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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1024 - (show annotations)
Wed Aug 29 11:08:24 2012 UTC (7 years, 1 month ago) by ph10
File size: 235060 byte(s)
Fix re_nsub bug in the POSIX wrapper.
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
66
67 Version 8.31 06-July-2012
68 -------------------------
69
70 1. Fixing a wrong JIT test case and some compiler warnings.
71
72 2. Removed a bashism from the RunTest script.
73
74 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
75 to unsigned type, result still unsigned" that was given by an MS compiler
76 on encountering the code "-sizeof(xxx)".
77
78 4. Partial matching support is added to the JIT compiler.
79
80 5. Fixed several bugs concerned with partial matching of items that consist
81 of more than one character:
82
83 (a) /^(..)\1/ did not partially match "aba" because checking references was
84 done on an "all or nothing" basis. This also applied to repeated
85 references.
86
87 (b) \R did not give a hard partial match if \r was found at the end of the
88 subject.
89
90 (c) \X did not give a hard partial match after matching one or more
91 characters at the end of the subject.
92
93 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
94 a partial match for the string "\r".
95
96 (e) When newline was set to CRLF, the metacharacter "." did not recognize
97 a partial match for a CR character at the end of the subject string.
98
99 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
100 running pcretest, the text "(JIT)" added to the output whenever JIT is
101 actually used to run the match.
102
103 7. Individual JIT compile options can be set in pcretest by following -s+[+]
104 or /S+[+] with a digit between 1 and 7.
105
106 8. OP_NOT now supports any UTF character not just single-byte ones.
107
108 9. (*MARK) control verb is now supported by the JIT compiler.
109
110 10. The command "./RunTest list" lists the available tests without actually
111 running any of them. (Because I keep forgetting what they all are.)
112
113 11. Add PCRE_INFO_MAXLOOKBEHIND.
114
115 12. Applied a (slightly modified) user-supplied patch that improves performance
116 when the heap is used for recursion (compiled with --disable-stack-for-
117 recursion). Instead of malloc and free for each heap frame each time a
118 logical recursion happens, frames are retained on a chain and re-used where
119 possible. This sometimes gives as much as 30% improvement.
120
121 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
122 call.
123
124 14. As documented, (*COMMIT) is now confined to within a positive assertion.
125
126 15. It is now possible to link pcretest with libedit as an alternative to
127 libreadline.
128
129 16. (*COMMIT) control verb is now supported by the JIT compiler.
130
131 17. The Unicode data tables have been updated to Unicode 6.1.0.
132
133 18. Added --file-list option to pcregrep.
134
135 19. Added binary file support to pcregrep, including the -a, --binary-files,
136 -I, and --text options.
137
138 20. The madvise function is renamed for posix_madvise for QNX compatibility
139 reasons. Fixed by Giuseppe D'Angelo.
140
141 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
142
143 22. Changed the option for creating symbolic links for 16-bit man pages from
144 -s to -sf so that re-installing does not cause issues.
145
146 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
147
148 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
149 matches in certain environments (the workspace was not being correctly
150 retained). Also added to pcre_dfa_exec() a simple plausibility check on
151 some of the workspace data at the beginning of a restart.
152
153 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
154 was not doing so when it should - probably a typo introduced by SVN 528
155 (change 8.10/14).
156
157 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
158 \w+ when the character tables indicated that \x{c4} was a word character.
159 There were several related cases, all because the tests for doing a table
160 lookup were testing for characters less than 127 instead of 255.
161
162 27. If a pattern contains capturing parentheses that are not used in a match,
163 their slots in the ovector are set to -1. For those that are higher than
164 any matched groups, this happens at the end of processing. In the case when
165 there were back references that the ovector was too small to contain
166 (causing temporary malloc'd memory to be used during matching), and the
167 highest capturing number was not used, memory off the end of the ovector
168 was incorrectly being set to -1. (It was using the size of the temporary
169 memory instead of the true size.)
170
171 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
172 ovector size, it uses memory at the end of the block that it has got.
173
174 29. Check for an overlong MARK name and give an error at compile time. The
175 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
176
177 30. JIT compiler update.
178
179 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
180 Rill for the patch.
181
182 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
183
184 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
185
186 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
187 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
188 was enabled.
189
190 35. Improve JIT code generation for greedy plus quantifier.
191
192 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
193 1 to "aa" instead of to an empty string. The bug affected repeated groups
194 that could potentially match an empty string.
195
196 37. Optimizing single character iterators in JIT.
197
198 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
199 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
200 codepoints that are too big for the mode are faulted, and in a UTF mode,
201 disallowed codepoints are also faulted.
202
203 39. If PCRE was compiled with UTF support, in three places in the DFA
204 matcher there was code that should only have been obeyed in UTF mode, but
205 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
206 processing when bytes with values greater than 127 were present. In 16-bit
207 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
208 both cases the values are those that cannot be the first data item in a UTF
209 character. The three items that might have provoked this were recursions,
210 possessively repeated groups, and atomic groups.
211
212 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
213 and pcregrep, because some OS require shared objects to be explicitly
214 passed to ld, causing the link step to fail if they are not.
215
216 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
217 mode, patterns that started with \h* or \R* might be incorrectly matched.
218
219
220 Version 8.30 04-February-2012
221 -----------------------------
222
223 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
224 name is defined in ctype.h.
225
226 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
227 only in quite long subpatterns.
228
229 3. Removed the function pcre_info(), which has been obsolete and deprecated
230 since it was replaced by pcre_fullinfo() in February 2000.
231
232 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
233 match a (*MARK), and the match failed at the start of the subject, a
234 reference to memory before the start of the subject could occur. This bug
235 was introduced by fix 17 of release 8.21.
236
237 5. A reference to an unset group with zero minimum repetition was giving
238 totally wrong answers (in non-JavaScript-compatibility mode). For example,
239 /(another)?(\1?)test/ matched against "hello world test". This bug was
240 introduced in release 8.13.
241
242 6. Add support for 16-bit character strings (a large amount of work involving
243 many changes and refactorings).
244
245 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
246 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
247 from a file.
248
249 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
250 rounding is not applied in this particular case).
251
252 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
253 if they appear, or are escaped, in patterns.
254
255 10. Get rid of a number of -Wunused-but-set-variable warnings.
256
257 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
258 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
259 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
260 also returns the mark "x". This bug applied to capturing parentheses,
261 non-capturing parentheses, and atomic parentheses. It also applied to some
262 assertions.
263
264 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
265 information out of configure.ac instead of relying on pcre.h.generic, which
266 is not stored in the repository.
267
268 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
269 -lreadline.
270
271 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
272
273 15. Applied Graycode's patch to put the top-level frame on the stack rather
274 than the heap when not using the stack for recursion. This gives a
275 performance improvement in many cases when recursion is not deep.
276
277 16. Experimental code added to "pcretest -C" to output the stack frame size.
278
279
280 Version 8.21 12-Dec-2011
281 ------------------------
282
283 1. Updating the JIT compiler.
284
285 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
286 are added as well.
287
288 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
289 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
290 calling _pcre_jit_exec. Some extra comments are added.
291
292 4. (*MARK) settings inside atomic groups that do not contain any capturing
293 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
294 was introduced by change 18 for 8.20.
295
296 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
297 ECMA-262 standard.
298
299 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
300 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
301 This bug was probably introduced by change 9 of 8.13.
302
303 7. While fixing 6 above, I noticed that a number of other items were being
304 incorrectly rejected as "not fixed length". This arose partly because newer
305 opcodes had not been added to the fixed-length checking code. I have (a)
306 corrected the bug and added tests for these items, and (b) arranged for an
307 error to occur if an unknown opcode is encountered while checking for fixed
308 length instead of just assuming "not fixed length". The items that were
309 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
310 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
311 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
312
313 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
314 being incorrectly compiled and would have given unpredicatble results.
315
316 9. A possessively repeated subpattern with minimum repeat count greater than
317 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
318 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
319 the first (A) could occur when it should not.
320
321 10. Add a cast and remove a redundant test from the code.
322
323 11. JIT should use pcre_malloc/pcre_free for allocation.
324
325 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
326 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
327 is anything other than /usr, -L is still shown).
328
329 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
330
331 14. Perl does not support \N without a following name in a [] class; PCRE now
332 also gives an error.
333
334 15. If a forward reference was repeated with an upper limit of around 2000,
335 it caused the error "internal error: overran compiling workspace". The
336 maximum number of forward references (including repeats) was limited by the
337 internal workspace, and dependent on the LINK_SIZE. The code has been
338 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
339 the default depends on LINK_SIZE. There is a new upper limit (for safety)
340 of around 200,000 forward references. While doing this, I also speeded up
341 the filling in of repeated forward references.
342
343 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
344 incorrectly expecting the subject to contain another "a" after the start.
345
346 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
347 in the match, the SKIP should be ignored. This was not happening; instead
348 the SKIP was being treated as NOMATCH. For patterns such as
349 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
350 tested.
351
352 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
353 now much more compatible with Perl, in particular in cases where the result
354 is a non-match for a non-anchored pattern. For example, if
355 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
356 "m", where previously it did not return a name. A side effect of this
357 change is that for partial matches, the last encountered mark name is
358 returned, as for non matches. A number of tests that were previously not
359 Perl-compatible have been moved into the Perl-compatible test files. The
360 refactoring has had the pleasing side effect of removing one argument from
361 the match() function, thus reducing its stack requirements.
362
363 19. If the /S+ option was used in pcretest to study a pattern using JIT,
364 subsequent uses of /S (without +) incorrectly behaved like /S+.
365
366 21. Retrieve executable code size support for the JIT compiler and fixing
367 some warnings.
368
369 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
370 not work when the shorter character appeared right at the end of the
371 subject string.
372
373 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
374 systems.
375
376 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
377 output it when the /M option is used in pcretest.
378
379 25. The CheckMan script was not being included in the distribution. Also, added
380 an explicit "perl" to run Perl scripts from the PrepareRelease script
381 because this is reportedly needed in Windows.
382
383 26. If study data was being save in a file and studying had not found a set of
384 "starts with" bytes for the pattern, the data written to the file (though
385 never used) was taken from uninitialized memory and so caused valgrind to
386 complain.
387
388 27. Updated RunTest.bat as provided by Sheri Pierce.
389
390 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
391
392 29. Computation of memory usage for the table of capturing group names was
393 giving an unnecessarily large value.
394
395
396 Version 8.20 21-Oct-2011
397 ------------------------
398
399 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
400 a POSIX class. After further experiments with Perl, which convinced me that
401 Perl has bugs and confusions, a closing square bracket is no longer allowed
402 in a POSIX name. This bug also affected patterns with classes that started
403 with full stops.
404
405 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
406 captured substring, but while checking the failing first alternative,
407 substring 1 is temporarily captured. If the output vector supplied to
408 pcre_exec() was not big enough for this capture, the yield of the function
409 was still zero ("insufficient space for captured substrings"). This cannot
410 be totally fixed without adding another stack variable, which seems a lot
411 of expense for a edge case. However, I have improved the situation in cases
412 such as /(a)(b)x|abc/ matched against "abc", where the return code
413 indicates that fewer than the maximum number of slots in the ovector have
414 been set.
415
416 3. Related to (2) above: when there are more back references in a pattern than
417 slots in the output vector, pcre_exec() uses temporary memory during
418 matching, and copies in the captures as far as possible afterwards. It was
419 using the entire output vector, but this conflicts with the specification
420 that only 2/3 is used for passing back captured substrings. Now it uses
421 only the first 2/3, for compatibility. This is, of course, another edge
422 case.
423
424 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
425 main code base, and can be used by building with --enable-jit. When this is
426 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
427 runtime --no-jit option is given.
428
429 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
430 ovector, the return from the function was zero, implying that there were
431 other matches that did not fit. The correct "exactly full" value is now
432 returned.
433
434 6. If a subpattern that was called recursively or as a subroutine contained
435 (*PRUNE) or any other control that caused it to give a non-standard return,
436 invalid errors such as "Error -26 (nested recursion at the same subject
437 position)" or even infinite loops could occur.
438
439 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
440 computing the minimum length on reaching *ACCEPT, and so ended up with the
441 wrong value of 1 rather than 0. Further investigation indicates that
442 computing a minimum subject length in the presence of *ACCEPT is difficult
443 (think back references, subroutine calls), and so I have changed the code
444 so that no minimum is registered for a pattern that contains *ACCEPT.
445
446 8. If (*THEN) was present in the first (true) branch of a conditional group,
447 it was not handled as intended. [But see 16 below.]
448
449 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
450 Sheri Pierce.
451
452 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
453 the first byte in a match must be "a".
454
455 11. Change 17 for 8.13 increased the recursion depth for patterns like
456 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
457 pattern contains any instances of (*THEN). If it does not, the old
458 optimizations are restored. It would be nice to do this on a per-group
459 basis, but at the moment that is not feasible.
460
461 12. In some environments, the output of pcretest -C is CRLF terminated. This
462 broke RunTest's code that checks for the link size. A single white space
463 character after the value is now allowed for.
464
465 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
466 For "fr", it uses the Windows-specific input and output files.
467
468 14. If (*THEN) appeared in a group that was called recursively or as a
469 subroutine, it did not work as intended. [But see next item.]
470
471 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
472 pattern fragments (but not containing any | characters). If A and B are
473 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
474 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
475 D. In other words, Perl considers parentheses that do not contain any |
476 characters to be part of a surrounding alternative, whereas PCRE was
477 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
478 differently. PCRE now behaves in the same way as Perl, except in the case
479 of subroutine/recursion calls such as (?1) which have in any case always
480 been different (but PCRE had them first :-).
481
482 16. Related to 15 above: Perl does not treat the | in a conditional group as
483 creating alternatives. Such a group is treated in the same way as an
484 ordinary group without any | characters when processing (*THEN). PCRE has
485 been changed to match Perl's behaviour.
486
487 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
488 RunGrepTest script failed.
489
490 18. Change 22 for version 13 caused atomic groups to use more stack. This is
491 inevitable for groups that contain captures, but it can lead to a lot of
492 stack use in large patterns. The old behaviour has been restored for atomic
493 groups that do not contain any capturing parentheses.
494
495 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
496 suppress the check for a minimum subject length at run time. (If it was
497 given to pcre_exec() or pcre_dfa_exec() it did work.)
498
499 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
500 fail to work when decoding hex characters in data strings in EBCDIC
501 environments.
502
503 21. It appears that in at least one Mac OS environment, the isxdigit() function
504 is implemented as a macro that evaluates to its argument more than once,
505 contravening the C 90 Standard (I haven't checked a later standard). There
506 was an instance in pcretest which caused it to go wrong when processing
507 \x{...} escapes in subject strings. The has been rewritten to avoid using
508 things like p++ in the argument of isxdigit().
509
510
511 Version 8.13 16-Aug-2011
512 ------------------------
513
514 1. The Unicode data tables have been updated to Unicode 6.0.0.
515
516 2. Two minor typos in pcre_internal.h have been fixed.
517
518 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
519 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
520 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
521
522 4. There were a number of related bugs in the code for matching backrefences
523 caselessly in UTF-8 mode when codes for the characters concerned were
524 different numbers of bytes. For example, U+023A and U+2C65 are an upper
525 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
526 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
527 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
528 2-byte code at the end of the subject (it thought there wasn't enough data
529 left).
530
531 5. Comprehensive information about what went wrong is now returned by
532 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
533 as the output vector has at least 2 elements. The offset of the start of
534 the failing character and a reason code are placed in the vector.
535
536 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
537 now returned is for the first byte of the failing character, instead of the
538 last byte inspected. This is an incompatible change, but I hope it is small
539 enough not to be a problem. It makes the returned offset consistent with
540 pcre_exec() and pcre_dfa_exec().
541
542 7. pcretest now gives a text phrase as well as the error number when
543 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
544 failure, the offset and reason code are output.
545
546 8. When \R was used with a maximizing quantifier it failed to skip backwards
547 over a \r\n pair if the subsequent match failed. Instead, it just skipped
548 back over a single character (\n). This seems wrong (because it treated the
549 two characters as a single entity when going forwards), conflicts with the
550 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
551 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
552 has been changed.
553
554 9. Some internal refactoring has changed the processing so that the handling
555 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
556 time (the PCRE_DOTALL option was changed this way some time ago: version
557 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
558 which was always a bit of a fudge. It also means that there is one less
559 argument for the match() function, which reduces its stack requirements
560 slightly. This change also fixes an incompatibility with Perl: the pattern
561 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
562
563 10. More internal refactoring has drastically reduced the number of recursive
564 calls to match() for possessively repeated groups such as (abc)++ when
565 using pcre_exec().
566
567 11. While implementing 10, a number of bugs in the handling of groups were
568 discovered and fixed:
569
570 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
571 (a|)*(?1) gave a compile-time internal error.
572 ((a|)+)+ did not notice that the outer group could match an empty string.
573 (^a|^)+ was not marked as anchored.
574 (.*a|.*)+ was not marked as matching at start or after a newline.
575
576 12. Yet more internal refactoring has removed another argument from the match()
577 function. Special calls to this function are now indicated by setting a
578 value in a variable in the "match data" data block.
579
580 13. Be more explicit in pcre_study() instead of relying on "default" for
581 opcodes that mean there is no starting character; this means that when new
582 ones are added and accidentally left out of pcre_study(), testing should
583 pick them up.
584
585 14. The -s option of pcretest has been documented for ages as being an old
586 synonym of -m (show memory usage). I have changed it to mean "force study
587 for every regex", that is, assume /S for every regex. This is similar to -i
588 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
589 using it. It makes it easier to run collections of tests with and without
590 study enabled, and thereby test pcre_study() more easily. All the standard
591 tests are now run with and without -s (but some patterns can be marked as
592 "never study" - see 20 below).
593
594 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
595 restoration of the capturing data to the outer values was not happening
596 correctly.
597
598 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
599 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
600 pattern had matched an empty string, and so incorrectly returned a no
601 match.
602
603 17. There was optimizing code for the last branch of non-capturing parentheses,
604 and also for the obeyed branch of a conditional subexpression, which used
605 tail recursion to cut down on stack usage. Unfortunately, now that there is
606 the possibility of (*THEN) occurring in these branches, tail recursion is
607 no longer possible because the return has to be checked for (*THEN). These
608 two optimizations have therefore been removed. [But see 8.20/11 above.]
609
610 18. If a pattern containing \R was studied, it was assumed that \R always
611 matched two bytes, thus causing the minimum subject length to be
612 incorrectly computed because \R can also match just one byte.
613
614 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
615 was incorrectly computed.
616
617 20. If /S is present twice on a test pattern in pcretest input, it now
618 *disables* studying, thereby overriding the use of -s on the command line
619 (see 14 above). This is necessary for one or two tests to keep the output
620 identical in both cases.
621
622 21. When (*ACCEPT) was used in an assertion that matched an empty string and
623 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
624
625 22. When an atomic group that contained a capturing parenthesis was
626 successfully matched, but the branch in which it appeared failed, the
627 capturing was not being forgotten if a higher numbered group was later
628 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
629 group 1 to "a", when in fact it should be unset. This applied to multi-
630 branched capturing and non-capturing groups, repeated or not, and also to
631 positive assertions (capturing in negative assertions does not happen
632 in PCRE) and also to nested atomic groups.
633
634 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
635 subject after a captured substring, to make it easier to tell which of a
636 number of identical substrings has been captured.
637
638 24. The way atomic groups are processed by pcre_exec() has been changed so that
639 if they are repeated, backtracking one repetition now resets captured
640 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
641 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
642 "aaa". Previously, it would have been "a". As part of this code
643 refactoring, the way recursive calls are handled has also been changed.
644
645 25. If an assertion condition captured any substrings, they were not passed
646 back unless some other capturing happened later. For example, if
647 (?(?=(a))a) was matched against "a", no capturing was returned.
648
649 26. When studying a pattern that contained subroutine calls or assertions,
650 the code for finding the minimum length of a possible match was handling
651 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
652 group 1 called group 2 while simultaneously a separate group 2 called group
653 1). A stack overflow occurred in this case. I have fixed this by limiting
654 the recursion depth to 10.
655
656 27. Updated RunTest.bat in the distribution to the version supplied by Tom
657 Fortmann. This supports explicit test numbers on the command line, and has
658 argument validation and error reporting.
659
660 28. An instance of \X with an unlimited repeat could fail if at any point the
661 first character it looked at was a mark character.
