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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 979 - (show annotations) (download)
Sun Jun 17 19:08:41 2012 UTC (2 years, 2 months ago) by ph10
File size: 232385 byte(s)
Fix DFA bug (3 cases) when UTF code was being obeyed in non-UTF mode.

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

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