigger vector, with some working space on the end. This means that
4342 the POSIX wrapper now has to get and free some memory, and copy the results.
4343
4344 6. Added some simple autoconf support, placing the test data and the
4345 documentation in separate directories, re-organizing some of the
4346 information files, and making it build pcre-config (a GNU standard). Also added
4347 libtool support for building PCRE as a shared library, which is now the
4348 default.
4349
4350 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4351 09 are not valid octal constants. Single digits will be used for minor values
4352 less than 10.
4353
4354 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4355 existing programs that set these in the POSIX interface can use PCRE without
4356 modification.
4357
4358 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4359 return all that pcre_info() returns, plus additional data. The pcre_info()
4360 function is retained for compatibility, but is considered to be obsolete.
4361
4362 10. Added experimental recursion feature (?R) to handle one common case that
4363 Perl 5.6 will be able to do with (?p{...}).
4364
4365 11. Added support for POSIX character classes like [:alpha:], which Perl is
4366 adopting.
4367
4368
4369 Version 2.08 31-Aug-99
4370 ----------------------
4371
4372 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4373 trying to match at the startoffset position, but instead was moving forward to
4374 the next newline as if a previous match had failed.
4375
4376 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4377 and could get into a loop if a null string was matched other than at the start
4378 of the subject.
4379
4380 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4381 be distinguished at compile time, and for completeness also added PCRE_DATE.
4382
4383 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4384 in GnuWin32 environments.
4385
4386
4387 Version 2.07 29-Jul-99
4388 ----------------------
4389
4390 1. The documentation is now supplied in plain text form and HTML as well as in
4391 the form of man page sources.
4392
4393 2. C++ compilers don't like assigning (void *) values to other pointer types.
4394 In particular this affects malloc(). Although there is no problem in Standard
4395 C, I've put in casts to keep C++ compilers happy.
4396
4397 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4398 should be (const char *).
4399
4400 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4401 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4402 However, I haven't made this a standard facility. The documentation doesn't
4403 mention it, and the Makefile doesn't support it.
4404
4405 5. The Makefile now contains an "install" target, with editable destinations at
4406 the top of the file. The pcretest program is not installed.
4407
4408 6. pgrep -V now gives the PCRE version number and date.
4409
4410 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4411 causing the entire string to be ignored, instead of just the last character.
4412
4413 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4414 non-matching string, it can take a very, very long time, even for strings of
4415 quite modest length, because of the nested recursion. PCRE now does better in
4416 some of these cases. It does this by remembering the last required literal
4417 character in the pattern, and pre-searching the subject to ensure it is present
4418 before running the real match. In other words, it applies a heuristic to detect
4419 some types of certain failure quickly, and in the above example, if presented
4420 with a string that has no trailing " it gives "no match" very quickly.
4421
4422 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4423 other alternatives are tried instead.
4424
4425
4426 Version 2.06 09-Jun-99
4427 ----------------------
4428
4429 1. Change pcretest's output for amount of store used to show just the code
4430 space, because the remainder (the data block) varies in size between 32-bit and
4431 64-bit systems.
4432
4433 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4434 start matching at. This allows lookbehinds to work when searching for multiple
4435 occurrences in a string.
4436
4437 3. Added additional options to pcretest for testing multiple occurrences:
4438
4439 /+ outputs the rest of the string that follows a match
4440 /g loops for multiple occurrences, using the new startoffset argument
4441 /G loops for multiple occurrences by passing an incremented pointer
4442
4443 4. PCRE wasn't doing the "first character" optimization for patterns starting
4444 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4445 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4446 the letter 'x'. On long subject strings, this gives a significant speed-up.
4447
4448
4449 Version 2.05 21-Apr-99
4450 ----------------------
4451
4452 1. Changed the type of magic_number from int to long int so that it works
4453 properly on 16-bit systems.
4454
4455 2. Fixed a bug which caused patterns starting with .* not to work correctly
4456 when the subject string contained newline characters. PCRE was assuming
4457 anchoring for such patterns in all cases, which is not correct because .* will
4458 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4459 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4460 must be retried after every newline in the subject.