662
663 29. Some minor code refactoring concerning Unicode properties and scripts
664 should reduce the stack requirement of match() slightly.
665
666 30. Added the '=' option to pcretest to check the setting of unused capturing
667 slots at the end of the pattern, which are documented as being -1, but are
668 not included in the return count.
669
670 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
671 compiled something random. Now it gives a compile-time error (as does
672 Perl).
673
674 32. A *MARK encountered during the processing of a positive assertion is now
675 recorded and passed back (compatible with Perl).
676
677 33. If --only-matching or --colour was set on a pcregrep call whose pattern
678 had alternative anchored branches, the search for a second match in a line
679 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
680 matched the line "0102" twice. The same bug affected patterns that started
681 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
682 twice.
683
684 34. Previously, PCRE did not allow quantification of assertions. However, Perl
685 does, and because of capturing effects, quantifying parenthesized
686 assertions may at times be useful. Quantifiers are now allowed for
687 parenthesized assertions.
688
689 35. A minor code tidy in pcre_compile() when checking options for \R usage.
690
691 36. \g was being checked for fancy things in a character class, when it should
692 just be a literal "g".
693
694 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
695 appearance of a nested POSIX class supersedes an apparent external class.
696 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
697 unescaped square brackets may also appear as part of class names. For
698 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
699 more like Perl. (But see 8.20/1 above.)
700
701 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
702 was because it thought it was \N{name}, which is not supported).
703
704 39. Add minix to OS list not supporting the -S option in pcretest.
705
706 40. PCRE tries to detect cases of infinite recursion at compile time, but it
707 cannot analyze patterns in sufficient detail to catch mutual recursions
708 such as ((?1))((?2)). There is now a runtime test that gives an error if a
709 subgroup is called recursively as a subpattern for a second time at the
710 same position in the subject string. In previous releases this might have
711 been caught by the recursion limit, or it might have run out of stack.
712
713 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
714 happen only once. PCRE was, however incorrectly giving a compile time error
715 "recursive call could loop indefinitely" because it cannot analyze the
716 pattern in sufficient detail. The compile time test no longer happens when
717 PCRE is compiling a conditional subpattern, but actual runaway loops are
718 now caught at runtime (see 40 above).
719
720 42. It seems that Perl allows any characters other than a closing parenthesis
721 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
722 has been changed to be the same.
723
724 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
725 as not to get warnings when autogen.sh is called. Also changed
726 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
727
728 44. To help people who use pcregrep to scan files containing exceedingly long
729 lines, the following changes have been made:
730
731 (a) The default value of the buffer size parameter has been increased from
732 8K to 20K. (The actual buffer used is three times this size.)
733
734 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
735 PCRE is built.
736
737 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
738 to be set at run time.
739
740 (d) Numerical values in pcregrep options can be followed by K or M, for
741 example --buffer-size=50K.
742
743 (e) If a line being scanned overflows pcregrep's buffer, an error is now
744 given and the return code is set to 2.
745
746 45. Add a pointer to the latest mark to the callout data block.
747
748 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
749 partial match of an empty string instead of no match. This was specific to
750 the use of ".".
751
752 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
753 complete match instead of a partial match. This bug was dependent on both
754 the PCRE_UTF8 and PCRE_DOTALL options being set.
755
756 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
757 starting byte set, because \b was not being ignored.
758
759
760 Version 8.12 15-Jan-2011
761 ------------------------
762
763 1. Fixed some typos in the markup of the man pages, and wrote a script that
764 checks for such things as part of the documentation building process.
765
766 2. On a big-endian 64-bit system, pcregrep did not correctly process the
767 --match-limit and --recursion-limit options (added for 8.11). In
768 particular, this made one of the standard tests fail. (The integer value
769 went into the wrong half of a long int.)
770
771 3. If the --colour option was given to pcregrep with -v (invert match), it
772 did strange things, either producing crazy output, or crashing. It should,
773 of course, ignore a request for colour when reporting lines that do not
774 match.
775
776 4. Another pcregrep bug caused similar problems if --colour was specified with
777 -M (multiline) and the pattern match finished with a line ending.
778
779 5. In pcregrep, when a pattern that ended with a literal newline sequence was
780 matched in multiline mode, the following line was shown as part of the
781 match. This seems wrong, so I have changed it.
782
783 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
784 the check for further matches in the same line (so they could be coloured)
785 to overrun the end of the current line. If another match was found, it was
786 incorrectly shown (and then shown again when found in the next line).
787
788 7. If pcregrep was compiled under Windows, there was a reference to the
789 function pcregrep_exit() before it was defined. I am assuming this was
790 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
791 reported by a user. I've moved the definition above the reference.
792
793
794 Version 8.11 10-Dec-2010
795 ------------------------
796
797 1. (*THEN) was not working properly if there were untried alternatives prior
798 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
799 backtracked to try for "b" instead of moving to the next alternative branch
800 at the same level (in this case, to look for "c"). The Perl documentation
801 is clear that when (*THEN) is backtracked onto, it goes to the "next
802 alternative in the innermost enclosing group".
803
804 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
805 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
806 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
807 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
808 (*THEN).
809
810 3. If \s appeared in a character class, it removed the VT character from
811 the class, even if it had been included by some previous item, for example
812 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
813 of \s, but is part of the POSIX "space" class.)
814
815 4. A partial match never returns an empty string (because you can always
816 match an empty string at the end of the subject); however the checking for
817 an empty string was starting at the "start of match" point. This has been
818 changed to the "earliest inspected character" point, because the returned
819 data for a partial match starts at this character. This means that, for
820 example, /(?<=abc)def/ gives a partial match for the subject "abc"
821 (previously it gave "no match").
822
823 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
824 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
825 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
826 has an implication that the given string is incomplete (because a partial
827 match is preferred over a full match). For this reason, these items now
828 give a partial match in this situation. [Aside: previously, the one case
829 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
830 match rather than a full match, which was wrong by the old rules, but is
831 now correct.]
832
833 6. There was a bug in the handling of #-introduced comments, recognized when
834 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
835 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
836 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
837 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
838 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
839 places in pcre_compile().
840
841 7. Related to (6) above, when pcre_compile() was skipping #-introduced
842 comments when looking ahead for named forward references to subpatterns,
843 the only newline sequence it recognized was NL. It now handles newlines
844 according to the set newline convention.
845
846 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
847 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
848 cater for a lack of strerror(). These oversights have been fixed.
849
850 9. Added --match-limit and --recursion-limit to pcregrep.
851
852 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
853
854 11. When the -o option was used, pcregrep was setting a return code of 1, even
855 when matches were found, and --line-buffered was not being honoured.
856
857 12. Added an optional parentheses number to the -o and --only-matching options
858 of pcregrep.
859
860 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
861 can match an empty string. The code to do it in pcretest and pcredemo
862 needed fixing:
863
864 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
865 only one byte after an empty string match just before CRLF (this case
866 just got forgotten; "any" and "anycrlf" were OK).
867
868 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
869 mode when an empty string match preceded an ASCII character followed by
870 a non-ASCII character. (The code for advancing by one character rather
871 than one byte was nonsense.)
872
873 (c) The pcredemo.c sample program did not have any code at all to handle
874 the cases when CRLF is a valid newline sequence.
875
876 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
877 as a starting offset was within the subject string. There is now a new
878 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
879 negative or greater than the length of the string. In order to test this,
880 pcretest is extended to allow the setting of negative starting offsets.
881
882 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
883 starting offset points to the beginning of a UTF-8 character was
884 unnecessarily clumsy. I tidied it up.
885
886 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
887 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
888
889 17. Nobody had reported that the --include_dir option, which was added in
890 release 7.7 should have been called --include-dir (hyphen, not underscore)
891 for compatibility with GNU grep. I have changed it to --include-dir, but
892 left --include_dir as an undocumented synonym, and the same for
893 --exclude-dir, though that is not available in GNU grep, at least as of
894 release 2.5.4.
895
896 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
897 characters from a string of bytes) have been redefined so as not to use
898 loops, in order to improve performance in some environments. At the same
899 time, I abstracted some of the common code into auxiliary macros to save
900 repetition (this should not affect the compiled code).
901
902 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
903 compile-time error is now given if \c is not followed by an ASCII
904 character, that is, a byte less than 128. (In EBCDIC mode, the code is
905 different, and any byte value is allowed.)
906
907 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
908 START_OPTIMIZE option, which is now allowed at compile time - but just
909 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
910 to pcregrep and other applications that have no direct access to PCRE
911 options. The new /Y option in pcretest sets this option when calling
912 pcre_compile().
913
914 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
915 back references. Groups containing recursive back references were forced to
916 be atomic by that change, but in the case of named groups, the amount of
917 memory required was incorrectly computed, leading to "Failed: internal
918 error: code overflow". This has been fixed.
919
920 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
921 pcretest.c, to avoid build problems in some Borland environments.
922
923
924 Version 8.10 25-Jun-2010
925 ------------------------
926
927 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
928 THEN.
929
930 2. (*ACCEPT) was not working when inside an atomic group.
931
932 3. Inside a character class, \B is treated as a literal by default, but
933 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
934 causes the error). The code is unchanged, but I tidied the documentation.
935
936 4. Inside a character class, PCRE always treated \R and \X as literals,
937 whereas Perl faults them if its -w option is set. I have changed PCRE so
938 that it faults them when PCRE_EXTRA is set.
939
940 5. Added support for \N, which always matches any character other than
941 newline. (It is the same as "." when PCRE_DOTALL is not set.)
942
943 6. When compiling pcregrep with newer versions of gcc which may have
944 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
945 declared with attribute warn_unused_result" were given. Just casting the
946 result to (void) does not stop the warnings; a more elaborate fudge is
947 needed. I've used a macro to implement this.
948
949 7. Minor change to pcretest.c to avoid a compiler warning.
950
951 8. Added four artifical Unicode properties to help with an option to make
952 \s etc use properties (see next item). The new properties are: Xan
953 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
954
955 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
956 use Unicode properties. (*UCP) at the start of a pattern can be used to set
957 this option. Modified pcretest to add /W to test this facility. Added
958 REG_UCP to make it available via the POSIX interface.
959
960 10. Added --line-buffered to pcregrep.
961
962 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
963 studied, and the match started with a letter with a code point greater than
964 127 whose first byte was different to the first byte of the other case of
965 the letter, the other case of this starting letter was not recognized
966 (#976).
967
968 12. If a pattern that was studied started with a repeated Unicode property
969 test, for example, \p{Nd}+, there was the theoretical possibility of
970 setting up an incorrect bitmap of starting bytes, but fortunately it could
971 not have actually happened in practice until change 8 above was made (it
972 added property types that matched character-matching opcodes).
973
974 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
975 possible starting bytes for non-anchored patterns.
976
977 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
978 \R, and also a number of cases that involve Unicode properties, both
979 explicit and implicit when PCRE_UCP is set.
980
981 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
982 input, it could crash or give wrong results if characters with values
983 greater than 0xc0 were present in the subject string. (Detail: it assumed
984 UTF-8 input when processing these items.)
985
986 16. Added a lot of (int) casts to avoid compiler warnings in systems where
987 size_t is 64-bit (#991).
988
989 17. Added a check for running out of memory when PCRE is compiled with
990 --disable-stack-for-recursion (#990).
991
992 18. If the last data line in a file for pcretest does not have a newline on
993 the end, a newline was missing in the output.
994
995 19. The default pcre_chartables.c file recognizes only ASCII characters (values
996 less than 128) in its various bitmaps. However, there is a facility for
997 generating tables according to the current locale when PCRE is compiled. It
998 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
999 characters, are recognized by isspace() and therefore were getting set in
1000 these tables, and indeed these tables seem to approximate to ISO 8859. This
1001 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1002 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1003 which of course cannot start UTF-8 characters. I have changed the code so
1004 that only real ASCII characters (less than 128) and the correct starting
1005 bytes for UTF-8 encodings are set for characters greater than 127 when in
1006 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1007 altogether.)
1008
1009 20. Added the /T option to pcretest so as to be able to run tests with non-
1010 standard character tables, thus making it possible to include the tests
1011 used for 19 above in the standard set of tests.
1012
1013 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1014 reference to a subpattern the other side of a comment that contains an
1015 opening parenthesis caused either an internal compiling error, or a
1016 reference to the wrong subpattern.
1017
1018
1019 Version 8.02 19-Mar-2010
1020 ------------------------
1021
1022 1. The Unicode data tables have been updated to Unicode 5.2.0.
1023
1024 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1025 configured.
1026
1027 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1028 original author of that file, following a query about its status.
1029
1030 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1031 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1032
1033 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1034 quantifier applied to a forward-referencing subroutine call, could compile
1035 incorrect code or give the error "internal error: previously-checked
1036 referenced subpattern not found".
1037
1038 6. Both MS Visual Studio and Symbian OS have problems with initializing
1039 variables to point to external functions. For these systems, therefore,
1040 pcre_malloc etc. are now initialized to local functions that call the
1041 relevant global functions.
1042
1043 7. There were two entries missing in the vectors called coptable and poptable
1044 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1045 I've fixed the data, and added a kludgy way of testing at compile time that
1046 the lengths are correct (equal to the number of opcodes).
1047
1048 8. Following on from 7, I added a similar kludge to check the length of the
1049 eint vector in pcreposix.c.
1050
1051 9. Error texts for pcre_compile() are held as one long string to avoid too
1052 much relocation at load time. To find a text, the string is searched,
1053 counting zeros. There was no check for running off the end of the string,
1054 which could happen if a new error number was added without updating the
1055 string.
1056
1057 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1058
1059 11. \K was not working if it appeared in an atomic group or in a group that
1060 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1061 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1062 the assertion is positive, but not if it is negative.
1063
1064 12. Change 11 fortuitously reduced the size of the stack frame used in the
1065 "match()" function of pcre_exec.c by one pointer. Forthcoming
1066 implementation of support for (*MARK) will need an extra pointer on the
1067 stack; I have reserved it now, so that the stack frame size does not
1068 decrease.
1069
1070 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1071 item in branch that calls a recursion is a subroutine call - as in the
1072 second branch in the above example - was incorrectly given the compile-
1073 time error "recursive call could loop indefinitely" because pcre_compile()
1074 was not correctly checking the subroutine for matching a non-empty string.
1075
1076 14. The checks for overrunning compiling workspace could trigger after an
1077 overrun had occurred. This is a "should never occur" error, but it can be
1078 triggered by pathological patterns such as hundreds of nested parentheses.
1079 The checks now trigger 100 bytes before the end of the workspace.
1080
1081 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1082
1083
1084 Version 8.01 19-Jan-2010
1085 ------------------------
1086
1087 1. If a pattern contained a conditional subpattern with only one branch (in
1088 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1089 computed the wrong minimum data length (which is of course zero for such
1090 subpatterns). This could cause incorrect "no match" results.
1091
1092 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1093 the pattern is reset in the first branch, pcre_compile() failed with
1094 "internal error: code overflow at offset...". This happened only when
1095 the reset was to the original external option setting. (An optimization
1096 abstracts leading options settings into an external setting, which was the
1097 cause of this.)
1098
1099 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1100 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1101 assertion pattern did not match (meaning that the assertion was true), it
1102 was incorrectly treated as false if the SKIP had been reached during the
1103 matching. This also applied to assertions used as conditions.
1104
1105 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1106 assertion subpattern, including such a pattern used as a condition,
1107 unpredictable results occurred, instead of the error return
1108 PCRE_ERROR_DFA_UITEM.
1109
1110 5. The C++ GlobalReplace function was not working like Perl for the special
1111 situation when an empty string is matched. It now does the fancy magic
1112 stuff that is necessary.
1113
1114 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1115 removed. (These were left over from very, very early versions of PCRE.)
1116
1117 7. Some cosmetic changes to the code to make life easier when compiling it
1118 as part of something else:
1119
1120 (a) Change DEBUG to PCRE_DEBUG.
1121
1122 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1123 called "current" as "current_branch", to prevent a collision with the
1124 Linux macro when compiled as a kernel module.
1125
1126 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1127 prevent a collision with the Linux macro when compiled as a kernel
1128 module.
1129
1130 8. In pcre_compile() there are some checks for integer overflows that used to
1131 cast potentially large values to (double). This has been changed to that
1132 when building, a check for int64_t is made, and if it is found, it is used
1133 instead, thus avoiding the use of floating point arithmetic. (There is no
1134 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1135 double.
1136
1137 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1138 2005 (difference between two addresses compared to an unsigned value).
1139
1140 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1141 custom one, because of the following reported problem in Windows:
1142
1143 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1144 under Win32.
1145 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1146 therefore missing the function definition.
1147 - The compiler thus generates a "C" signature for the test function.
1148 - The linker fails to find the "C" function.
1149 - PCRE fails to configure if asked to do so against libbz2.
1150
1151 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1152 messages were output:
1153
1154 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1155 rerunning libtoolize, to keep the correct libtool macros in-tree.
1156 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1157
1158 I have done both of these things.
1159
1160 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1161 most of the time, it *can* run out if it is given a pattern that contains a
1162 runaway infinite recursion. I updated the discussion in the pcrestack man
1163 page.
1164
1165 13. Now that we have gone to the x.xx style of version numbers, the minor
1166 version may start with zero. Using 08 or 09 is a bad idea because users
1167 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1168 interpreted as invalid octal numbers. I've updated the previous comment in
1169 configure.ac, and also added a check that gives an error if 08 or 09 are
1170 used.
1171
1172 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1173 causing partial matching to fail when the end of the subject matched \W
1174 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1175
1176 15. There were some discrepancies between the declarations in pcre_internal.h
1177 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1178 their definitions. The declarations used "const uschar *" and the
1179 definitions used USPTR. Even though USPTR is normally defined as "const
1180 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1181 reported that: "This difference in casting confuses some C++ compilers, for
1182 example, SunCC recognizes above declarations as different functions and
1183 generates broken code for hbpcre." I have changed the declarations to use
1184 USPTR.
1185
1186 16. GNU libtool is named differently on some systems. The autogen.sh script now
1187 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1188 (FreeBSD).
1189
1190 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1191 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1192 comment: "Figure out how to create a longlong from a string: strtoll and
1193 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1194 instance, but it only takes 2 args instead of 3!"
1195
1196 18. A subtle bug concerned with back references has been fixed by a change of
1197 specification, with a corresponding code fix. A pattern such as
1198 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1199 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1200 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1201 same bug. Such groups have to be quantified to be useful, or contained
1202 inside another quantified group. (If there's no repetition, the reference
1203 can never match.) The problem arises because, having left the group and
1204 moved on to the rest of the pattern, a later failure that backtracks into
1205 the group uses the captured value from the final iteration of the group
1206 rather than the correct earlier one. I have fixed this in PCRE by forcing
1207 any group that contains a reference to itself to be an atomic group; that
1208 is, there cannot be any backtracking into it once it has completed. This is
1209 similar to recursive and subroutine calls.
1210
1211
1212 Version 8.00 19-Oct-09
1213 ----------------------
1214
1215 1. The table for translating pcre_compile() error codes into POSIX error codes
1216 was out-of-date, and there was no check on the pcre_compile() error code
1217 being within the table. This could lead to an OK return being given in
1218 error.
1219
1220 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1221 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1222 in a Windows environment.
1223
1224 3. The pcregrep --count option prints the count for each file even when it is
1225 zero, as does GNU grep. However, pcregrep was also printing all files when
1226 --files-with-matches was added. Now, when both options are given, it prints
1227 counts only for those files that have at least one match. (GNU grep just
1228 prints the file name in this circumstance, but including the count seems
1229 more useful - otherwise, why use --count?) Also ensured that the
1230 combination -clh just lists non-zero counts, with no names.
1231
1232 4. The long form of the pcregrep -F option was incorrectly implemented as
1233 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1234 but it seems right to fix it, and I didn't think it was worth preserving
1235 the old behaviour.
1236
1237 5. The command line items --regex=pattern and --regexp=pattern were not
1238 recognized by pcregrep, which required --regex pattern or --regexp pattern
1239 (with a space rather than an '='). The man page documented the '=' forms,
1240 which are compatible with GNU grep; these now work.
1241
1242 6. No libpcreposix.pc file was created for pkg-config; there was just
1243 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1244
1245 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1246 when UCP support is not needed, by modifying the Python script that
1247 generates it from Unicode data files. This should not matter if the module
1248 is correctly used as a library, but I received one complaint about 50K of
1249 unwanted data. My guess is that the person linked everything into his
1250 program rather than using a library. Anyway, it does no harm.
