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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 975 - (show annotations) (download)
Sat Jun 2 11:03:06 2012 UTC (2 years, 4 months ago) by ph10
File size: 231239 byte(s)
Document update for 8.31-RC1 test release.

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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12