4461
4462
4463 Version 2.04 18-Feb-99
4464 ----------------------
4465
4466 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4467 computation of the store needed to hold the pattern was incorrect (too large).
4468 If such patterns were nested a few deep, this could multiply and become a real
4469 problem.
4470
4471 2. Added /M option to pcretest to show the memory requirement of a specific
4472 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4473
4474 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4475 compiled in such a way that the backtracking after subsequent failure was
4476 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4477 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4478
4479
4480 Version 2.03 02-Feb-99
4481 ----------------------
4482
4483 1. Fixed typo and small mistake in man page.
4484
4485 2. Added 4th condition (GPL supersedes if conflict) and created separate
4486 LICENCE file containing the conditions.
4487
4488 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4489 Perl, that is the internal \ allows the delimiter to be included in the
4490 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4491 the final delimiter, add \ to the end of the pattern (to test the error).
4492
4493 4. Added the convenience functions for extracting substrings after a successful
4494 match. Updated pcretest to make it able to test these functions.
4495
4496
4497 Version 2.02 14-Jan-99
4498 ----------------------
4499
4500 1. Initialized the working variables associated with each extraction so that
4501 their saving and restoring doesn't refer to uninitialized store.
4502
4503 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4504 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4505 fix the problem.
4506
4507 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4508 calls, and wasn't printing the correct value for compiling calls. Increased the
4509 default value of LOOPREPEAT, and the number of significant figures in the
4510 times.
4511
4512 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4513
4514 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4515 a building problem on Windows NT with a FAT file system.
4516
4517
4518 Version 2.01 21-Oct-98
4519 ----------------------
4520
4521 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4522 to character tables built by pcre_maketables() in the current locale. If NULL
4523 is passed, the default tables are used.
4524
4525
4526 Version 2.00 24-Sep-98
4527 ----------------------
4528
4529 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4530 it any more.
4531
4532 2. Allow quantification of (?>) groups, and make it work correctly.
4533
4534 3. The first character computation wasn't working for (?>) groups.
4535
4536 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4537 end of the subject) and add 5.005's \z, which really does match only at the
4538 very end of the subject.
4539
4540 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4541
4542 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4543 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4544 localized options. All options to pcre_study() were also removed.
4545
4546 7. Add other new features from 5.005:
4547
4548 $(?<= positive lookbehind
4549 $(?<! negative lookbehind
4550 (?imsx-imsx) added the unsetting capability
4551 such a setting is global if at outer level; local otherwise
4552 (?imsx-imsx:) non-capturing groups with option setting
4553 (?(cond)re|re) conditional pattern matching
4554
4555 A backreference to itself in a repeated group matches the previous
4556 captured string.
4557
4558 8. General tidying up of studying (both automatic and via "study")
4559 consequential on the addition of new assertions.
4560
4561 9. As in 5.005, unlimited repeated groups that could match an empty substring
4562 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4563 runtime if any iteration does actually match an empty substring.
4564
4565 10. Include the RunTest script in the distribution.
4566
4567 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4568 discrepancies, some of which were old and were also with respect to 5.004. They
4569 have now been fixed.
4570
4571
4572 Version 1.09 28-Apr-98
4573 ----------------------
4574
4575 1. A negated single character class followed by a quantifier with a minimum
4576 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4577 program crashes, or just wrong answers. This did not apply to negated classes
4578 containing more than one character, or to minima other than one.
4579
4580
4581 Version 1.08 27-Mar-98
4582 ----------------------
4583
4584 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4585
4586 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4587 latter must appear before anything that relies on it in the pattern.
4588
4589
4590 Version 1.07 16-Feb-98
4591 ----------------------
4592
4593 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4594 repeat of a potentially empty string).
4595
4596
4597 Version 1.06 23-Jan-98
4598 ----------------------
4599
4600 1. Added Markus Oberhumer's little patches for C++.
4601
4602 2. Literal strings longer than 255 characters were broken.
4603
4604
4605 Version 1.05 23-Dec-97
4606 ----------------------
4607
4608 1. Negated character classes containing more than one character were failing if
4609 PCRE_CASELESS was set at run time.
4610
4611
4612 Version 1.04 19-Dec-97
4613 ----------------------
4614
4615 1. Corrected the man page, where some "const" qualifiers had been omitted.