1251
1252 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1253 was a minimum greater than 1 for a wide character in a possessive
1254 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1255 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1256 character. Chaos in the form of incorrect output or a compiling loop could
1257 result.
1258
1259 9. The restrictions on what a pattern can contain when partial matching is
1260 requested for pcre_exec() have been removed. All patterns can now be
1261 partially matched by this function. In addition, if there are at least two
1262 slots in the offset vector, the offset of the earliest inspected character
1263 for the match and the offset of the end of the subject are set in them when
1264 PCRE_ERROR_PARTIAL is returned.
1265
1266 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1267 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1268 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1269 and may be more useful for multi-segment matching.
1270
1271 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1272 used to be given if ever the end of the subject was reached; now it is
1273 given only if matching could not proceed because another character was
1274 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1275 string "Z", which now yields "no match" instead of "partial match". In the
1276 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1277 final character ended with (*FAIL).
1278
1279 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1280 if the pattern had a "must contain" character that was already found in the
1281 earlier partial match, unless partial matching was again requested. For
1282 example, with the pattern /dog.(body)?/, the "must contain" character is
1283 "g". If the first part-match was for the string "dog", restarting with
1284 "sbody" failed. This bug has been fixed.
1285
1286 13. The string returned by pcre_dfa_exec() after a partial match has been
1287 changed so that it starts at the first inspected character rather than the
1288 first character of the match. This makes a difference only if the pattern
1289 starts with a lookbehind assertion or \b or \B (\K is not supported by
1290 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1291 matching functions compatible, and I think it's the right thing to do.
1292
1293 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1294 so that the demonstration program is easily available in environments where
1295 PCRE has not been installed from source.
1296
1297 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1298 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1299 library.
1300
1301 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1302 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1303 is not the first non-POSIX option to be added. Clearly some people find
1304 these options useful.
1305
1306 17. If a caller to the POSIX matching function regexec() passes a non-zero
1307 value for nmatch with a NULL value for pmatch, the value of
1308 nmatch is forced to zero.
1309
1310 18. RunGrepTest did not have a test for the availability of the -u option of
1311 the diff command, as RunTest does. It now checks in the same way as
1312 RunTest, and also checks for the -b option.
1313
1314 19. If an odd number of negated classes containing just a single character
1315 interposed, within parentheses, between a forward reference to a named
1316 subpattern and the definition of the subpattern, compilation crashed with
1317 an internal error, complaining that it could not find the referenced
1318 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1319 [The bug was that it was starting one character too far in when skipping
1320 over the character class, thus treating the ] as data rather than
1321 terminating the class. This meant it could skip too much.]
1322
1323 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1324 /g option in pcretest when the pattern contains \K, which makes it possible
1325 to have an empty string match not at the start, even when the pattern is
1326 anchored. Updated pcretest and pcredemo to use this option.
1327
1328 21. If the maximum number of capturing subpatterns in a recursion was greater
1329 than the maximum at the outer level, the higher number was returned, but
1330 with unset values at the outer level. The correct (outer level) value is
1331 now given.
1332
1333 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1334 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1335 make it do so. The string so far is captured, making this feature
1336 compatible with Perl.
1337
1338 23. The tests have been re-organized, adding tests 11 and 12, to make it
1339 possible to check the Perl 5.10 features against Perl 5.10.
1340
1341 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1342 pattern matches a fixed length string. PCRE did not allow this; now it
1343 does. Neither allows recursion.
1344
1345 25. I finally figured out how to implement a request to provide the minimum
1346 length of subject string that was needed in order to match a given pattern.
1347 (It was back references and recursion that I had previously got hung up
1348 on.) This code has now been added to pcre_study(); it finds a lower bound
1349 to the length of subject needed. It is not necessarily the greatest lower
1350 bound, but using it to avoid searching strings that are too short does give
1351 some useful speed-ups. The value is available to calling programs via
1352 pcre_fullinfo().
1353
1354 26. While implementing 25, I discovered to my embarrassment that pcretest had
1355 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1356 study optimizations had never been tested with that matching function.
1357 Oops. What is worse, even when it was passed study data, there was a bug in
1358 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1359 were also very few tests of studied patterns with pcre_dfa_exec().
1360
1361 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1362 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1363 on the other side of the coin, they are no longer allowed to have different
1364 names, because these cannot be distinguished in PCRE, and this has caused
1365 confusion. (This is a difference from Perl.)
1366
1367 28. When duplicate subpattern names are present (necessarily with different
1368 numbers, as required by 27 above), and a test is made by name in a
1369 conditional pattern, either for a subpattern having been matched, or for
1370 recursion in such a pattern, all the associated numbered subpatterns are
1371 tested, and the overall condition is true if the condition is true for any
1372 one of them. This is the way Perl works, and is also more like the way
1373 testing by number works.
1374
1375
1376 Version 7.9 11-Apr-09
1377 ---------------------
1378
1379 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1380 (pcretest), all targets were linked against these libraries. This included
1381 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1382 libraries. This caused unwanted dependencies to be created. This problem
1383 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1384 pcretest is linked with readline.
1385
1386 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1387 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1388 moved outside it again, because FALSE and TRUE are already defined in AIX,
1389 but BOOL is not.
1390
1391 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1392 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1393
1394 4. The pcregrep documentation said spaces were inserted as well as colons (or
1395 hyphens) following file names and line numbers when outputting matching
1396 lines. This is not true; no spaces are inserted. I have also clarified the
1397 wording for the --colour (or --color) option.
1398
1399 5. In pcregrep, when --colour was used with -o, the list of matching strings
1400 was not coloured; this is different to GNU grep, so I have changed it to be
1401 the same.
1402
1403 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1404 each matching line was coloured. Now it goes on to look for further matches
1405 of any of the test patterns, which is the same behaviour as GNU grep.
1406
1407 7. A pattern that could match an empty string could cause pcregrep to loop; it
1408 doesn't make sense to accept an empty string match in pcregrep, so I have
1409 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1410 seems to be how GNU grep behaves.
1411
1412 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1413 start or after a newline", because the conditional assertion was not being
1414 correctly handled. The rule now is that both the assertion and what follows
1415 in the first alternative must satisfy the test.
1416
1417 9. If auto-callout was enabled in a pattern with a conditional group whose
1418 condition was an assertion, PCRE could crash during matching, both with
1419 pcre_exec() and pcre_dfa_exec().
1420
1421 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1422 used for matching.
1423
1424 11. Unicode property support in character classes was not working for
1425 characters (bytes) greater than 127 when not in UTF-8 mode.
1426
1427 12. Added the -M command line option to pcretest.
1428
1429 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1430
1431 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1432
1433 16. Added comments and documentation about mis-use of no_arg in the C++
1434 wrapper.
1435
1436 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1437 from Martin Jerabek that uses macro names for all relevant character and
1438 string constants.
1439
1440 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1441 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1442 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1443 these, but not everybody uses configure.
1444
1445 19. A conditional group that had only one branch was not being correctly
1446 recognized as an item that could match an empty string. This meant that an
1447 enclosing group might also not be so recognized, causing infinite looping
1448 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1449 with the subject "ab", where knowledge that the repeated group can match
1450 nothing is needed in order to break the loop.
1451
1452 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1453 exec(), but without supplying a callout function, matching went wrong.
1454
1455 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1456 leak if the size of the offset vector was greater than 30. When the vector
1457 is smaller, the saved offsets during recursion go onto a local stack
1458 vector, but for larger vectors malloc() is used. It was failing to free
1459 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1460 error, in fact).
1461
1462 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1463 heapframe that is used only when UTF-8 support is enabled. This caused no
1464 problem, but was untidy.
1465
1466 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1467 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1468 included within another project.
1469
1470 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1471 slightly modified by me:
1472
1473 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1474 not building pcregrep.
1475
1476 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1477 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1478
1479 25. Forward references, both numeric and by name, in patterns that made use of
1480 duplicate group numbers, could behave incorrectly or give incorrect errors,
1481 because when scanning forward to find the reference group, PCRE was not
1482 taking into account the duplicate group numbers. A pattern such as
1483 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1484
1485 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1486 the feature of a custom pointer more persuasive (as requested by a user).
1487
1488 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1489 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1490 pre-defined.
1491
1492 28. Added support for (*UTF8) at the start of a pattern.
1493
1494 29. Arrange for flags added by the "release type" setting in CMake to be shown
1495 in the configuration summary.
1496
1497
1498 Version 7.8 05-Sep-08
1499 ---------------------
1500
1501 1. Replaced UCP searching code with optimized version as implemented for Ad
1502 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1503 stage table and inline lookup instead of a function, giving speed ups of 2
1504 to 5 times on some simple patterns that I tested. Permission was given to
1505 distribute the MultiStage2.py script that generates the tables (it's not in
1506 the tarball, but is in the Subversion repository).
1507
1508 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1509 scripts.
1510
1511 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1512 a group with a zero qualifier. The result of the study could be incorrect,
1513 or the function might crash, depending on the pattern.
1514
1515 4. Caseless matching was not working for non-ASCII characters in back
1516 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1517 It now works when Unicode Property Support is available.
1518
1519 5. In pcretest, an escape such as \x{de} in the data was always generating
1520 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1521 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1522 truncation.
1523
1524 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1525
1526 7. Added two (int) casts to pcregrep when printing the difference of two
1527 pointers, in case they are 64-bit values.
1528
1529 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1530 test 2 if it fails.
1531
1532 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1533 and a #define of that name to empty if it is not externally set. This is to
1534 allow users of MSVC to set it if necessary.
1535
1536 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1537 the convenience functions in the pcre_get.c source file.
1538
1539 11. An option change at the start of a pattern that had top-level alternatives
1540 could cause overwriting and/or a crash. This command provoked a crash in
1541 some environments:
1542
1543 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1544
1545 This potential security problem was recorded as CVE-2008-2371.
1546
1547 12. For a pattern where the match had to start at the beginning or immediately
1548 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1549 pcre_dfa_exec() could read past the end of the passed subject if there was
1550 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1551 pcretest so that it places the subject at the end of its malloc-ed buffer.
1552
1553 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1554 exec() might read past the end of the data buffer in UTF-8 mode.
1555
1556 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1557 the data contained the byte 0x85 as part of a UTF-8 character within its
1558 first line. This applied both to normal and DFA matching.
1559
1560 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1561 /^[^d]*?$/8 failed to match "abc".
1562
1563 16. Added a missing copyright notice to pcrecpp_internal.h.
1564
1565 17. Make it more clear in the documentation that values returned from
1566 pcre_exec() in ovector are byte offsets, not character counts.
1567
1568 18. Tidied a few places to stop certain compilers from issuing warnings.
1569
1570 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1571 supplied by Stefan Weber. I made a further small update for 7.8 because
1572 there is a change of source arrangements: the pcre_searchfuncs.c module is
1573 replaced by pcre_ucd.c.
1574
1575
1576 Version 7.7 07-May-08
1577 ---------------------
1578
1579 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1580 a string to a long long, pretend we don't even have a long long." This is
1581 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1582
1583 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1584 pre-7.6 versions, which defined a global no_arg variable instead of putting
1585 it in the RE class. (See also #8 below.)
1586
1587 3. Remove a line of dead code, identified by coverity and reported by Nuno
1588 Lopes.
1589
1590 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1591
1592 (1) The include/exclude patterns were being applied to the whole pathnames
1593 of files, instead of just to the final components.
1594
1595 (2) If there was more than one level of directory, the subdirectories were
1596 skipped unless they satisfied the include/exclude conditions. This is
1597 inconsistent with GNU grep (and could even be seen as contrary to the
1598 pcregrep specification - which I improved to make it absolutely clear).
1599 The action now is always to scan all levels of directory, and just
1600 apply the include/exclude patterns to regular files.
1601
1602 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1603 --exclude_dir in the tests to avoid scanning .svn directories.
1604
1605 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1606 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1607 doesn't support NULs in patterns.
1608
1609 7. Added some missing "const"s to declarations of static tables in
1610 pcre_compile.c and pcre_dfa_exec.c.
1611
1612 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1613 caused by fix #2 above. (Subsequently also a second patch to fix the
1614 first patch. And a third patch - this was a messy problem.)
1615
1616 9. Applied Craig's patch to remove the use of push_back().
1617
1618 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1619 matching function regexec().
1620
1621 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1622 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1623 references. PCRE supports relative numbers with this syntax (I don't think
1624 Oniguruma does).
1625
1626 12. Previously, a group with a zero repeat such as (...){0} was completely
1627 omitted from the compiled regex. However, this means that if the group
1628 was called as a subroutine from elsewhere in the pattern, things went wrong
1629 (an internal error was given). Such groups are now left in the compiled
1630 pattern, with a new opcode that causes them to be skipped at execution
1631 time.
1632
1633 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1634 to the way PCRE behaves:
1635
1636 (a) A lone ] character is dis-allowed (Perl treats it as data).
1637
1638 (b) A back reference to an unmatched subpattern matches an empty string
1639 (Perl fails the current match path).
1640
1641 (c) A data ] in a character class must be notated as \] because if the
1642 first data character in a class is ], it defines an empty class. (In
1643 Perl it is not possible to have an empty class.) The empty class []
1644 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1645 The negative empty class [^] matches any one character, independently
1646 of the DOTALL setting.
1647
1648 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1649 non-existent subpattern following a character class starting with ']' and
1650 containing () gave an internal compiling error instead of "reference to
1651 non-existent subpattern". Fortunately, when the pattern did exist, the
1652 compiled code was correct. (When scanning forwards to check for the
1653 existencd of the subpattern, it was treating the data ']' as terminating
1654 the class, so got the count wrong. When actually compiling, the reference
1655 was subsequently set up correctly.)
1656
1657 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1658 it was being rejected as not supported by pcre_dfa_exec(), even though
1659 other assertions are supported. I have made pcre_dfa_exec() support
1660 (*FAIL).
1661
1662 16. The implementation of 13c above involved the invention of a new opcode,
1663 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1664 cannot be changed at match time, I realized I could make a small
1665 improvement to matching performance by compiling OP_ALLANY instead of
1666 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1667 on the OP_ANY path.
1668
1669 17. Compiling pcretest on Windows with readline support failed without the
1670 following two fixes: (1) Make the unistd.h include conditional on
1671 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1672
1673 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1674 ncurses library to be included for pcretest when ReadLine support is
1675 requested, but also to allow for it to be overridden. This patch came from
1676 Daniel Bergström.
1677
1678 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1679 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1680 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1681 spotting this.
1682
1683
1684 Version 7.6 28-Jan-08
1685 ---------------------
1686
1687 1. A character class containing a very large number of characters with
1688 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1689 overflow.
1690
1691 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1692 HAVE_LONG_LONG is not defined.
1693
1694 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1695 bring them up to date and include new features. This patch includes:
1696
1697 - Fixed PH's badly added libz and libbz2 support.
1698 - Fixed a problem with static linking.
1699 - Added pcredemo. [But later removed - see 7 below.]
1700 - Fixed dftables problem and added an option.
1701 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1702 HAVE_LONG_LONG.
1703 - Added readline support for pcretest.
1704 - Added an listing of the option settings after cmake has run.
1705
1706 4. A user submitted a patch to Makefile that makes it easy to create
1707 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1708 Makefile.am that cause it to include this special target, without
1709 affecting anything else. Note that the same mingw target plus all
1710 the other distribution libraries and programs are now supported
1711 when configuring with CMake (see 6 below) instead of with
1712 Configure/Make.
1713
1714 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1715 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1716 exported in the Windows port". It has not yet been confirmed that the patch
1717 solves the problem, but it does no harm.
1718
1719 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1720 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1721 with CMake, and also correct the comment about stack recursion.
1722
1723 7. Remove the automatic building of pcredemo from the ./configure system and
1724 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1725 of a program that users should build themselves after PCRE is installed, so
1726 building it automatically is not really right. What is more, it gave
1727 trouble in some build environments.
1728
1729 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1730
1731
1732 Version 7.5 10-Jan-08
1733 ---------------------
1734
1735 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1736 values in parens when parsing an RE using the C++ wrapper."
1737
1738 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1739 Characters greater than 255 were excluded from the class instead of being
1740 included.
1741
1742 3. The same bug as (2) above applied to negated POSIX classes such as
1743 [:^space:].
1744
1745 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1746 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1747 I have changed it.
1748
1749 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1750 first named subpattern) and a construct such as (?&a) would reference the
1751 first named subpattern whose name started with "a" (in other words, the
1752 length check was missing). Both these problems are fixed. "Subpattern name
1753 expected" is now given for (?&) (a zero-length name), and this patch also
1754 makes it give the same error for \k'' (previously it complained that that
1755 was a reference to a non-existent subpattern).
1756
1757 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1758 this is right because (?- can be followed by option settings as well as by
1759 digits. I have, however, made the messages clearer.
1760
1761 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1762 than the number used in the conditional) now cause a compile-time error.
1763 This is actually not compatible with Perl, which accepts such patterns, but
1764 treats the conditional as always being FALSE (as PCRE used to), but it
1765 seems to me that giving a diagnostic is better.
1766
1767 8. Change "alphameric" to the more common word "alphanumeric" in comments
1768 and messages.
1769
1770 9. Fix two occurrences of "backslash" in comments that should have been
1771 "backspace".
1772
1773 10. Remove two redundant lines of code that can never be obeyed (their function
1774 was moved elsewhere).
1775
1776 11. The program that makes PCRE's Unicode character property table had a bug
1777 which caused it to generate incorrect table entries for sequences of
1778 characters that have the same character type, but are in different scripts.
1779 It amalgamated them into a single range, with the script of the first of
1780 them. In other words, some characters were in the wrong script. There were
1781 thirteen such cases, affecting characters in the following ranges:
1782
1783 U+002b0 - U+002c1
1784 U+0060c - U+0060d
1785 U+0061e - U+00612
1786 U+0064b - U+0065e
1787 U+0074d - U+0076d
1788 U+01800 - U+01805
1789 U+01d00 - U+01d77
1790 U+01d9b - U+01dbf
1791 U+0200b - U+0200f
1792 U+030fc - U+030fe
1793 U+03260 - U+0327f
1794 U+0fb46 - U+0fbb1
1795 U+10450 - U+1049d
1796
1797 12. The -o option (show only the matching part of a line) for pcregrep was not
1798 compatible with GNU grep in that, if there was more than one match in a
1799 line, it showed only the first of them. It now behaves in the same way as
1800 GNU grep.
1801
1802 13. If the -o and -v options were combined for pcregrep, it printed a blank
1803 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1804 does the same. The return code can be used to tell if there were any
1805 non-matching lines.
1806
1807 14. Added --file-offsets and --line-offsets to pcregrep.
1808
1809 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1810 infinitely looping recursion. The bug was that positive lookaheads were not
1811 being skipped when checking for a possible empty match (negative lookaheads
1812 and both kinds of lookbehind were skipped).
1813
1814 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1815 inclusion of <windows.h> to before rather than after the definition of
1816 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1817
1818 17. Specifying a possessive quantifier with a specific limit for a Unicode
1819 character property caused pcre_compile() to compile bad code, which led at
1820 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1821 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1822 caused the error; without that there was no problem.
1823
1824 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1825
1826 19. Added --enable-pcretest-libreadline.
1827
1828 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1829 RE::GlobalReplace(). As a result, the number of replacements returned was
1830 double what it should be. I removed one of the increments, but Craig sent a
1831 later patch that removed the other one (the right fix) and added unit tests
1832 that check the return values (which was not done before).
1833
1834 21. Several CMake things:
1835
1836 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1837 the names libpcre and libpcreposix, not just pcre and pcreposix.
1838
1839 (2) The above change means that pcretest and pcregrep are now correctly
1840 linked with the newly-built libraries, not previously installed ones.
1841
1842 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1843
1844 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1845 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1846 UTF-8 newline character). The key issue is that the pattern starts .*;
1847 this means that the match must be either at the beginning, or after a
1848 newline. The bug was in the code for advancing after a failed match and
1849 checking that the new position followed a newline. It was not taking
1850 account of UTF-8 characters correctly.
1851
1852 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1853 character classes. PCRE was not treating the sequence [:...:] as a
1854 character class unless the ... were all letters. Perl, however, seems to
1855 allow any characters between [: and :], though of course it rejects as
1856 unknown any "names" that contain non-letters, because all the known class
1857 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1858 for example, whereas PCRE did not - it did not recognize a POSIX character
1859 class. This seemed a bit dangerous, so the code has been changed to be
1860 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1861 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1862 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1863 Perl does, and where it didn't before.
1864
1865 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1866 Windows environments %n is disabled by default.
1867
1868
1869 Version 7.4 21-Sep-07
1870 ---------------------
1871
1872 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1873 means that a class such as [\s] counted as "explicit reference to CR or
1874 LF". That isn't really right - the whole point of the change was to try to
1875 help when there was an actual mention of one of the two characters. So now
1876 the change happens only if \r or \n (or a literal CR or LF) character is
1877 encountered.
1878
1879 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1880 of both had grown to the point where there were only 3 bits left.
1881 Fortunately, there was spare space in the data structure, and so I have
1882 moved the internal flags into a new 16-bit field to free up more option
1883 bits.
1884
1885 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1886 but did not set the internal JCHANGED flag - either of these is enough to
1887 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1888 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1889 start sets both bits.
1890
1891 4. Added options (at build time, compile time, exec time) to change \R from
1892 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1893
1894 5. doc/pcresyntax.html was missing from the distribution.
1895
1896 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1897 compatibility, even though it is no longer used.
1898
1899 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1900 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1901 windows.h file is present (where different names are used). [This was
1902 reversed later after testing - see 16 below.]
1903
1904 8. Changed all #include <config.h> to #include "config.h". There were also
1905 some further <pcre.h> cases that I changed to "pcre.h".
1906
1907 9. When pcregrep was used with the --colour option, it missed the line ending
1908 sequence off the lines that it output.
1909
1910 10. It was pointed out to me that arrays of string pointers cause lots of
1911 relocations when a shared library is dynamically loaded. A technique of
1912 using a single long string with a table of offsets can drastically reduce
1913 these. I have refactored PCRE in four places to do this. The result is
1914 dramatic:
1915
1916 Originally: 290
1917 After changing UCP table: 187
1918 After changing error message table: 43
1919 After changing table of "verbs" 36
1920 After changing table of Posix names 22
1921
1922 Thanks to the folks working on Gregex for glib for this insight.
1923
1924 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1925 unicode-properties was also set.
1926
1927 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1928
1929 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1930 checked only for CRLF.
1931
1932 14. Added casts to pcretest.c to avoid compiler warnings.
1933
1934 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1935
1936 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1937 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1938 entirely. This removes changes made in 7 above.
1939
1940 17. The CMake files have been updated, and there is now more information about
1941 building with CMake in the NON-UNIX-USE document.
1942
1943
1944 Version 7.3 28-Aug-07
1945 ---------------------
1946
1947 1. In the rejigging of the build system that eventually resulted in 7.1, the
1948 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1949 brackets there is not right, since it causes compilers to look for an
1950 installed pcre.h, not the version that is in the source that is being
1951 compiled (which of course may be different). I have changed it back to:
1952
1953 #include "pcre.h"
1954
1955 I have a vague recollection that the change was concerned with compiling in
1956 different directories, but in the new build system, that is taken care of
1957 by the VPATH setting the Makefile.
1958
1959 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1960 when the subject happened to end in the byte 0x85 (e.g. if the last
1961 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1962 characters but of course it shouldn't be taken as a newline when it is part
1963 of another character. The bug was that, for an unlimited repeat of . in
1964 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1965 characters when looking for a newline.
1966
1967 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1968
1969 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1970 in debug output.
1971
1972 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1973 long printing in the pcrecpp unittest when running under MinGW.
1974
1975 6. ESC_K was left out of the EBCDIC table.
1976
1977 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1978 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1979 limit also applies to "virtual nesting" when a pattern is recursive, and in
1980 this case 1000 isn't so big. I have been able to remove this limit at the
1981 expense of backing off one optimization in certain circumstances. Normally,
1982 when pcre_exec() would call its internal match() function recursively and
1983 immediately return the result unconditionally, it uses a "tail recursion"
1984 feature to save stack. However, when a subpattern that can match an empty
1985 string has an unlimited repetition quantifier, it no longer makes this
1986 optimization. That gives it a stack frame in which to save the data for
1987 checking that an empty string has been matched. Previously this was taken
1988 from the 1000-entry workspace that had been reserved. So now there is no
1989 explicit limit, but more stack is used.
1990
1991 8. Applied Daniel's patches to solve problems with the import/export magic
1992 syntax that is required for Windows, and which was going wrong for the
1993 pcreposix and pcrecpp parts of the library. These were overlooked when this
1994 problem was solved for the main library.
1995
1996 9. There were some crude static tests to avoid integer overflow when computing
1997 the size of patterns that contain repeated groups with explicit upper
1998 limits. As the maximum quantifier is 65535, the maximum group length was
1999 set at 30,000 so that the product of these two numbers did not overflow a
2000 32-bit integer. However, it turns out that people want to use groups that
2001 are longer than 30,000 bytes (though not repeat them that many times).
2002 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2003 made it possible to implement the integer overflow checks in a much more
2004 dynamic way, which I have now done. The artificial limitation on group
2005 length has been removed - we now have only the limit on the total length of
2006 the compiled pattern, which depends on the LINK_SIZE setting.
2007
2008 10. Fixed a bug in the documentation for get/copy named substring when
2009 duplicate names are permitted. If none of the named substrings are set, the
2010 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2011 empty string.
2012
2013 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2014 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2015 because the ] is interpreted as the first data character and the
2016 terminating ] is not found. PCRE has been made compatible with Perl in this
2017 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2018 cause memory overwriting.
2019
2020 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2021 string has been matched (to stop an infinite loop). It was not recognizing
2022 a conditional subpattern that could match an empty string if that
2023 subpattern was within another subpattern. For example, it looped when
2024 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2025 condition was not nested. This bug has been fixed.
2026
2027 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2028 past the start of the subject in the presence of bytes with the top bit
2029 set, for example "\x8aBCD".
2030
2031 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2032 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2033
2034 14. Optimized (?!) to (*FAIL).
2035
2036 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2037 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2038 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2039 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2040 does: it's just the validity check that is more restrictive.
2041
2042 16. Inserted checks for integer overflows during escape sequence (backslash)
2043 processing, and also fixed erroneous offset values for syntax errors during
2044 backslash processing.
2045
2046 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2047 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2048
2049 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2050 caused an overrun.
2051
2052 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2053 something other than just ASCII characters) inside a group that had an
2054 unlimited repeat caused a loop at compile time (while checking to see
2055 whether the group could match an empty string).
2056
2057 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2058 [\P{Any}] did so. (Error in the code for printing property names.)
2059
2060 21. An orphan \E inside a character class could cause a crash.
2061
2062 22. A repeated capturing bracket such as (A)? could cause a wild memory
2063 reference during compilation.
2064
2065 23. There are several functions in pcre_compile() that scan along a compiled
2066 expression for various reasons (e.g. to see if it's fixed length for look
2067 behind). There were bugs in these functions when a repeated \p or \P was
2068 present in the pattern. These operators have additional parameters compared
2069 with \d, etc, and these were not being taken into account when moving along
2070 the compiled data. Specifically:
2071
2072 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2073 length.
2074
2075 (b) An item such as \pL+ within a repeated group could cause crashes or
2076 loops.
2077
2078 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2079 "reference to non-existent subpattern" error.
2080
2081 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2082
2083 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2084 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2085
2086 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2087
2088 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2089 character were causing crashes (broken optimization).
2090
2091 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2092 \p or \P) caused a compile-time loop.
2093
2094 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2095 break. For example, the unstudied pattern [\r\n]A does not match the string
2096 "\r\nA" because change 7.0/46 below moves the current point on by two
2097 characters after failing to match at the start. However, the pattern \nA
2098 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2099 the same is true. There doesn't seem any very clean way out of this, but
2100 what I have chosen to do makes the common cases work: PCRE now takes note
2101 of whether there can be an explicit match for \r or \n anywhere in the
2102 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2103 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2104 pattern has explicit CR or LF references.
2105
2106 29. Added (*CR) etc for changing newline setting at start of pattern.
2107
2108
2109 Version 7.2 19-Jun-07
2110 ---------------------
2111
2112 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2113 which is apparently normally available under Windows.
2114
2115 2. Re-jig the pcregrep tests with different newline settings in an attempt
2116 to make them independent of the local environment's newline setting.
2117
2118 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2119
2120 4. Some of the "internals" tests were previously cut out when the link size
2121 was not 2, because the output contained actual offsets. The recent new
2122 "Z" feature of pcretest means that these can be cut out, making the tests
2123 usable with all link sizes.
2124
2125 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2126 stack recursion. This gives a massive performance boost under BSD, but just
2127 a small improvement under Linux. However, it saves one field in the frame
2128 in all cases.
2129
2130 6. Added more features from the forthcoming Perl 5.10:
2131
2132 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2133 recursion call. It refers to the nth most recently opened parentheses.
2134
2135 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2136 to be opened parentheses.
2137
2138 (c) Conditions that refer to capturing parentheses can be specified
2139 relatively, for example, (?(-2)... or (?(+3)...
2140
2141 (d) \K resets the start of the current match so that everything before
2142 is not part of it.
2143
2144 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2145
2146 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2147 reference syntax.
2148
2149 (g) (?| introduces a group in which the numbering of parentheses in each
2150 alternative starts with the same number.
2151
2152 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2153
2154 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2155 PCRE_INFO_JCHANGED.
2156
2157 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2158 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2159 for detecting groups that can match an empty string.
2160
2161 9. A pattern with a very large number of alternatives (more than several
2162 hundred) was running out of internal workspace during the pre-compile
2163 phase, where pcre_compile() figures out how much memory will be needed. A
2164 bit of new cunning has reduced the workspace needed for groups with
2165 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2166 workspace instead of running out of the 4096 that are available.
2167
2168 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2169
2170 11. Applied patch from Google to remove an optimization that didn't quite work.
2171 The report of the bug said:
2172
2173 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2174 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2175 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2176
2177 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2178 it matched the wrong number of bytes.
2179
2180
2181 Version 7.1 24-Apr-07
2182 ---------------------
2183
2184 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2185 that is more "standard", making use of automake and other Autotools. There
2186 is some re-arrangement of the files and adjustment of comments consequent
2187 on this.
2188
2189 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2190 for recursive directory scanning broke on some systems because the files
2191 are not scanned in any specific order and on different systems the order
2192 was different. A call to "sort" has been inserted into RunGrepTest for the
2193 approprate test as a short-term fix. In the longer term there may be an
2194 alternative.
2195
2196 3. I had an email from Eric Raymond about problems translating some of PCRE's
2197 man pages to HTML (despite the fact that I distribute HTML pages, some
2198 people do their own conversions for various reasons). The problems
2199 concerned the use of low-level troff macros .br and .in. I have therefore
2200 removed all such uses from the man pages (some were redundant, some could
2201 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2202 HTML has been updated to handle .nf/.fi and to complain if it encounters
2203 .br or .in.
2204
2205 4. Updated comments in configure.ac that get placed in config.h.in and also
2206 arranged for config.h to be included in the distribution, with the name
2207 config.h.generic, for the benefit of those who have to compile without
2208 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2209
2210 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2211 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2212 makevp.bat for the current PCRE, using the additional files
2213 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2214
2215 6. A Windows user reported a minor discrepancy with test 2, which turned out
2216 to be caused by a trailing space on an input line that had got lost in his
2217 copy. The trailing space was an accident, so I've just removed it.
2218
2219 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2220 that is needed.
2221
2222 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2223 as "const" (a) because they are and (b) because it helps the PHP
2224 maintainers who have recently made a script to detect big data structures
2225 in the php code that should be moved to the .rodata section. I remembered
2226 to update Builducptable as well, so it won't revert if ucptable.h is ever
2227 re-created.
2228
2229 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2230 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2231 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2232 support is not required. This saves 1.5-2K of code, which is important in
2233 some applications.
2234
2235 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2236 so as not to refer to the tables, even though these functions will never be
2237 called when UTF-8 support is disabled. Otherwise there are problems with a
2238 shared library.
2239
2240 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2241
2242 (a) It was defining its arguments as char * instead of void *.
2243
2244 (b) It was assuming that all moves were upwards in memory; this was true
2245 a long time ago when I wrote it, but is no longer the case.
2246
2247 The emulated memove() is provided for those environments that have neither
2248 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2249 is clearly not the case, as these two bugs were recently reported.
2250
2251 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2252 and Detrail to create the HTML documentation, the .txt form of the man
2253 pages, and it removes trailing spaces from listed files. It also creates
2254 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2255 case, it wraps all the #defines with #ifndefs. This script should be run
2256 before "make dist".
2257
2258 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2259 with Unicode property support.
2260
2261 (a) For a maximizing quantifier, if the two different cases of the
2262 character were of different lengths in their UTF-8 codings (there are
2263 some cases like this - I found 11), and the matching function had to
2264 back up over a mixture of the two cases, it incorrectly assumed they
2265 were both the same length.
2266
2267 (b) When PCRE was configured to use the heap rather than the stack for
2268 recursion during matching, it was not correctly preserving the data for
2269 the other case of a UTF-8 character when checking ahead for a match
2270 while processing a minimizing repeat. If the check also involved
2271 matching a wide character, but failed, corruption could cause an
2272 erroneous result when trying to check for a repeat of the original
2273 character.
2274
2275 13. Some tidying changes to the testing mechanism:
2276
2277 (a) The RunTest script now detects the internal link size and whether there
2278 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2279 values substituted by "configure". (The RunGrepTest script already did
2280 this for UTF-8.) The configure.ac script no longer substitutes the
2281 relevant variables.
2282
2283 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2284 with length and offset values. This means that the output is different
2285 for different internal link sizes. Test 2 is skipped for link sizes
2286 other than 2 because of this, bypassing the problem. Unfortunately,
2287 there was also a test in test 3 (the locale tests) that used /B and
2288 failed for link sizes other than 2. Rather than cut the whole test out,
2289 I have added a new /Z option to pcretest that replaces the length and
2290 offset values with spaces. This is now used to make test 3 independent
2291 of link size. (Test 2 will be tidied up later.)
2292
2293 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2294 segmentation fault instead of returning the appropriate error message.
2295
2296 15. In multiline mode when the newline sequence was set to "any", the pattern
2297 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2298 This doesn't seem right; it now treats the CRLF combination as the line
2299 ending, and so does not match in that case. It's only a pattern such as ^$
2300 that would hit this one: something like ^ABC$ would have failed after \r
2301 and then tried again after \r\n.
2302
2303 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2304 in an attempt to make files that differ only in their line terminators
2305 compare equal. This works on Linux.
2306
2307 17. Under certain error circumstances pcregrep might try to free random memory
2308 as it exited. This is now fixed, thanks to valgrind.
2309
2310 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2311 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2312 was because its rules for how to advance for /g after matching an empty
2313 string at the end of a line did not allow for this case. They now check for
2314 it specially.
2315
2316 20. pcretest is supposed to handle patterns and data of any length, by
2317 extending its buffers when necessary. It was getting this wrong when the
2318 buffer for a data line had to be extended.
2319
2320 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2321 CRLF as a newline sequence.
2322
2323 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2324 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2325 I have nevertheless tidied it up.
2326
2327 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2328
2329 24. Added a man page for pcre-config.
2330
2331
2332 Version 7.0 19-Dec-06
2333 ---------------------
2334
2335 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2336 moving to gcc 4.1.1.
2337
2338 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2339 sys/time.h, which is documented as needed for this function. It doesn't
2340 seem to matter on Linux, but it showed up on some releases of OS X.
2341
2342 3. It seems that there are systems where bytes whose values are greater than
2343 127 match isprint() in the "C" locale. The "C" locale should be the
2344 default when a C program starts up. In most systems, only ASCII printing
2345 characters match isprint(). This difference caused the output from pcretest
2346 to vary, making some of the tests fail. I have changed pcretest so that:
2347
2348 (a) When it is outputting text in the compiled version of a pattern, bytes
2349 other than 32-126 are always shown as hex escapes.
2350
2351 (b) When it is outputting text that is a matched part of a subject string,
2352 it does the same, unless a different locale has been set for the match
2353 (using the /L modifier). In this case, it uses isprint() to decide.
2354
2355 4. Fixed a major bug that caused incorrect computation of the amount of memory
2356 required for a compiled pattern when options that changed within the
2357 pattern affected the logic of the preliminary scan that determines the
2358 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2359 that the computed length was too small. The symptoms of this bug were
2360 either the PCRE error "internal error: code overflow" from pcre_compile(),
2361 or a glibc crash with a message such as "pcretest: free(): invalid next
2362 size (fast)". Examples of patterns that provoked this bug (shown in
2363 pcretest format) are:
2364
2365 /(?-x: )/x
2366 /(?x)(?-x: \s*#\s*)/
2367 /((?i)[\x{c0}])/8
2368 /(?i:[\x{c0}])/8
2369
2370 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2371 is now done differently.
2372
2373 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2374 wrapper classes; (b) implement a new function in the C++ scanner that is
2375 more efficient than the old way of doing things because it avoids levels of
2376 recursion in the regex matching; (c) add a paragraph to the documentation
2377 for the FullMatch() function.
2378
2379 6. The escape sequence \n was being treated as whatever was defined as
2380 "newline". Not only was this contrary to the documentation, which states
2381 that \n is character 10 (hex 0A), but it also went horribly wrong when
2382 "newline" was defined as CRLF. This has been fixed.
2383
2384 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2385 was being set to -1 for the "end of line" case (supposedly a value that no
2386 character can have). Though this value is never used (the check for end of
2387 line is "zero bytes in current character"), it caused compiler complaints.
2388 I've changed it to 0xffffffff.
2389
2390 8. In pcre_version.c, the version string was being built by a sequence of
2391 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2392 string (as it is for production releases) called a macro with an empty
2393 argument. The C standard says the result of this is undefined. The gcc
2394 compiler treats it as an empty string (which was what was wanted) but it is
2395 reported that Visual C gives an error. The source has been hacked around to
2396 avoid this problem.
2397
2398 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2399 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2400 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2401 of them did).
2402
2403 10. Originally, pcretest opened its input and output without "b"; then I was
2404 told that "b" was needed in some environments, so it was added for release
2405 5.0 to both the input and output. (It makes no difference on Unix-like
2406 systems.) Later I was told that it is wrong for the input on Windows. I've
2407 now abstracted the modes into two macros, to make it easier to fiddle with
2408 them, and removed "b" from the input mode under Windows.
2409
2410 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2411
2412 12. Added -help and --help to pcretest as an official way of being reminded
2413 of the options.
2414
2415 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2416 and pcrecpp.cc because they annoy compilers at high warning levels.
2417
2418 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2419
2420 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2421 scripts are not C programs :-) and which was not noticed because it works
2422 on Linux.
2423
2424 16. pcretest is supposed to handle any length of pattern and data line (as one
2425 line or as a continued sequence of lines) by extending its input buffer if
2426 necessary. This feature was broken for very long pattern lines, leading to
2427 a string of junk being passed to pcre_compile() if the pattern was longer
2428 than about 50K.
2429
2430 17. I have done a major re-factoring of the way pcre_compile() computes the
2431 amount of memory needed for a compiled pattern. Previously, there was code
2432 that made a preliminary scan of the pattern in order to do this. That was
2433 OK when PCRE was new, but as the facilities have expanded, it has become
2434 harder and harder to keep it in step with the real compile phase, and there
2435 have been a number of bugs (see for example, 4 above). I have now found a
2436 cunning way of running the real compile function in a "fake" mode that
2437 enables it to compute how much memory it would need, while actually only
2438 ever using a few hundred bytes of working memory and without too many
2439 tests of the mode. This should make future maintenance and development
2440 easier. A side effect of this work is that the limit of 200 on the nesting
2441 depth of parentheses has been removed (though this was never a serious
2442 limitation, I suspect). However, there is a downside: pcre_compile() now
2443 runs more slowly than before (30% or more, depending on the pattern). I
2444 hope this isn't a big issue. There is no effect on runtime performance.
2445
2446 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2447 newline (only possible for the last line of a file) and it was a
2448 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2449
2450 19. Added additional timing features to pcretest. (1) The -tm option now times
2451 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2452 separate command line item, by a number that specifies the number of
2453 repeats to use when timing. The default is 50000; this gives better
2454 precision, but takes uncomfortably long for very large patterns.
2455
2456 20. Extended pcre_study() to be more clever in cases where a branch of a
2457 subpattern has no definite first character. For example, (a*|b*)[cd] would
2458 previously give no result from pcre_study(). Now it recognizes that the
2459 first character must be a, b, c, or d.