4616
4617 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4618 input syntax.
4619
4620 3. Fixed memory leak which occurred when a regex with back references was
4621 matched with an offsets vector that wasn't big enough. The temporary memory
4622 that is used in this case wasn't being freed if the match failed.
4623
4624 4. Tidied pcretest to ensure it frees memory that it gets.
4625
4626 5. Temporary memory was being obtained in the case where the passed offsets
4627 vector was exactly big enough.
4628
4629 6. Corrected definition of offsetof() from change 5 below.
4630
4631 7. I had screwed up change 6 below and broken the rules for the use of
4632 setjmp(). Now fixed.
4633
4634
4635 Version 1.03 18-Dec-97
4636 ----------------------
4637
4638 1. A erroneous regex with a missing opening parenthesis was correctly
4639 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4640 on some systems.
4641
4642 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4643 it was reported that one broken compiler failed on the former because "code" is
4644 also an independent variable.
4645
4646 3. The erroneous regex a[]b caused an array overrun reference.
4647
4648 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4649 fail on data ending with that character. (It was going on too far, and checking
4650 the next character, typically a binary zero.) This was specific to the
4651 optimized code for single-character negative classes.
4652
4653 5. Added a contributed patch from the TIN world which does the following:
4654
4655 + Add an undef for memmove, in case the the system defines a macro for it.
4656
4657 + Add a definition of offsetof(), in case there isn't one. (I don't know
4658 the reason behind this - offsetof() is part of the ANSI standard - but
4659 it does no harm).
4660
4661 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4662 most of the places where whitespace preceded '#'. I have given up and
4663 allowed the remaining 2 cases to be at the margin.
4664
4665 + Rename some variables in pcre to eliminate shadowing. This seems very
4666 pedantic, but does no harm, of course.
4667
4668 6. Moved the call to setjmp() into its own function, to get rid of warnings
4669 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4670
4671 7. Constructs such as \d{8,} were compiling into the equivalent of
4672 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4673 outcome, but in this particular case used more store than had been allocated,
4674 which caused the bug to be discovered because it threw up an internal error.
4675
4676 8. The debugging code in both pcre and pcretest for outputting the compiled
4677 form of a regex was going wrong in the case of back references followed by
4678 curly-bracketed repeats.
4679
4680
4681 Version 1.02 12-Dec-97
4682 ----------------------
4683
4684 1. Typos in pcre.3 and comments in the source fixed.
4685
4686 2. Applied a contributed patch to get rid of places where it used to remove
4687 'const' from variables, and fixed some signed/unsigned and uninitialized
4688 variable warnings.
4689
4690 3. Added the "runtest" target to Makefile.
4691
4692 4. Set default compiler flag to -O2 rather than just -O.
4693
4694
4695 Version 1.01 19-Nov-97
4696 ----------------------
4697
4698 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4699 like /([ab]*)*/, that is, for classes with more than one character in them.
4700
4701 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4702 as /((?>a*))*/ (a PCRE_EXTRA facility).
4703
4704
4705 Version 1.00 18-Nov-97
4706 ----------------------
4707
4708 1. Added compile-time macros to support systems such as SunOS4 which don't have
4709 memmove() or strerror() but have other things that can be used instead.
4710
4711 2. Arranged that "make clean" removes the executables.
4712
4713
4714 Version 0.99 27-Oct-97
4715 ----------------------
4716
4717 1. Fixed bug in code for optimizing classes with only one character. It was
4718 initializing a 32-byte map regardless, which could cause it to run off the end
4719 of the memory it had got.
4720
4721 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4722
4723
4724 Version 0.98 22-Oct-97
4725 ----------------------
4726
4727 1. Fixed bug in code for handling temporary memory usage when there are more
4728 back references than supplied space in the ovector. This could cause segfaults.
4729
4730
4731 Version 0.97 21-Oct-97
4732 ----------------------
4733
4734 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4735
4736 2. Optimized negated single characters not to use a bit map.
4737
4738 3. Brought error texts together as macro definitions; clarified some of them;
4739 fixed one that was wrong - it said "range out of order" when it meant "invalid
4740 escape sequence".