2460
2461 21. There was an incorrect error "recursive call could loop indefinitely" if
2462 a subpattern (or the entire pattern) that was being tested for matching an
2463 empty string contained only one non-empty item after a nested subpattern.
2464 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2465 incorrectly, because the \d was being skipped in the check.
2466
2467 22. The pcretest program now has a new pattern option /B and a command line
2468 option -b, which is equivalent to adding /B to every pattern. This causes
2469 it to show the compiled bytecode, without the additional information that
2470 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2471 is the same as /B/I).
2472
2473 23. A new optimization is now able automatically to treat some sequences such
2474 as a*b as a*+b. More specifically, if something simple (such as a character
2475 or a simple class like \d) has an unlimited quantifier, and is followed by
2476 something that cannot possibly match the quantified thing, the quantifier
2477 is automatically "possessified".
2478
2479 24. A recursive reference to a subpattern whose number was greater than 39
2480 went wrong under certain circumstances in UTF-8 mode. This bug could also
2481 have affected the operation of pcre_study().
2482
2483 25. Realized that a little bit of performance could be had by replacing
2484 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2485
2486 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2487
2488 27. Possessive quantifiers such as a++ were previously implemented by turning
2489 them into atomic groups such as ($>a+). Now they have their own opcodes,
2490 which improves performance. This includes the automatically created ones
2491 from 23 above.
2492
2493 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2494 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2495 the first matched character to be a colon. This applied both to named and
2496 numbered groups.
2497
2498 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2499
2500 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2501 building PCRE on Windows easier, so I have included it in the distribution.
2502
2503 31. There is now a check in pcretest against a ridiculously large number being
2504 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2505 loop, the loop is abandoned.
2506
2507 32. Forward references to subpatterns in conditions such as (?(2)...) where
2508 subpattern 2 is defined later cause pcre_compile() to search forwards in
2509 the pattern for the relevant set of parentheses. This search went wrong
2510 when there were unescaped parentheses in a character class, parentheses
2511 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2512
2513 33. "Subroutine" calls and backreferences were previously restricted to
2514 referencing subpatterns earlier in the regex. This restriction has now
2515 been removed.
2516
2517 34. Added a number of extra features that are going to be in Perl 5.10. On the
2518 whole, these are just syntactic alternatives for features that PCRE had
2519 previously implemented using the Python syntax or my own invention. The
2520 other formats are all retained for compatibility.
2521
2522 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2523 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2524 also .NET compatible.
2525
2526 (b) A recursion or subroutine call to a named group can now be defined as
2527 (?&name) as well as (?P>name).
2528
2529 (c) A backreference to a named group can now be defined as \k<name> or
2530 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2531 5.10, are also .NET compatible.
2532
2533 (d) A conditional reference to a named group can now use the syntax
2534 (?(<name>) or (?('name') as well as (?(name).
2535
2536 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2537 groups (named and numbered) that are never evaluated inline, but can be
2538 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2539 is always false. There may be only one alternative in such a group.
2540
2541 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2542 as the simple (?(R). The condition is true only if the most recent
2543 recursion is that of the given number or name. It does not search out
2544 through the entire recursion stack.
2545
2546 (g) The escape \gN or \g{N} has been added, where N is a positive or
2547 negative number, specifying an absolute or relative reference.
2548
2549 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2550 some "unreachable code" warnings.
2551
2552 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2553 things, this adds five new scripts.
2554
2555 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2556 There were also incompatibilities regarding the handling of \Q..\E inside
2557 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2558 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2559
2560 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2561 matches an empty string, and forcibly breaks the loop. There were bugs in
2562 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2563 against aaaa the result was just "a" rather than "aaaa", for example. Two
2564 separate and independent bugs (that affected different cases) have been
2565 fixed.
2566
2567 39. Refactored the code to abolish the use of different opcodes for small
2568 capturing bracket numbers. This is a tidy that I avoided doing when I
2569 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2570 The new approach is not only tidier, it makes it possible to reduce the
2571 memory needed to fix the previous bug (38).
2572
2573 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2574 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2575 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2576 mode.
2577
2578 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2579 report.
2580
2581 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2582 copy construction and assignment in the C++ wrapper.
2583
2584 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2585 couple of bugs that could have given wrong results in the "--newline=crlf"
2586 case.
2587
2588 44. Added a number of casts and did some reorganization of signed/unsigned int
2589 variables following suggestions from Dair Grant. Also renamed the variable
2590 "this" as "item" because it is a C++ keyword.
2591
2592 45. Arranged for dftables to add
2593
2594 #include "pcre_internal.h"
2595
2596 to pcre_chartables.c because without it, gcc 4.x may remove the array
2597 definition from the final binary if PCRE is built into a static library and
2598 dead code stripping is activated.
2599
2600 46. For an unanchored pattern, if a match attempt fails at the start of a
2601 newline sequence, and the newline setting is CRLF or ANY, and the next two
2602 characters are CRLF, advance by two characters instead of one.
2603
2604
2605 Version 6.7 04-Jul-06
2606 ---------------------
2607
2608 1. In order to handle tests when input lines are enormously long, pcretest has
2609 been re-factored so that it automatically extends its buffers when
2610 necessary. The code is crude, but this _is_ just a test program. The
2611 default size has been increased from 32K to 50K.
2612
2613 2. The code in pcre_study() was using the value of the re argument before
2614 testing it for NULL. (Of course, in any sensible call of the function, it
2615 won't be NULL.)
2616
2617 3. The memmove() emulation function in pcre_internal.h, which is used on
2618 systems that lack both memmove() and bcopy() - that is, hardly ever -
2619 was missing a "static" storage class specifier.
2620
2621 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2622 containing an extended class (one that cannot be represented by a bitmap
2623 because it contains high-valued characters or Unicode property items, e.g.
2624 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2625 pattern, but PCRE should not loop if you do not (it no longer does).
2626 [Detail: two cases were found: (a) a repeated subpattern containing an
2627 extended class; (b) a recursive reference to a subpattern that followed a
2628 previous extended class. It wasn't skipping over the extended class
2629 correctly when UTF-8 mode was not set.]
2630
2631 5. A negated single-character class was not being recognized as fixed-length
2632 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2633 compile error "lookbehind assertion is not fixed length".
2634
2635 6. The RunPerlTest auxiliary script was showing an unexpected difference
2636 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2637 write a Perl script that can interpret lines of an input file either as
2638 byte characters or as UTF-8, which is what "perltest" was being required to
2639 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2640 can't do is switch easily at run time between having the "use utf8;" pragma
2641 or not. In the end, I fudged it by using the RunPerlTest script to insert
2642 "use utf8;" explicitly for the UTF-8 tests.
2643
2644 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2645 the end of the subject string, contrary to the documentation and to what
2646 Perl does. This was true of both matching functions. Now it matches only at
2647 the start of the subject and immediately after *internal* newlines.
2648
2649 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2650 a pointer to an int instead of a pointer to an unsigned long int. This
2651 caused problems on 64-bit systems.
2652
2653 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2654 instance of the 'standard' template library not being so standard".
2655
2656 10. There was no check on the number of named subpatterns nor the maximum
2657 length of a subpattern name. The product of these values is used to compute
2658 the size of the memory block for a compiled pattern. By supplying a very
2659 long subpattern name and a large number of named subpatterns, the size
2660 computation could be caused to overflow. This is now prevented by limiting
2661 the length of names to 32 characters, and the number of named subpatterns
2662 to 10,000.
2663
2664 11. Subpatterns that are repeated with specific counts have to be replicated in
2665 the compiled pattern. The size of memory for this was computed from the
2666 length of the subpattern and the repeat count. The latter is limited to
2667 65535, but there was no limit on the former, meaning that integer overflow
2668 could in principle occur. The compiled length of a repeated subpattern is
2669 now limited to 30,000 bytes in order to prevent this.
2670
2671 12. Added the optional facility to have named substrings with the same name.
2672
2673 13. Added the ability to use a named substring as a condition, using the
2674 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2675 are numbers (not recommended). Forward references are permitted.
2676
2677 14. Added forward references in named backreferences (if you see what I mean).
2678
2679 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2680 pattern could run off the end of the subject. For example, the pattern
2681 "(?s)(.{1,5})"8 did this with the subject "ab".
2682
2683 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2684 PCRE_CASELESS was set when matching characters that were quantified with ?
2685 or *.
2686
2687 17. A character class other than a single negated character that had a minimum
2688 but no maximum quantifier - for example [ab]{6,} - was not handled
2689 correctly by pce_dfa_exec(). It would match only one character.
2690
2691 18. A valid (though odd) pattern that looked like a POSIX character
2692 class but used an invalid character after [ (for example [[,abc,]]) caused
2693 pcre_compile() to give the error "Failed: internal error: code overflow" or
2694 in some cases to crash with a glibc free() error. This could even happen if
2695 the pattern terminated after [[ but there just happened to be a sequence of
2696 letters, a binary zero, and a closing ] in the memory that followed.
2697
2698 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2699 over the years. Originally (before any Unicode support), just the bottom 8
2700 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2701 output from "man perlunicode" includes this:
2702
2703 The regular expression compiler produces polymorphic opcodes. That
2704 is, the pattern adapts to the data and automatically switches to
2705 the Unicode character scheme when presented with Unicode data--or
2706 instead uses a traditional byte scheme when presented with byte
2707 data.
2708
2709 Sadly, a wide octal escape does not cause a switch, and in a string with
2710 no other multibyte characters, these octal escapes are treated as before.
2711 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2712 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2713 Unicode string.
2714
2715 I have not perpetrated such confusion in PCRE. Up till now, it took just
2716 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2717 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2718 translate to the appropriate multibyte character.
2719
2720 29. Applied some refactoring to reduce the number of warnings from Microsoft
2721 and Borland compilers. This has included removing the fudge introduced
2722 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2723 a warning about an unused variable.
2724
2725 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2726 characters since release 4.0, because Perl (from release 5.004) does not.
2727 [Or at least, is documented not to: some releases seem to be in conflict
2728 with the documentation.] However, when a pattern was studied with
2729 pcre_study() and all its branches started with \s, PCRE still included VT
2730 as a possible starting character. Of course, this did no harm; it just
2731 caused an unnecessary match attempt.
2732
2733 22. Removed a now-redundant internal flag bit that recorded the fact that case
2734 dependency changed within the pattern. This was once needed for "required
2735 byte" processing, but is no longer used. This recovers a now-scarce options
2736 bit. Also moved the least significant internal flag bit to the most-
2737 significant bit of the word, which was not previously used (hangover from
2738 the days when it was an int rather than a uint) to free up another bit for
2739 the future.
2740
2741 23. Added support for CRLF line endings as well as CR and LF. As well as the
2742 default being selectable at build time, it can now be changed at runtime
2743 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2744 specify that it is scanning data with non-default line endings.
2745
2746 24. Changed the definition of CXXLINK to make it agree with the definition of
2747 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2748
2749 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2750 recursions. This makes a big different to stack usage for some patterns.
2751
2752 26. If a subpattern containing a named recursion or subroutine reference such
2753 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2754 the space required for the compiled pattern went wrong and gave too small a
2755 value. Depending on the environment, this could lead to "Failed: internal
2756 error: code overflow at offset 49" or "glibc detected double free or
2757 corruption" errors.
2758
2759 27. Applied patches from Google (a) to support the new newline modes and (b) to
2760 advance over multibyte UTF-8 characters in GlobalReplace.
2761
2762 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2763 difference for some implementation of PCRE in some Windows version.
2764
2765 29. Added some extra testing facilities to pcretest:
2766
2767 \q<number> in a data line sets the "match limit" value
2768 \Q<number> in a data line sets the "match recursion limt" value
2769 -S <number> sets the stack size, where <number> is in megabytes
2770
2771 The -S option isn't available for Windows.
2772
2773
2774 Version 6.6 06-Feb-06
2775 ---------------------
2776
2777 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2778 in pcreposix.h. I have copied the definition from pcre.h.
2779
2780 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2781 because pcre.h is no longer a built file.
2782
2783 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2784 not normally included in the compiled code.
2785
2786
2787 Version 6.5 01-Feb-06
2788 ---------------------
2789
2790 1. When using the partial match feature with pcre_dfa_exec(), it was not
2791 anchoring the second and subsequent partial matches at the new starting
2792 point. This could lead to incorrect results. For example, with the pattern
2793 /1234/, partially matching against "123" and then "a4" gave a match.
2794
2795 2. Changes to pcregrep:
2796
2797 (a) All non-match returns from pcre_exec() were being treated as failures
2798 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2799 error message is output. Some extra information is given for the
2800 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2801 probably the only errors that are likely to be caused by users (by
2802 specifying a regex that has nested indefinite repeats, for instance).
2803 If there are more than 20 of these errors, pcregrep is abandoned.
2804
2805 (b) A binary zero was treated as data while matching, but terminated the
2806 output line if it was written out. This has been fixed: binary zeroes
2807 are now no different to any other data bytes.
2808
2809 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2810 used to set a locale for matching. The --locale=xxxx long option has
2811 been added (no short equivalent) to specify a locale explicitly on the
2812 pcregrep command, overriding the environment variables.
2813
2814 (d) When -B was used with -n, some line numbers in the output were one less
2815 than they should have been.
2816
2817 (e) Added the -o (--only-matching) option.
2818
2819 (f) If -A or -C was used with -c (count only), some lines of context were
2820 accidentally printed for the final match.
2821
2822 (g) Added the -H (--with-filename) option.
2823
2824 (h) The combination of options -rh failed to suppress file names for files
2825 that were found from directory arguments.
2826
2827 (i) Added the -D (--devices) and -d (--directories) options.
2828
2829 (j) Added the -F (--fixed-strings) option.
2830
2831 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2832
2833 (l) Added the --colo(u)r option.
2834
2835 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2836 is not present by default.
2837
2838 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2839 items such as (?R) or (?1), when the recursion could match a number of
2840 alternatives. If it matched one of the alternatives, but subsequently,
2841 outside the recursion, there was a failure, the code tried to back up into
2842 the recursion. However, because of the way PCRE is implemented, this is not
2843 possible, and the result was an incorrect result from the match.
2844
2845 In order to prevent this happening, the specification of recursion has
2846 been changed so that all such subpatterns are automatically treated as
2847 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2848
2849 4. I had overlooked the fact that, in some locales, there are characters for
2850 which isalpha() is true but neither isupper() nor islower() are true. In
2851 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2852 and ordfeminine) are like this. This affected the treatment of \w and \W
2853 when they appeared in character classes, but not when they appeared outside
2854 a character class. The bit map for "word" characters is now created
2855 separately from the results of isalnum() instead of just taking it from the
2856 upper, lower, and digit maps. (Plus the underscore character, of course.)
2857
2858 5. The above bug also affected the handling of POSIX character classes such as
2859 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2860 permanent tables. Instead, the bit maps for such a class were previously
2861 created as the appropriate unions of the upper, lower, and digit bitmaps.
2862 Now they are created by subtraction from the [[:word:]] class, which has
2863 its own bitmap.
2864
2865 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2866 It is created by subtracting the vertical space characters (\x09, \x0a,
2867 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2868 subtraction was done in the overall bitmap for a character class, meaning
2869 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2870 be recognized. This bug has been fixed.
2871
2872 7. Patches from the folks at Google:
2873
2874 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2875 real life, but is still worth protecting against".
2876
2877 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2878 regular expressions".
2879
2880 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2881 have it.
2882
2883 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2884 "configure" and the latter not, in order to fix a problem somebody had
2885 with compiling the Arg class on HP-UX.
2886
2887 (e) Improve the error-handling of the C++ wrapper a little bit.
2888
2889 (f) New tests for checking recursion limiting.
2890
2891 8. The pcre_memmove() function, which is used only if the environment does not
2892 have a standard memmove() function (and is therefore rarely compiled),
2893 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2894 returning a result (though PCRE never actually uses the result).
2895
2896 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2897 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2898 returned instead of calling malloc() with an overflowing number that would
2899 most likely cause subsequent chaos.
2900
2901 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2902
2903 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2904 with this option is matched, the nmatch and pmatch options of regexec() are
2905 ignored.
2906
2907 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2908 provided in case anyone wants to the the POSIX interface with UTF-8
2909 strings.
2910
2911 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2912 C++ linking (needed for some HP-UX environments).
2913
2914 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2915 (unused parameter) and in the pcre_printint() function (omitted "default"
2916 switch label when the default is to do nothing).
2917
2918 15. Added some code to make it possible, when PCRE is compiled as a C++
2919 library, to replace subject pointers for pcre_exec() with a smart pointer
2920 class, thus making it possible to process discontinuous strings.
2921
2922 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2923 much the same function. They were added by different people who were trying
2924 to make PCRE easy to compile on non-Unix systems. It has been suggested
2925 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2926 for compiling on Windows systems. I have therefore replaced it with
2927 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2928 defaults to "extern" for C or "extern C" for C++, which works fine on
2929 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2930 SCOPE with something explicit in config.h. In addition:
2931
2932 (a) pcreposix.h still had just "extern" instead of either of these macros;
2933 I have replaced it with PCRE_DATA_SCOPE.
2934
2935 (b) Functions such as _pcre_xclass(), which are internal to the library,
2936 but external in the C sense, all had PCRE_EXPORT in their definitions.
2937 This is apparently wrong for the Windows case, so I have removed it.
2938 (It makes no difference on Unix-like systems.)
2939
2940 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2941 of recursive calls to match(). This is different to MATCH_LIMIT because
2942 that limits the total number of calls to match(), not all of which increase
2943 the depth of recursion. Limiting the recursion depth limits the amount of
2944 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2945 when PCRE is compiled, and changed at run time. A patch from Google adds
2946 this functionality to the C++ interface.
2947
2948 18. Changes to the handling of Unicode character properties:
2949
2950 (a) Updated the table to Unicode 4.1.0.
2951
2952 (b) Recognize characters that are not in the table as "Cn" (undefined).
2953
2954 (c) I revised the way the table is implemented to a much improved format
2955 which includes recognition of ranges. It now supports the ranges that
2956 are defined in UnicodeData.txt, and it also amalgamates other
2957 characters into ranges. This has reduced the number of entries in the
2958 table from around 16,000 to around 3,000, thus reducing its size
2959 considerably. I realized I did not need to use a tree structure after
2960 all - a binary chop search is just as efficient. Having reduced the
2961 number of entries, I extended their size from 6 bytes to 8 bytes to
2962 allow for more data.
2963
2964 (d) Added support for Unicode script names via properties such as \p{Han}.
2965
2966 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2967 matching that character.
2968
2969 20. When matching a repeated Unicode property with a minimum greater than zero,
2970 (for example \pL{2,}), PCRE could look past the end of the subject if it
2971 reached it while seeking the minimum number of characters. This could
2972 happen only if some of the characters were more than one byte long, because
2973 there is a check for at least the minimum number of bytes.
2974
2975 21. Refactored the implementation of \p and \P so as to be more general, to
2976 allow for more different types of property in future. This has changed the
2977 compiled form incompatibly. Anybody with saved compiled patterns that use
2978 \p or \P will have to recompile them.
2979
2980 22. Added "Any" and "L&" to the supported property types.
2981
2982 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2983 but give a compile time error if the value is greater than 0xff.
2984
2985 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2986 accidentally not being installed or uninstalled.
2987
2988 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2989 made were to insert the current release number. This seemed silly, because
2990 it made things harder for people building PCRE on systems that don't run
2991 "configure". I have turned pcre.h into a distributed file, no longer built
2992 by "configure", with the version identification directly included. There is
2993 no longer a pcre.h.in file.
2994
2995 However, this change necessitated a change to the pcre-config script as
2996 well. It is built from pcre-config.in, and one of the substitutions was the
2997 release number. I have updated configure.ac so that ./configure now finds
2998 the release number by grepping pcre.h.
2999
3000 26. Added the ability to run the tests under valgrind.
3001
3002
3003 Version 6.4 05-Sep-05
3004 ---------------------
3005
3006 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3007 "--" to be printed when multiple files were scanned, even when none of the
3008 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3009 consider it to be a bug, and have restored the previous behaviour.
3010
3011 2. A couple of code tidies to get rid of compiler warnings.
3012
3013 3. The pcretest program used to cheat by referring to symbols in the library
3014 whose names begin with _pcre_. These are internal symbols that are not
3015 really supposed to be visible externally, and in some environments it is
3016 possible to suppress them. The cheating is now confined to including
3017 certain files from the library's source, which is a bit cleaner.
3018
3019 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3020 file's purpose clearer.
3021
3022 5. Reorganized pcre_ucp_findchar().
3023
3024
3025 Version 6.3 15-Aug-05
3026 ---------------------
3027
3028 1. The file libpcre.pc.in did not have general read permission in the tarball.
3029
3030 2. There were some problems when building without C++ support:
3031
3032 (a) If C++ support was not built, "make install" and "make test" still
3033 tried to test it.
3034
3035 (b) There were problems when the value of CXX was explicitly set. Some
3036 changes have been made to try to fix these, and ...
3037
3038 (c) --disable-cpp can now be used to explicitly disable C++ support.
3039
3040 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3041 backslash in a target when C++ was disabled. This confuses some
3042 versions of "make", apparently. Using an intermediate variable solves
3043 this. (Same for CPP_LOBJ.)
3044
3045 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3046 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3047 necessary on certain architectures.
3048
3049 4. Added a setting of -export-symbols-regex to the link command to remove
3050 those symbols that are exported in the C sense, but actually are local
3051 within the library, and not documented. Their names all begin with
3052 "_pcre_". This is not a perfect job, because (a) we have to except some
3053 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3054 available (and never for static libraries). I have made a note to try to
3055 find a way round (a) in the future.
3056
3057
3058 Version 6.2 01-Aug-05
3059 ---------------------
3060
3061 1. There was no test for integer overflow of quantifier values. A construction
3062 such as {1111111111111111} would give undefined results. What is worse, if
3063 a minimum quantifier for a parenthesized subpattern overflowed and became
3064 negative, the calculation of the memory size went wrong. This could have
3065 led to memory overwriting.
3066
3067 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3068
3069 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3070 operating environments where this matters.
3071
3072 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3073 PCRE options from within the C++ wrapper.
3074
3075 5. Named capturing subpatterns were not being correctly counted when a pattern
3076 was compiled. This caused two problems: (a) If there were more than 100
3077 such subpatterns, the calculation of the memory needed for the whole
3078 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3079 back references of the form \12, where the number was greater than 9, were
3080 not recognized as back references, even though there were sufficient
3081 previous subpatterns.
3082
3083 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3084 versions of gcc, e.g. 2.95.4.
3085
3086
3087 Version 6.1 21-Jun-05
3088 ---------------------
3089
3090 1. There was one reference to the variable "posix" in pcretest.c that was not
3091 surrounded by "#if !defined NOPOSIX".
3092
3093 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3094 the cross-check on the old pcre_info() function, for the benefit of the
3095 cut-down version of PCRE that is currently imported into Exim.
3096
3097 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3098 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3099 patterns that start (?i) but I don't think that matters. The use of (?i) is
3100 just an example; this all applies to the other options as well.
3101
3102 4. Since libtool seems to echo the compile commands it is issuing, the output
3103 from "make" can be reduced a bit by putting "@" in front of each libtool
3104 compile command.
3105
3106 5. Patch from the folks at Google for configure.in to be a bit more thorough
3107 in checking for a suitable C++ installation before trying to compile the
3108 C++ stuff. This should fix a reported problem when a compiler was present,
3109 but no suitable headers.
3110
3111 6. The man pages all had just "PCRE" as their title. I have changed them to
3112 be the relevant file name. I have also arranged that these names are
3113 retained in the file doc/pcre.txt, which is a concatenation in text format
3114 of all the man pages except the little individual ones for each function.
3115
3116 7. The NON-UNIX-USE file had not been updated for the different set of source
3117 files that come with release 6. I also added a few comments about the C++
3118 wrapper.
3119
3120
3121 Version 6.0 07-Jun-05
3122 ---------------------
3123
3124 1. Some minor internal re-organization to help with my DFA experiments.
3125
3126 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3127 didn't matter for the library itself when fully configured, but did matter
3128 when compiling without UCP support, or within Exim, where the ucp files are
3129 not imported.
3130
3131 3. Refactoring of the library code to split up the various functions into
3132 different source modules. The addition of the new DFA matching code (see
3133 below) to a single monolithic source would have made it really too
3134 unwieldy, quite apart from causing all the code to be include in a
3135 statically linked application, when only some functions are used. This is
3136 relevant even without the DFA addition now that patterns can be compiled in
3137 one application and matched in another.
3138
3139 The downside of splitting up is that there have to be some external
3140 functions and data tables that are used internally in different modules of
3141 the library but which are not part of the API. These have all had their
3142 names changed to start with "_pcre_" so that they are unlikely to clash
3143 with other external names.
3144
3145 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3146 a different (DFA) algorithm. Although it is slower than the original
3147 function, it does have some advantages for certain types of matching
3148 problem.
3149
3150 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3151 including restarting after a partial match.
3152
3153 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3154 defined when compiling for Windows was sent to me. I have put it into the
3155 code, though I have no means of testing or verifying it.
3156
3157 7. Added the pcre_refcount() auxiliary function.
3158
3159 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3160 match before or at the first newline in the subject string. In pcretest,
3161 the /f option on a pattern can be used to set this.
3162
3163 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3164 would behave wrongly. This has been present in PCRE since release 4.0.
3165
3166 10. A number of changes to the pcregrep command:
3167
3168 (a) Refactored how -x works; insert ^(...)$ instead of setting
3169 PCRE_ANCHORED and checking the length, in preparation for adding
3170 something similar for -w.
3171
3172 (b) Added the -w (match as a word) option.
3173
3174 (c) Refactored the way lines are read and buffered so as to have more
3175 than one at a time available.
3176
3177 (d) Implemented a pcregrep test script.
3178
3179 (e) Added the -M (multiline match) option. This allows patterns to match
3180 over several lines of the subject. The buffering ensures that at least
3181 8K, or the rest of the document (whichever is the shorter) is available
3182 for matching (and similarly the previous 8K for lookbehind assertions).
3183
3184 (f) Changed the --help output so that it now says
3185
3186 -w, --word-regex(p)
3187
3188 instead of two lines, one with "regex" and the other with "regexp"
3189 because that confused at least one person since the short forms are the
3190 same. (This required a bit of code, as the output is generated
3191 automatically from a table. It wasn't just a text change.)
3192
3193 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3194 option but starts with a hyphen. Could be a pattern or a path name
3195 starting with a hyphen, for instance.
3196
3197 (h) "-" can be given as a file name to represent stdin.
3198
3199 (i) When file names are being printed, "(standard input)" is used for
3200 the standard input, for compatibility with GNU grep. Previously
3201 "<stdin>" was used.
3202
3203 (j) The option --label=xxx can be used to supply a name to be used for
3204 stdin when file names are being printed. There is no short form.
3205
3206 (k) Re-factored the options decoding logic because we are going to add
3207 two more options that take data. Such options can now be given in four
3208 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3209
3210 (l) Added the -A, -B, and -C options for requesting that lines of context
3211 around matches be printed.
3212
3213 (m) Added the -L option to print the names of files that do not contain
3214 any matching lines, that is, the complement of -l.
3215
3216 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3217 continue to scan other files.
3218
3219 (o) The -s option was incorrectly implemented. For compatibility with other
3220 greps, it now suppresses the error message for a non-existent or non-
3221 accessible file (but not the return code). There is a new option called
3222 -q that suppresses the output of matching lines, which was what -s was
3223 previously doing.
3224
3225 (p) Added --include and --exclude options to specify files for inclusion
3226 and exclusion when recursing.
3227
3228 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3229 Hopefully, it now does.
3230
3231 12. Missing cast in pcre_study().
3232
3233 13. Added an "uninstall" target to the makefile.
3234
3235 14. Replaced "extern" in the function prototypes in Makefile.in with
3236 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3237 world, but is set differently for Windows.
3238
3239 15. Added a second compiling function called pcre_compile2(). The only
3240 difference is that it has an extra argument, which is a pointer to an
3241 integer error code. When there is a compile-time failure, this is set
3242 non-zero, in addition to the error test pointer being set to point to an
3243 error message. The new argument may be NULL if no error number is required
3244 (but then you may as well call pcre_compile(), which is now just a
3245 wrapper). This facility is provided because some applications need a
3246 numeric error indication, but it has also enabled me to tidy up the way
3247 compile-time errors are handled in the POSIX wrapper.
3248
3249 16. Added VPATH=.libs to the makefile; this should help when building with one
3250 prefix path and installing with another. (Or so I'm told by someone who
3251 knows more about this stuff than I do.)
3252
3253 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3254 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3255 match everything, including newlines. This is not POSIX-compatible, but
3256 somebody wanted the feature. From pcretest it can be activated by using
3257 both the P and the s flags.
3258
3259 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3260
3261 19. libpcre.pc was being incorrectly installed as executable.
3262
3263 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3264 it now also looks for '\r' so that it will work unmodified on Windows.
3265
3266 21. Added Google's contributed C++ wrapper to the distribution.
3267
3268 22. Added some untidy missing memory free() calls in pcretest, to keep
3269 Electric Fence happy when testing.
3270
3271
3272
3273 Version 5.0 13-Sep-04
3274 ---------------------
3275
3276 1. Internal change: literal characters are no longer packed up into items
3277 containing multiple characters in a single byte-string. Each character
3278 is now matched using a separate opcode. However, there may be more than one
3279 byte in the character in UTF-8 mode.
3280
3281 2. The pcre_callout_block structure has two new fields: pattern_position and
3282 next_item_length. These contain the offset in the pattern to the next match
3283 item, and its length, respectively.
3284
3285 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3286 insertion of callouts before each pattern item. Added the /C option to
3287 pcretest to make use of this.
3288
3289 4. On the advice of a Windows user, the lines
3290
3291 #if defined(_WIN32) || defined(WIN32)
3292 _setmode( _fileno( stdout ), 0x8000 );
3293 #endif /* defined(_WIN32) || defined(WIN32) */
3294
3295 have been added to the source of pcretest. This apparently does useful
3296 magic in relation to line terminators.
3297
3298 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3299 for the benefit of those environments where the "b" makes a difference.
3300
3301 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3302 to know about it. I have put a hack into configure.in that adds in code
3303 to set GCC=yes if CC=icc. This seems to end up at a point in the
3304 generated configure script that is early enough to affect the setting of
3305 compiler options, which is what is needed, but I have no means of testing
3306 whether it really works. (The user who reported this had patched the
3307 generated configure script, which of course I cannot do.)
3308
3309 LATER: After change 22 below (new libtool files), the configure script
3310 seems to know about icc (and also ecc). Therefore, I have commented out
3311 this hack in configure.in.
3312
3313 7. Added support for pkg-config (2 patches were sent in).
3314
3315 8. Negated POSIX character classes that used a combination of internal tables
3316 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3317 [[:^ascii]]. Typically, they would match almost any characters. The other
3318 POSIX classes were not broken in this way.
3319
3320 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3321 to find the match, as PCRE was deluded into thinking that the match had to
3322 start at the start point or following a newline. The same bug applied to
3323 patterns with negative forward assertions or any backward assertions
3324 preceding ".*" at the start, unless the pattern required a fixed first
3325 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3326
3327 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3328 starting at the last subject character, bytes beyond the end of the subject
3329 string were read.
3330
3331 11. Renamed the variable "class" as "classbits" to make life easier for C++
3332 users. (Previously there was a macro definition, but it apparently wasn't
3333 enough.)
3334
3335 12. Added the new field "tables" to the extra data so that tables can be passed
3336 in at exec time, or the internal tables can be re-selected. This allows
3337 a compiled regex to be saved and re-used at a later time by a different
3338 program that might have everything at different addresses.
3339
3340 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3341 -R library as well as a -L library.
3342
3343 14. The debugging options of pcretest (-d on the command line or D on a
3344 pattern) showed incorrect output for anything following an extended class
3345 that contained multibyte characters and which was followed by a quantifier.
3346
3347 15. Added optional support for general category Unicode character properties
3348 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3349 support. It adds about 90K to the size of the library. The meanings of the
3350 inbuilt class escapes such as \d and \s have NOT been changed.
3351
3352 16. Updated pcredemo.c to include calls to free() to release the memory for the
3353 compiled pattern.
3354
3355 17. The generated file chartables.c was being created in the source directory
3356 instead of in the building directory. This caused the build to fail if the
3357 source directory was different from the building directory, and was
3358 read-only.
3359
3360 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3361 file. No doubt somebody will tell me if they don't make sense... Also added
3362 Dan Mooney's comments about building on OpenVMS.
3363
3364 19. Added support for partial matching via the PCRE_PARTIAL option for
3365 pcre_exec() and the \P data escape in pcretest.
3366
3367 20. Extended pcretest with 3 new pattern features:
3368
3369 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3370 write the compiled pattern to the file whose name is "rest-of-line".
3371 This is a straight binary dump of the data, with the saved pointer to
3372 the character tables forced to be NULL. The study data, if any, is
3373 written too. After writing, pcretest reads a new pattern.
3374
3375 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3376 compiled pattern from the given file. There must not be any
3377 occurrences of "<" in the file name (pretty unlikely); if there are,
3378 pcretest will instead treat the initial "<" as a pattern delimiter.
3379 After reading in the pattern, pcretest goes on to read data lines as
3380 usual.
3381
3382 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3383 and 16-bit fields in a compiled pattern, to simulate a pattern that
3384 was compiled on a host of opposite endianness.
3385
3386 21. The pcre-exec() function can now cope with patterns that were compiled on
3387 hosts of opposite endianness, with this restriction:
3388
3389 As for any compiled expression that is saved and used later, the tables
3390 pointer field cannot be preserved; the extra_data field in the arguments
3391 to pcre_exec() should be used to pass in a tables address if a value
3392 other than the default internal tables were used at compile time.
3393
3394 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3395 now diagnosed as an error. Previously, most of the time, a negative number
3396 would have been treated as zero, but if in addition "ovector" was passed as
3397 NULL, a crash could occur.
3398
3399 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3400 new versions from the libtool 1.5 distribution (the last one is a copy of
3401 a file called libtool.m4). This seems to have fixed the need to patch
3402 "configure" to support Darwin 1.3 (which I used to do). However, I still
3403 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3404 workstation).
3405
3406 24. Changed the PCRE licence to be the more standard "BSD" licence.
3407
3408
3409 Version 4.5 01-Dec-03
3410 ---------------------
3411
3412 1. There has been some re-arrangement of the code for the match() function so
3413 that it can be compiled in a version that does not call itself recursively.
3414 Instead, it keeps those local variables that need separate instances for
3415 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3416 needs to "recurse". Keeping track of where control must go is done by means
3417 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3418 hide most of the details from the main code, and operates only if
3419 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3420 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3421 operating.
3422
3423 To make it easier for callers to provide specially tailored get/free
3424 functions for this usage, two new functions, pcre_stack_malloc, and
3425 pcre_stack_free, are used. They are always called in strict stacking order,
3426 and the size of block requested is always the same.
3427
3428 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3429 PCRE has been compiled to use the stack or the heap for recursion. The
3430 -C option of pcretest uses this to show which version is compiled.
3431
3432 A new data escape \S, is added to pcretest; it causes the amounts of store
3433 obtained and freed by both kinds of malloc/free at match time to be added
3434 to the output.
3435
3436 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3437 what's available on my current Linux desktop machine.
3438
3439 3. When matching a UTF-8 string, the test for a valid string at the start has
3440 been extended. If start_offset is not zero, PCRE now checks that it points
3441 to a byte that is the start of a UTF-8 character. If not, it returns
3442 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3443 this is necessary because there may be backward assertions in the pattern.
3444 When matching the same subject several times, it may save resources to use
3445 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3446
3447 4. The code for checking the validity of UTF-8 strings has been tightened so
3448 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3449 containing "overlong sequences".
3450
3451 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3452 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3453 should have been "&", but it just so happened that all the cases this let
3454 through by mistake were picked up later in the function.
3455
3456 6. I had used a variable called "isblank" - this is a C99 function, causing
3457 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3458
3459 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3460 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3461 so the version is shown; (c) stop "make test" echoing "./RunTest".
3462
3463 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3464
3465 9. The prototype for memmove() for systems that don't have it was using
3466 size_t, but the inclusion of the header that defines size_t was later. I've
3467 moved the #includes for the C headers earlier to avoid this.
3468
3469 10. Added some adjustments to the code to make it easier to compiler on certain
3470 special systems:
3471
3472 (a) Some "const" qualifiers were missing.
3473 (b) Added the macro EXPORT before all exported functions; by default this
3474 is defined to be empty.
3475 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3476 that it reads its output file name as an argument instead of writing
3477 to the standard output and assuming this can be redirected.
3478
3479 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3480 class containing characters with values greater than 255, PCRE compilation
3481 went into a loop.
3482
3483 12. A recursive reference to a subpattern that was within another subpattern
3484 that had a minimum quantifier of zero caused PCRE to crash. For example,
3485 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3486 recursion. If the recursively-called subpattern itself had a zero repeat,
3487 that was OK.
3488
3489 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3490 buffer into which it was copied (for escape processing) was still set at
3491 1024, so long lines caused crashes.
3492
3493 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3494 "internal error: code overflow...". This applied to any character class
3495 that was followed by a possessive quantifier.
3496
3497 15. Modified the Makefile to add libpcre.la as a prerequisite for
3498 libpcreposix.la because I was told this is needed for a parallel build to
3499 work.
3500
3501 16. If a pattern that contained .* following optional items at the start was
3502 studied, the wrong optimizing data was generated, leading to matching
3503 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3504 matching string must start with a or b or c. The correct conclusion for
3505 this pattern is that a match can start with any character.
3506
3507
3508 Version 4.4 13-Aug-03
3509 ---------------------
3510
3511 1. In UTF-8 mode, a character class containing characters with values between
3512 127 and 255 was not handled correctly if the compiled pattern was studied.
3513 In fixing this, I have also improved the studying algorithm for such
3514 classes (slightly).
3515
3516 2. Three internal functions had redundant arguments passed to them. Removal
3517 might give a very teeny performance improvement.
3518
3519 3. Documentation bug: the value of the capture_top field in a callout is *one
3520 more than* the number of the hightest numbered captured substring.
3521
3522 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3523 in incorrectly linking with a previously installed version. They now link
3524 explicitly with libpcre.la.
3525
3526 5. configure.in no longer needs to recognize Cygwin specially.
3527
3528 6. A problem in pcre.in for Windows platforms is fixed.
3529
3530 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3531 pcretest, it used to include the size of the study block as part of its
3532 output. Unfortunately, the structure contains a field that has a different
3533 size on different hardware architectures. This meant that the tests that
3534 showed this size failed. As the block is currently always of a fixed size,
3535 this information isn't actually particularly useful in pcretest output, so
3536 I have just removed it.
3537
3538 8. Three pre-processor statements accidentally did not start in column 1.
3539 Sadly, there are *still* compilers around that complain, even though
3540 standard C has not required this for well over a decade. Sigh.
3541
3542 9. In pcretest, the code for checking callouts passed small integers in the
3543 callout_data field, which is a void * field. However, some picky compilers
3544 complained about the casts involved for this on 64-bit systems. Now
3545 pcretest passes the address of the small integer instead, which should get
3546 rid of the warnings.
3547
3548 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3549 both compile and run time, and gives an error if an invalid UTF-8 sequence
3550 is found. There is a option for disabling this check in cases where the
3551 string is known to be correct and/or the maximum performance is wanted.
3552
3553 11. In response to a bug report, I changed one line in Makefile.in from
3554
3555 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3556 to
3557 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3558
3559 to look similar to other lines, but I have no way of telling whether this
3560 is the right thing to do, as I do not use Windows. No doubt I'll get told
3561 if it's wrong...
3562
3563
3564 Version 4.3 21-May-03
3565 ---------------------
3566
3567 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3568 Makefile.
3569
3570 2. Some refactoring to improve the quality of the code:
3571
3572 (i) The utf8_table... variables are now declared "const".
3573
3574 (ii) The code for \cx, which used the "case flipping" table to upper case
3575 lower case letters, now just substracts 32. This is ASCII-specific,
3576 but the whole concept of \cx is ASCII-specific, so it seems
3577 reasonable.
3578
3579 (iii) PCRE was using its character types table to recognize decimal and
3580 hexadecimal digits in the pattern. This is silly, because it handles
3581 only 0-9, a-f, and A-F, but the character types table is locale-
3582 specific, which means strange things might happen. A private
3583 table is now used for this - though it costs 256 bytes, a table is
3584 much faster than multiple explicit tests. Of course, the standard
3585 character types table is still used for matching digits in subject
3586 strings against \d.