4741
4742 4. Changed some char * arguments to const char *.
4743
4744 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4745
4746 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4747 pcretest.
4748
4749
4750 Version 0.96 16-Oct-97
4751 ----------------------
4752
4753 1. Added a simple "pgrep" utility to the distribution.
4754
4755 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4756 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4757 where "ddd" means "one or more decimal digits".
4758
4759 3. Fixed serious bug. If a pattern had a back reference, but the call to
4760 pcre_exec() didn't supply a large enough ovector to record the related
4761 identifying subpattern, the match always failed. PCRE now remembers the number
4762 of the largest back reference, and gets some temporary memory in which to save
4763 the offsets during matching if necessary, in order to ensure that
4764 backreferences always work.
4765
4766 4. Increased the compatibility with Perl in a number of ways:
4767
4768 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4769 to request this handling. The option can be set at compile or exec time.
4770
4771 (b) $ matches before a terminating newline by default; an option
4772 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4773 mode). The option can be set at compile or exec time.
4774
4775 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4776 the same as Perl's. If the decimal number it represents is less than 10
4777 or there aren't that many previous left capturing parentheses, an octal
4778 escape is read. Inside a character class, it's always an octal escape,
4779 even if it is a single digit.
4780
4781 (d) An escaped but undefined alphabetic character is taken as a literal,
4782 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4783 escapes.
4784
4785 (e) {0} is now permitted. (The previous item is removed from the compiled
4786 pattern).
4787
4788 5. Changed all the names of code files so that the basic parts are no longer
4789 than 10 characters, and abolished the teeny "globals.c" file.
4790
4791 6. Changed the handling of character classes; they are now done with a 32-byte
4792 bit map always.
4793
4794 7. Added the -d and /D options to pcretest to make it possible to look at the
4795 internals of compilation without having to recompile pcre.
4796
4797
4798 Version 0.95 23-Sep-97
4799 ----------------------
4800
4801 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4802 \x20 at the start of a run of normal characters. These were being treated as
4803 real characters, instead of the source characters being re-checked.
4804
4805
4806 Version 0.94 18-Sep-97
4807 ----------------------
4808
4809 1. The functions are now thread-safe, with the caveat that the global variables
4810 containing pointers to malloc() and free() or alternative functions are the
4811 same for all threads.
4812
4813 2. Get pcre_study() to generate a bitmap of initial characters for non-
4814 anchored patterns when this is possible, and use it if passed to pcre_exec().
4815
4816
4817 Version 0.93 15-Sep-97
4818 ----------------------
4819
4820 1. /(b)|(:+)/ was computing an incorrect first character.
4821
4822 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4823 but not actually doing anything yet.
4824
4825 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4826 as Perl does (e.g. [-az] or [az-]).
4827
4828 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4829 all possible positions.
4830
4831 5. Split up into different modules to avoid including unneeded functions in a
4832 compiled binary. However, compile and exec are still in one module. The "study"
4833 function is split off.
4834
4835 6. The character tables are now in a separate module whose source is generated
4836 by an auxiliary program - but can then be edited by hand if required. There are
4837 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4838 toupper() in the code.
4839
4840 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4841 make them global. Abolish the function for setting them, as the caller can now
4842 set them directly.
4843
4844
4845 Version 0.92 11-Sep-97
4846 ----------------------
4847
4848 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4849 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4850
4851 2. Caseless matching was not working in character classes if the characters in
4852 the pattern were in upper case.
4853
4854 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4855
4856 4. Make PCRE_ANCHORED public and accept as a compile option.
4857
4858 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4859 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4860 pass them.
4861
4862 6. Give an error if bad option bits passed at compile or run time.
4863
4864 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4865 pcretest to cause it to pass that flag.
4866
4867 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4868 options, and the first character, if set.
4869
4870 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4871
4872
4873 Version 0.91 10-Sep-97
4874 ----------------------
4875
4876 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4877 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4878
4879 2. PCRE was looping on encountering an indefinitely repeated back reference to
4880 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4881 Perl does - treats the match as successful.
4882
4883 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5