3587
3588 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3589 ending in _t are). So I've renamed it as ESC_tee.
3590
3591 3. The first argument for regexec() in the POSIX wrapper should have been
3592 defined as "const".
3593
3594 4. Changed pcretest to use malloc() for its buffers so that they can be
3595 Electric Fenced for debugging.
3596
3597 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3598 to read one or more bytes before the start of the subject string. Often this
3599 had no effect on PCRE's behaviour, but in some circumstances it could
3600 provoke a segmentation fault.
3601
3602 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3603 to try to read one or more bytes before the start of the subject string.
3604
3605 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3606 UTF-8 support could misbehave in various ways if the subject string
3607 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3608 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3609 back over UTF-8 characters.)
3610
3611
3612 Version 4.2 14-Apr-03
3613 ---------------------
3614
3615 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3616
3617 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3618 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3619 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3620 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3621 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3622 and BUILD_EXEEXT
3623 Note: automatic setting of the BUILD variables is not yet working
3624 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3625 compile-time but not at link-time
3626 [LINK]: use for linking executables only
3627 make different versions for Windows and non-Windows
3628 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3629 libraries
3630 [LINK_FOR_BUILD]: new variable
3631 [OBJEXT]: use throughout
3632 [EXEEXT]: use throughout
3633 <winshared>: new target
3634 <wininstall>: new target
3635 <dftables.o>: use native compiler
3636 <dftables>: use native linker
3637 <install>: handle Windows platform correctly
3638 <clean>: ditto
3639 <check>: ditto
3640 copy DLL to top builddir before testing
3641
3642 As part of these changes, -no-undefined was removed again. This was reported
3643 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3644 in any case.
3645
3646 3. Some tidies to get rid of compiler warnings:
3647
3648 . In the match_data structure, match_limit was an unsigned long int, whereas
3649 match_call_count was an int. I've made them both unsigned long ints.
3650
3651 . In pcretest the fact that a const uschar * doesn't automatically cast to
3652 a void * provoked a warning.
3653
3654 . Turning on some more compiler warnings threw up some "shadow" variables
3655 and a few more missing casts.
3656
3657 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3658 option, a class that contained a single character with a value between 128
3659 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3660
3661 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3662 option, a class that contained several characters, but with at least one
3663 whose value was between 128 and 255 caused PCRE to crash.
3664
3665
3666 Version 4.1 12-Mar-03
3667 ---------------------
3668
3669 1. Compiling with gcc -pedantic found a couple of places where casts were
3670 needed, and a string in dftables.c that was longer than standard compilers are
3671 required to support.
3672
3673 2. Compiling with Sun's compiler found a few more places where the code could
3674 be tidied up in order to avoid warnings.
3675
3676 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3677 first of these names is deprecated in the latest Autoconf in favour of the name
3678 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3679 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3680 analogy I have changed it to CFLAGS_FOR_BUILD.
3681
3682 4. Added -no-undefined to the linking command in the Makefile, because this is
3683 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3684 linking step for the pcreposix library.
3685
3686 5. PCRE was failing to diagnose the case of two named groups with the same
3687 name.
3688
3689 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3690 literal character that is needed in the subject for a match, and scans along to
3691 ensure that it is present before embarking on the full matching process. This
3692 saves time in cases of nested unlimited repeats that are never going to match.
3693 Problem: the scan can take a lot of time if the subject is very long (e.g.
3694 megabytes), thus penalizing straightforward matches. It is now done only if the
3695 amount of subject to be scanned is less than 1000 bytes.
3696
3697 7. A lesser problem with the same optimization is that it was recording the
3698 first character of an anchored pattern as "needed", thus provoking a search
3699 right along the subject, even when the first match of the pattern was going to
3700 fail. The "needed" character is now not set for anchored patterns, unless it
3701 follows something in the pattern that is of non-fixed length. Thus, it still
3702 fulfils its original purpose of finding quick non-matches in cases of nested
3703 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3704
3705
3706 Version 4.0 17-Feb-03
3707 ---------------------
3708
3709 1. If a comment in an extended regex that started immediately after a meta-item
3710 extended to the end of string, PCRE compiled incorrect data. This could lead to
3711 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3712
3713 2. Moved to autoconf 2.53 and libtool 1.4.2.
3714
3715 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3716 the special perltest8 script is no longer needed - all the tests can be run
3717 from a single perltest script.
3718
3719 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3720 by \s. It has now been removed in PCRE. This means it isn't recognized as
3721 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3722 class [:space:] *does* include VT, thereby creating a mess.
3723
3724 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3725 space and tab.
3726
3727 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3728 its new features into the main test script, reducing the number of scripts.
3729
3730 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3731 were backward compatible, and made the (?i) apply to the whole pattern, as if
3732 /i were given. Now it behaves more logically, and applies the option setting
3733 only to what follows. PCRE has been changed to follow suit. However, if it
3734 finds options settings right at the start of the pattern, it extracts them into
3735 the global options, as before. Thus, they show up in the info data.
3736
3737 8. Added support for the \Q...\E escape sequence. Characters in between are
3738 treated as literals. This is slightly different from Perl in that $ and @ are
3739 also handled as literals inside the quotes. In Perl, they will cause variable
3740 interpolation. Note the following examples:
3741
3742 Pattern PCRE matches Perl matches
3743
3744 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3745 \Qabc\$xyz\E abc\$xyz abc\$xyz
3746 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3747
3748 For compatibility with Perl, \Q...\E sequences are recognized inside character
3749 classes as well as outside them.
3750
3751 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3752 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3753 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3754 signed/unsigned warnings.
3755
3756 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3757 option for pcretest, so I've replaced it by a simple function that does just
3758 that job.
3759
3760 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3761 "pcregrep -".
3762
3763 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3764 Java package. This provides some syntactic sugar for simple cases of what my
3765 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3766 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3767 item, you can use this simplified notation. Note that only makes sense with
3768 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3769 greediness, whatever the setting of the PCRE_UNGREEDY option.
3770
3771 13. A change of greediness default within a pattern was not taking effect at
3772 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3773 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3774 was abstracted outside.
3775
3776 14. PCRE now supports the \G assertion. It is true when the current matching
3777 position is at the start point of the match. This differs from \A when the
3778 starting offset is non-zero. Used with the /g option of pcretest (or similar
3779 code), it works in the same way as it does for Perl's /g option. If all
3780 alternatives of a regex begin with \G, the expression is anchored to the start
3781 match position, and the "anchored" flag is set in the compiled expression.
3782
3783 15. Some bugs concerning the handling of certain option changes within patterns
3784 have been fixed. These applied to options other than (?ims). For example,
3785 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3786 been the other way round. Some of this was related to change 7 above.
3787
3788 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3789 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3790 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3791 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3792
3793 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3794 mode. Unlike ".", it always matches newline, whatever the setting of
3795 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3796 assertions. Perl allows it, but it doesn't (in general) work because it can't
3797 calculate the length of the lookbehind. At least, that's the case for Perl
3798 5.8.0 - I've been told they are going to document that it doesn't work in
3799 future.
3800
3801 18. Added an error diagnosis for escapes that PCRE does not support: these are
3802 \L, \l, \N, \P, \p, \U, \u, and \X.
3803
3804 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3805 reading past the end of the pattern in cases such as /[abcd/.
3806
3807 20. PCRE was getting more memory than necessary for patterns with classes that
3808 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3809
3810 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3811 compiling PCRE for use with Virtual Pascal.
3812
3813 22. Small fix to the Makefile to make it work properly if the build is done
3814 outside the source tree.
3815
3816 23. Added a new extension: a condition to go with recursion. If a conditional
3817 subpattern starts with (?(R) the "true" branch is used if recursion has
3818 happened, whereas the "false" branch is used only at the top level.
3819
3820 24. When there was a very long string of literal characters (over 255 bytes
3821 without UTF support, over 250 bytes with UTF support), the computation of how
3822 much memory was required could be incorrect, leading to segfaults or other
3823 strange effects.
3824
3825 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3826 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3827 there was a subsequent back reference to those brackets. This meant that, for
3828 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3829 possible to check for precisely this case. All we can do is abandon the
3830 optimization if .* occurs inside capturing brackets when there are any back
3831 references whatsoever. (See below for a better fix that came later.)
3832
3833 26. The handling of the optimization for finding the first character of a
3834 non-anchored pattern, and for finding a character that is required later in the
3835 match were failing in some cases. This didn't break the matching; it just
3836 failed to optimize when it could. The way this is done has been re-implemented.
3837
3838 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3839
3840 28. Added a new feature that provides some of the functionality that Perl
3841 provides with (?{...}). The facility is termed a "callout". The way it is done
3842 in PCRE is for the caller to provide an optional function, by setting
3843 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3844 global variable. By default it is unset, which disables all calling out. To get
3845 the function called, the regex must include (?C) at appropriate points. This
3846 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3847 This provides a means of identifying different callout points. When PCRE
3848 reaches such a point in the regex, if pcre_callout has been set, the external
3849 function is called. It is provided with data in a structure called
3850 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3851 matching continues; if it returns a non-zero value, the match at the current
3852 point fails. However, backtracking will occur if possible. [This was changed
3853 later and other features added - see item 49 below.]
3854
3855 29. pcretest is upgraded to test the callout functionality. It provides a
3856 callout function that displays information. By default, it shows the start of
3857 the match and the current position in the text. There are some new data escapes
3858 to vary what happens:
3859
3860 \C+ in addition, show current contents of captured substrings
3861 \C- do not supply a callout function
3862 \C!n return 1 when callout number n is reached
3863 \C!n!m return 1 when callout number n is reached for the mth time
3864
3865 30. If pcregrep was called with the -l option and just a single file name, it
3866 output "<stdin>" if a match was found, instead of the file name.
3867
3868 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3869 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3870 pcre_exec(). This saves a malloc/free per call. The default value of
3871 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3872 when configuring.
3873
3874 32. The default maximum size of a compiled pattern is 64K. There have been a
3875 few cases of people hitting this limit. The code now uses macros to handle the
3876 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3877 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3878 configuring. Tests 2 and 5 work only with 2-byte links because they output
3879 debugging information about compiled patterns.
3880
3881 33. Internal code re-arrangements:
3882
3883 (a) Moved the debugging function for printing out a compiled regex into
3884 its own source file (printint.c) and used #include to pull it into
3885 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3886 separate copies.
3887
3888 (b) Defined the list of op-code names for debugging as a macro in
3889 internal.h so that it is next to the definition of the opcodes.
3890
3891 (c) Defined a table of op-code lengths for simpler skipping along compiled
3892 code. This is again a macro in internal.h so that it is next to the
3893 definition of the opcodes.
3894
3895 34. Added support for recursive calls to individual subpatterns, along the
3896 lines of Robin Houston's patch (but implemented somewhat differently).
3897
3898 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3899 allow it to read and process whole directories in Win32. This code was
3900 contributed by Lionel Fourquaux; it has not been tested by me.
3901
3902 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3903 used to name a group. Names consist of alphanumerics and underscores, and must
3904 be unique. Back references use the syntax (?P=name) and recursive calls use
3905 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3906 numbers. The function pcre_fullinfo() can be used after compilation to extract
3907 a name/number map. There are three relevant calls:
3908
3909 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3910 PCRE_INFO_NAMECOUNT yields the number of entries
3911 PCRE_INFO_NAMETABLE yields a pointer to the map.
3912
3913 The map is a vector of fixed-size entries. The size of each entry depends on
3914 the length of the longest name used. The first two bytes of each entry are the
3915 group number, most significant byte first. There follows the corresponding
3916 name, zero terminated. The names are in alphabetical order.
3917
3918 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3919 case instead of 255. Making it the same both with and without UTF-8 support
3920 means that the same test output works with both.
3921
3922 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3923 calling malloc() with a zero argument.
3924
3925 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3926 optimization. I've improved things by keeping a bitmap of backreferences with
3927 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3928 fact referenced, the optimization can be applied. It is unlikely that a
3929 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3930 the match to follow \n) will appear inside brackets with a number greater than
3931 31, but if it does, any back reference > 31 suppresses the optimization.
3932
3933 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3934 of disabling numbered capturing parentheses. Any opening parenthesis that is
3935 not followed by ? behaves as if it were followed by ?: but named parentheses
3936 can still be used for capturing (and they will acquire numbers in the usual
3937 way).
3938
3939 41. Redesigned the return codes from the match() function into yes/no/error so
3940 that errors can be passed back from deep inside the nested calls. A malloc
3941 failure while inside a recursive subpattern call now causes the
3942 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3943
3944 42. It is now possible to set a limit on the number of times the match()
3945 function is called in a call to pcre_exec(). This facility makes it possible to
3946 limit the amount of recursion and backtracking, though not in a directly
3947 obvious way, because the match() function is used in a number of different
3948 circumstances. The count starts from zero for each position in the subject
3949 string (for non-anchored patterns). The default limit is, for compatibility, a
3950 large number, namely 10 000 000. You can change this in two ways:
3951
3952 (a) When configuring PCRE before making, you can use --with-match-limit=n
3953 to set a default value for the compiled library.
3954
3955 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3956 a different value is set. See 45 below.
3957
3958 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3959
3960 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3961 of things that can be changed at compile time. The first argument specifies
3962 what is wanted and the second points to where the information is to be placed.
3963 The current list of available information is:
3964
3965 PCRE_CONFIG_UTF8
3966
3967 The output is an integer that is set to one if UTF-8 support is available;
3968 otherwise it is set to zero.
3969
3970 PCRE_CONFIG_NEWLINE
3971
3972 The output is an integer that it set to the value of the code that is used for
3973 newline. It is either LF (10) or CR (13).
3974
3975 PCRE_CONFIG_LINK_SIZE
3976
3977 The output is an integer that contains the number of bytes used for internal
3978 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3979
3980 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3981
3982 The output is an integer that contains the threshold above which the POSIX
3983 interface uses malloc() for output vectors. See item 31 above.
3984
3985 PCRE_CONFIG_MATCH_LIMIT
3986
3987 The output is an unsigned integer that contains the default limit of the number
3988 of match() calls in a pcre_exec() execution. See 42 above.
3989
3990 44. pcretest has been upgraded by the addition of the -C option. This causes it
3991 to extract all the available output from the new pcre_config() function, and to
3992 output it. The program then exits immediately.
3993
3994 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3995 order to support additional features. One way would have been to define
3996 pcre_exec2() (for example) with extra arguments, but this would not have been
3997 extensible, and would also have required all calls to the original function to
3998 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3999 is used for passing in "extra" data from pcre_study().
4000
4001 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4002 contains the following fields:
4003
4004 flags a bitmap indicating which of the following fields are set
4005 study_data opaque data from pcre_study()
4006 match_limit a way of specifying a limit on match() calls for a specific
4007 call to pcre_exec()
4008 callout_data data for callouts (see 49 below)
4009
4010 The flag bits are also defined in pcre.h, and are
4011
4012 PCRE_EXTRA_STUDY_DATA
4013 PCRE_EXTRA_MATCH_LIMIT
4014 PCRE_EXTRA_CALLOUT_DATA
4015
4016 The pcre_study() function now returns one of these new pcre_extra blocks, with
4017 the actual study data pointed to by the study_data field, and the
4018 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4019 before. That is, this change is entirely upwards-compatible and requires no
4020 change to existing code.
4021
4022 If you want to pass in additional data to pcre_exec(), you can either place it
4023 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4024 block.
4025
4026 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4027 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4028 times with different match limits, until it finds the minimum value needed for
4029 pcre_exec() to complete. The value is then output. This can be instructive; for
4030 most simple matches the number is quite small, but for pathological cases it
4031 gets very large very quickly.
4032
4033 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4034 returns the size of the data block pointed to by the study_data field in a
4035 pcre_extra block, that is, the value that was passed as the argument to
4036 pcre_malloc() when PCRE was getting memory in which to place the information
4037 created by pcre_study(). The fourth argument should point to a size_t variable.
4038 pcretest has been extended so that this information is shown after a successful
4039 pcre_study() call when information about the compiled regex is being displayed.
4040
4041 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4042 because what follows is always an absolute path. (Later: it turns out that this
4043 is more than cosmetic for MinGW, because it doesn't like empty path
4044 components.)
4045
4046 49. Some changes have been made to the callout feature (see 28 above):
4047
4048 (i) A callout function now has three choices for what it returns:
4049
4050 0 => success, carry on matching
4051 > 0 => failure at this point, but backtrack if possible
4052 < 0 => serious error, return this value from pcre_exec()
4053
4054 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4055 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4056 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4057 use by callout functions. It will never be used by PCRE itself.
4058
4059 (ii) The pcre_extra structure (see 45 above) has a void * field called
4060 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4061 pcre_callout_block structure has a field of the same name. The contents of
4062 the field passed in the pcre_extra structure are passed to the callout
4063 function in the corresponding field in the callout block. This makes it
4064 easier to use the same callout-containing regex from multiple threads. For
4065 testing, the pcretest program has a new data escape
4066
4067 \C*n pass the number n (may be negative) as callout_data
4068
4069 If the callout function in pcretest receives a non-zero value as
4070 callout_data, it returns that value.
4071
4072 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4073 there were some redundant $(CFLAGS) in commands that are now specified as
4074 $(LINK), which already includes $(CFLAGS).
4075
4076 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4077 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4078 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4079 one-byte characters throughout. Note that case-insensitive matching applies
4080 only to characters whose values are less than 256. PCRE doesn't support the
4081 notion of cases for higher-valued characters.
4082
4083 (i) A character class whose characters are all within 0-255 is handled as
4084 a bit map, and the map is inverted for negative classes. Previously, a
4085 character > 255 always failed to match such a class; however it should
4086 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4087
4088 (ii) A negated character class with a single character < 255 is coded as
4089 "not this character" (OP_NOT). This wasn't working properly when the test
4090 character was multibyte, either singly or repeated.
4091
4092 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4093 mode, for example: \x{100}{2,3}.
4094
4095 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4096 singly or repeated) now correctly test multibyte characters. However,
4097 PCRE doesn't recognize any characters with values greater than 255 as
4098 digits, spaces, or word characters. Such characters always match \D, \S,
4099 and \W, and never match \d, \s, or \w.
4100
4101 (v) Classes may now contain characters and character ranges with values
4102 greater than 255. For example: [ab\x{100}-\x{400}].
4103
4104 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4105 PCRE in UTF-8 mode.
4106
4107 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4108 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4109 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4110 value.)
4111
4112 53. The single man page has become too large. I have therefore split it up into
4113 a number of separate man pages. These also give rise to individual HTML pages;
4114 these are now put in a separate directory, and there is an index.html page that
4115 lists them all. Some hyperlinking between the pages has been installed.
4116
4117 54. Added convenience functions for handling named capturing parentheses.
4118
4119 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4120 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4121 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4122 are faulted.
4123
4124 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4125 calling configure. These values are used when compiling the dftables.c program
4126 which is run to generate the source of the default character tables. They
4127 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4128 you will need to set these values.
4129
4130 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4131
4132
4133 Version 3.9 02-Jan-02
4134 ---------------------
4135
4136 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4137
4138 2. If --disable-static was given, the building process failed when trying to
4139 build pcretest and pcregrep. (For some reason it was using libtool to compile
4140 them, which is not right, as they aren't part of the library.)
4141
4142
4143 Version 3.8 18-Dec-01
4144 ---------------------
4145
4146 1. The experimental UTF-8 code was completely screwed up. It was packing the
4147 bytes in the wrong order. How dumb can you get?
4148
4149
4150 Version 3.7 29-Oct-01
4151 ---------------------
4152
4153 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4154 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4155 this didn't happen under Solaris 8, where I normally test things.
4156
4157 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4158 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4159 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4160 this fix an hour or so after the initial 3.7 release.)
4161
4162
4163 Version 3.6 23-Oct-01
4164 ---------------------
4165
4166 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4167 offsets passed as NULL with zero offset count.
4168
4169 2. The config.guess and config.sub files had not been updated when I moved to
4170 the latest autoconf.
4171
4172
4173 Version 3.5 15-Aug-01
4174 ---------------------
4175
4176 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4177 had been forgotten.
4178
4179 2. By using declared but undefined structures, we can avoid using "void"
4180 definitions in pcre.h while keeping the internal definitions of the structures
4181 private.
4182
4183 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4184 user point of view, this means that both static and shared libraries are built
4185 by default, but this can be individually controlled. More of the work of
4186 handling this static/shared cases is now inside libtool instead of PCRE's make
4187 file.
4188
4189 4. The pcretest utility is now installed along with pcregrep because it is
4190 useful for users (to test regexs) and by doing this, it automatically gets
4191 relinked by libtool. The documentation has been turned into a man page, so
4192 there are now .1, .txt, and .html versions in /doc.
4193
4194 5. Upgrades to pcregrep:
4195 (i) Added long-form option names like gnu grep.
4196 (ii) Added --help to list all options with an explanatory phrase.
4197 (iii) Added -r, --recursive to recurse into sub-directories.
4198 (iv) Added -f, --file to read patterns from a file.
4199
4200 6. pcre_exec() was referring to its "code" argument before testing that
4201 argument for NULL (and giving an error if it was NULL).
4202
4203 7. Upgraded Makefile.in to allow for compiling in a different directory from
4204 the source directory.
4205
4206 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4207 options bits, the pointer it was passed was to an int instead of to an unsigned
4208 long int. This mattered only on 64-bit systems.
4209
4210 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4211 generated) instead of pcre.in, which it its source. Also made the same change
4212 in several of the .c files.
4213
4214 10. A new release of gcc defines printf() as a macro, which broke pcretest
4215 because it had an ifdef in the middle of a string argument for printf(). Fixed
4216 by using separate calls to printf().
4217
4218 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4219 script, to force use of CR or LF instead of \n in the source. On non-Unix
4220 systems, the value can be set in config.h.
4221
4222 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4223 absolute limit. Changed the text of the error message to make this clear, and
4224 likewise updated the man page.
4225
4226 13. The limit of 99 on the number of capturing subpatterns has been removed.
4227 The new limit is 65535, which I hope will not be a "real" limit.
4228
4229
4230 Version 3.4 22-Aug-00
4231 ---------------------
4232
4233 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4234
4235 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4236
4237
4238 Version 3.3 01-Aug-00
4239 ---------------------
4240
4241 1. If an octal character was given, but the value was greater than \377, it
4242 was not getting masked to the least significant bits, as documented. This could
4243 lead to crashes in some systems.
4244
4245 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4246 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4247
4248 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4249 These just pass their arguments on to (pcre_free)(), but they are provided
4250 because some uses of PCRE bind it to non-C systems that can call its functions,
4251 but cannot call free() or pcre_free() directly.
4252
4253 4. Add "make test" as a synonym for "make check". Corrected some comments in
4254 the Makefile.
4255
4256 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4257 Makefile.
4258
4259 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4260 command called pgrep for grepping around the active processes.
4261
4262 7. Added the beginnings of support for UTF-8 character strings.
4263
4264 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4265 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4266 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4267 out for the ar command.)
4268
4269
4270 Version 3.2 12-May-00
4271 ---------------------
4272
4273 This is purely a bug fixing release.
4274
4275 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4276 of ZA. This was just one example of several cases that could provoke this bug,
4277 which was introduced by change 9 of version 2.00. The code for breaking
4278 infinite loops after an iteration that matches an empty string was't working
4279 correctly.
4280
4281 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4282 when matched against abbab (for example). After matching an empty string, it
4283 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4284 caused it to match further down the string than it should.
4285
4286 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4287 was there because it doesn't seem to be needed, and it causes trouble on some
4288 systems, as it is not a Standard C header. It has been removed.
4289
4290 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4291 were reported on the Macintosh. The changes were from
4292
4293 while ((c = *(++ptr)) != 0 && c != '\n');
4294 to
4295 while ((c = *(++ptr)) != 0 && c != '\n') ;
4296
4297 Totally extraordinary, but if that's what it takes...
4298
4299 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4300 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4301 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4302 assumes the way PCRE uses memmove() (always moving upwards).
4303
4304 6. PCRE is being used in one environment where strchr() is not available. There
4305 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4306 faster code anyway.
4307
4308
4309 Version 3.1 09-Feb-00
4310 ---------------------
4311
4312 The only change in this release is the fixing of some bugs in Makefile.in for
4313 the "install" target:
4314
4315 (1) It was failing to install pcreposix.h.
4316
4317 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4318
4319
4320 Version 3.0 01-Feb-00
4321 ---------------------
4322
4323 1. Add support for the /+ modifier to perltest (to output $` like it does in
4324 pcretest).
4325
4326 2. Add support for the /g modifier to perltest.
4327
4328 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4329 matches null strings.
4330
4331 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4332 pattern. Perl treats empty patterns specially - it reuses the most recent
4333 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4334 effect.
4335
4336 5. The POSIX interface was broken in that it was just handing over the POSIX
4337 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4338 required a bigger vector, with some working space on the end. This means that
4339 the POSIX wrapper now has to get and free some memory, and copy the results.
4340
4341 6. Added some simple autoconf support, placing the test data and the
4342 documentation in separate directories, re-organizing some of the
4343 information files, and making it build pcre-config (a GNU standard). Also added
4344 libtool support for building PCRE as a shared library, which is now the
4345 default.
4346
4347 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4348 09 are not valid octal constants. Single digits will be used for minor values
4349 less than 10.
4350
4351 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4352 existing programs that set these in the POSIX interface can use PCRE without
4353 modification.
4354
4355 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4356 return all that pcre_info() returns, plus additional data. The pcre_info()
4357 function is retained for compatibility, but is considered to be obsolete.
4358
4359 10. Added experimental recursion feature (?R) to handle one common case that
4360 Perl 5.6 will be able to do with (?p{...}).
4361
4362 11. Added support for POSIX character classes like [:alpha:], which Perl is
4363 adopting.
4364
4365
4366 Version 2.08 31-Aug-99
4367 ----------------------
4368
4369 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4370 trying to match at the startoffset position, but instead was moving forward to
4371 the next newline as if a previous match had failed.
4372
4373 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4374 and could get into a loop if a null string was matched other than at the start
4375 of the subject.
4376
4377 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4378 be distinguished at compile time, and for completeness also added PCRE_DATE.
4379
4380 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4381 in GnuWin32 environments.
4382
4383
4384 Version 2.07 29-Jul-99
4385 ----------------------
4386
4387 1. The documentation is now supplied in plain text form and HTML as well as in
4388 the form of man page sources.
4389
4390 2. C++ compilers don't like assigning (void *) values to other pointer types.
4391 In particular this affects malloc(). Although there is no problem in Standard
4392 C, I've put in casts to keep C++ compilers happy.
4393
4394 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4395 should be (const char *).
4396
4397 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4398 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4399 However, I haven't made this a standard facility. The documentation doesn't
4400 mention it, and the Makefile doesn't support it.
4401
4402 5. The Makefile now contains an "install" target, with editable destinations at
4403 the top of the file. The pcretest program is not installed.
4404
4405 6. pgrep -V now gives the PCRE version number and date.
4406
4407 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4408 causing the entire string to be ignored, instead of just the last character.
4409
4410 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4411 non-matching string, it can take a very, very long time, even for strings of
4412 quite modest length, because of the nested recursion. PCRE now does better in
4413 some of these cases. It does this by remembering the last required literal
4414 character in the pattern, and pre-searching the subject to ensure it is present
4415 before running the real match. In other words, it applies a heuristic to detect
4416 some types of certain failure quickly, and in the above example, if presented
4417 with a string that has no trailing " it gives "no match" very quickly.
4418
4419 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4420 other alternatives are tried instead.
4421
4422
4423 Version 2.06 09-Jun-99
4424 ----------------------
4425
4426 1. Change pcretest's output for amount of store used to show just the code
4427 space, because the remainder (the data block) varies in size between 32-bit and
4428 64-bit systems.
4429
4430 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4431 start matching at. This allows lookbehinds to work when searching for multiple
4432 occurrences in a string.
4433
4434 3. Added additional options to pcretest for testing multiple occurrences:
4435
4436 /+ outputs the rest of the string that follows a match
4437 /g loops for multiple occurrences, using the new startoffset argument
4438 /G loops for multiple occurrences by passing an incremented pointer
4439
4440 4. PCRE wasn't doing the "first character" optimization for patterns starting
4441 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4442 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4443 the letter 'x'. On long subject strings, this gives a significant speed-up.
4444
4445
4446 Version 2.05 21-Apr-99
4447 ----------------------
4448
4449 1. Changed the type of magic_number from int to long int so that it works
4450 properly on 16-bit systems.
4451
4452 2. Fixed a bug which caused patterns starting with .* not to work correctly
4453 when the subject string contained newline characters. PCRE was assuming
4454 anchoring for such patterns in all cases, which is not correct because .* will
4455 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4456 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4457 must be retried after every newline in the subject.
4458
4459
4460 Version 2.04 18-Feb-99
4461 ----------------------
4462
4463 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4464 computation of the store needed to hold the pattern was incorrect (too large).
4465 If such patterns were nested a few deep, this could multiply and become a real
4466 problem.
4467
4468 2. Added /M option to pcretest to show the memory requirement of a specific
4469 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4470
4471 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4472 compiled in such a way that the backtracking after subsequent failure was
4473 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4474 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4475
4476
4477 Version 2.03 02-Feb-99
4478 ----------------------
4479
4480 1. Fixed typo and small mistake in man page.
4481
4482 2. Added 4th condition (GPL supersedes if conflict) and created separate
4483 LICENCE file containing the conditions.
4484
4485 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4486 Perl, that is the internal \ allows the delimiter to be included in the
4487 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4488 the final delimiter, add \ to the end of the pattern (to test the error).
4489
4490 4. Added the convenience functions for extracting substrings after a successful
4491 match. Updated pcretest to make it able to test these functions.
4492
4493
4494 Version 2.02 14-Jan-99
4495 ----------------------
4496
4497 1. Initialized the working variables associated with each extraction so that
4498 their saving and restoring doesn't refer to uninitialized store.
4499
4500 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4501 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4502 fix the problem.
4503
4504 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4505 calls, and wasn't printing the correct value for compiling calls. Increased the
4506 default value of LOOPREPEAT, and the number of significant figures in the
4507 times.
4508
4509 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4510
4511 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4512 a building problem on Windows NT with a FAT file system.
4513
4514
4515 Version 2.01 21-Oct-98
4516 ----------------------
4517
4518 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4519 to character tables built by pcre_maketables() in the current locale. If NULL
4520 is passed, the default tables are used.
4521
4522
4523 Version 2.00 24-Sep-98
4524 ----------------------
4525
4526 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4527 it any more.
4528
4529 2. Allow quantification of (?>) groups, and make it work correctly.
4530
4531 3. The first character computation wasn't working for (?>) groups.
4532
4533 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4534 end of the subject) and add 5.005's \z, which really does match only at the
4535 very end of the subject.
4536
4537 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4538
4539 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4540 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4541 localized options. All options to pcre_study() were also removed.
4542
4543 7. Add other new features from 5.005:
4544
4545 $(?<= positive lookbehind
4546 $(?<! negative lookbehind
4547 (?imsx-imsx) added the unsetting capability
4548 such a setting is global if at outer level; local otherwise
4549 (?imsx-imsx:) non-capturing groups with option setting
4550 (?(cond)re|re) conditional pattern matching
4551
4552 A backreference to itself in a repeated group matches the previous
4553 captured string.
4554
4555 8. General tidying up of studying (both automatic and via "study")
4556 consequential on the addition of new assertions.
4557
4558 9. As in 5.005, unlimited repeated groups that could match an empty substring
4559 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4560 runtime if any iteration does actually match an empty substring.
4561
4562 10. Include the RunTest script in the distribution.
4563
4564 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4565 discrepancies, some of which were old and were also with respect to 5.004. They
4566 have now been fixed.
4567
4568
4569 Version 1.09 28-Apr-98
4570 ----------------------
4571
4572 1. A negated single character class followed by a quantifier with a minimum
4573 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4574 program crashes, or just wrong answers. This did not apply to negated classes
4575 containing more than one character, or to minima other than one.
4576
4577
4578 Version 1.08 27-Mar-98
4579 ----------------------
4580
4581 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4582
4583 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4584 latter must appear before anything that relies on it in the pattern.
4585
4586
4587 Version 1.07 16-Feb-98
4588 ----------------------
4589
4590 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4591 repeat of a potentially empty string).
4592
4593
4594 Version 1.06 23-Jan-98
4595 ----------------------
4596
4597 1. Added Markus Oberhumer's little patches for C++.
4598
4599 2. Literal strings longer than 255 characters were broken.
4600
4601
4602 Version 1.05 23-Dec-97
4603 ----------------------
4604
4605 1. Negated character classes containing more than one character were failing if
4606 PCRE_CASELESS was set at run time.
4607
4608
4609 Version 1.04 19-Dec-97
4610 ----------------------
4611
4612 1. Corrected the man page, where some "const" qualifiers had been omitted.
4613
4614 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4615 input syntax.
4616
4617 3. Fixed memory leak which occurred when a regex with back references was
4618 matched with an offsets vector that wasn't big enough. The temporary memory
4619 that is used in this case wasn't being freed if the match failed.
4620
4621 4. Tidied pcretest to ensure it frees memory that it gets.
4622
4623 5. Temporary memory was being obtained in the case where the passed offsets
4624 vector was exactly big enough.
4625
4626 6. Corrected definition of offsetof() from change 5 below.
4627
4628 7. I had screwed up change 6 below and broken the rules for the use of
4629 setjmp(). Now fixed.
4630
4631
4632 Version 1.03 18-Dec-97
4633 ----------------------
4634
4635 1. A erroneous regex with a missing opening parenthesis was correctly
4636 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4637 on some systems.
4638
4639 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4640 it was reported that one broken compiler failed on the former because "code" is
4641 also an independent variable.
4642
4643 3. The erroneous regex a[]b caused an array overrun reference.
4644
4645 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4646 fail on data ending with that character. (It was going on too far, and checking
4647 the next character, typically a binary zero.) This was specific to the
4648 optimized code for single-character negative classes.
4649
4650 5. Added a contributed patch from the TIN world which does the following:
4651
4652 + Add an undef for memmove, in case the the system defines a macro for it.
4653
4654 + Add a definition of offsetof(), in case there isn't one. (I don't know
4655 the reason behind this - offsetof() is part of the ANSI standard - but
4656 it does no harm).
4657
4658 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4659 most of the places where whitespace preceded '#'. I have given up and
4660 allowed the remaining 2 cases to be at the margin.
4661
4662 + Rename some variables in pcre to eliminate shadowing. This seems very
4663 pedantic, but does no harm, of course.
4664
4665 6. Moved the call to setjmp() into its own function, to get rid of warnings
4666 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4667
4668 7. Constructs such as \d{8,} were compiling into the equivalent of
4669 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4670 outcome, but in this particular case used more store than had been allocated,
4671 which caused the bug to be discovered because it threw up an internal error.
4672
4673 8. The debugging code in both pcre and pcretest for outputting the compiled
4674 form of a regex was going wrong in the case of back references followed by
4675 curly-bracketed repeats.
4676
4677
4678 Version 1.02 12-Dec-97
4679 ----------------------
4680
4681 1. Typos in pcre.3 and comments in the source fixed.
4682
4683 2. Applied a contributed patch to get rid of places where it used to remove
4684 'const' from variables, and fixed some signed/unsigned and uninitialized
4685 variable warnings.
4686
4687 3. Added the "runtest" target to Makefile.
4688
4689 4. Set default compiler flag to -O2 rather than just -O.
4690
4691
4692 Version 1.01 19-Nov-97
4693 ----------------------
4694
4695 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4696 like /([ab]*)*/, that is, for classes with more than one character in them.
4697
4698 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4699 as /((?>a*))*/ (a PCRE_EXTRA facility).
4700
4701
4702 Version 1.00 18-Nov-97
4703 ----------------------
4704
4705 1. Added compile-time macros to support systems such as SunOS4 which don't have
4706 memmove() or strerror() but have other things that can be used instead.
4707
4708 2. Arranged that "make clean" removes the executables.
4709
4710
4711 Version 0.99 27-Oct-97
4712 ----------------------
4713
4714 1. Fixed bug in code for optimizing classes with only one character. It was
4715 initializing a 32-byte map regardless, which could cause it to run off the end
4716 of the memory it had got.
4717
4718 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4719
4720
4721 Version 0.98 22-Oct-97
4722 ----------------------
4723
4724 1. Fixed bug in code for handling temporary memory usage when there are more
4725 back references than supplied space in the ovector. This could cause segfaults.
4726
4727
4728 Version 0.97 21-Oct-97
4729 ----------------------
4730
4731 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4732
4733 2. Optimized negated single characters not to use a bit map.
4734
4735 3. Brought error texts together as macro definitions; clarified some of them;
4736 fixed one that was wrong - it said "range out of order" when it meant "invalid
4737 escape sequence".
4738
4739 4. Changed some char * arguments to const char *.
4740
4741 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4742
4743 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4744 pcretest.
4745
4746
4747 Version 0.96 16-Oct-97
4748 ----------------------
4749
4750 1. Added a simple "pgrep" utility to the distribution.
4751
4752 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4753 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4754 where "ddd" means "one or more decimal digits".
4755
4756 3. Fixed serious bug. If a pattern had a back reference, but the call to
4757 pcre_exec() didn't supply a large enough ovector to record the related
4758 identifying subpattern, the match always failed. PCRE now remembers the number
4759 of the largest back reference, and gets some temporary memory in which to save
4760 the offsets during matching if necessary, in order to ensure that
4761 backreferences always work.
4762
4763 4. Increased the compatibility with Perl in a number of ways:
4764
4765 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4766 to request this handling. The option can be set at compile or exec time.
4767
4768 (b) $ matches before a terminating newline by default; an option
4769 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4770 mode). The option can be set at compile or exec time.
4771
4772 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4773 the same as Perl's. If the decimal number it represents is less than 10
4774 or there aren't that many previous left capturing parentheses, an octal
4775 escape is read. Inside a character class, it's always an octal escape,
4776 even if it is a single digit.
4777
4778 (d) An escaped but undefined alphabetic character is taken as a literal,
4779 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4780 escapes.
4781
4782 (e) {0} is now permitted. (The previous item is removed from the compiled
4783 pattern).
4784
4785 5. Changed all the names of code files so that the basic parts are no longer
4786 than 10 characters, and abolished the teeny "globals.c" file.
4787
4788 6. Changed the handling of character classes; they are now done with a 32-byte
4789 bit map always.
4790
4791 7. Added the -d and /D options to pcretest to make it possible to look at the
4792 internals of compilation without having to recompile pcre.
4793
4794
4795 Version 0.95 23-Sep-97
4796 ----------------------
4797
4798 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4799 \x20 at the start of a run of normal characters. These were being treated as
4800 real characters, instead of the source characters being re-checked.
4801
4802
4803 Version 0.94 18-Sep-97
4804 ----------------------
4805
4806 1. The functions are now thread-safe, with the caveat that the global variables
4807 containing pointers to malloc() and free() or alternative functions are the
4808 same for all threads.
4809
4810 2. Get pcre_study() to generate a bitmap of initial characters for non-
4811 anchored patterns when this is possible, and use it if passed to pcre_exec().
4812
4813
4814 Version 0.93 15-Sep-97
4815 ----------------------
4816
4817 1. /(b)|(:+)/ was computing an incorrect first character.
4818
4819 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4820 but not actually doing anything yet.
4821
4822 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4823 as Perl does (e.g. [-az] or [az-]).
4824
4825 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4826 all possible positions.
4827
4828 5. Split up into different modules to avoid including unneeded functions in a
4829 compiled binary. However, compile and exec are still in one module. The "study"
4830 function is split off.
4831
4832 6. The character tables are now in a separate module whose source is generated
4833 by an auxiliary program - but can then be edited by hand if required. There are
4834 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4835 toupper() in the code.
4836
4837 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4838 make them global. Abolish the function for setting them, as the caller can now
4839 set them directly.
4840
4841
4842 Version 0.92 11-Sep-97
4843 ----------------------
4844
4845 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4846 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4847
4848 2. Caseless matching was not working in character classes if the characters in
4849 the pattern were in upper case.
4850
4851 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4852
4853 4. Make PCRE_ANCHORED public and accept as a compile option.
4854
4855 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4856 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4857 pass them.
4858
4859 6. Give an error if bad option bits passed at compile or run time.
4860
4861 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4862 pcretest to cause it to pass that flag.
4863
4864 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4865 options, and the first character, if set.
4866
4867 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4868
4869
4870 Version 0.91 10-Sep-97
4871 ----------------------
4872
4873 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4874 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4875
4876 2. PCRE was looping on encountering an indefinitely repeated back reference to
4877 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4878 Perl does - treats the match as successful.
4879
4880 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5