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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1009 - (show annotations)
Wed Aug 22 12:01:22 2012 UTC (7 years, 3 months ago) by zherczeg
File size: 234595 byte(s)
Improve the matching speed of capturing brackets.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.32
5 ------------
6
7 1. Improved JIT compiler optimizations for first character search and single
8 character iterators.
9
10 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11 Patch by Daniel Richard G.
12
13 3. Single character iterator optimizations in the JIT compiler.
14
15 4. Improved JIT compiler optimizations for character ranges.
16
17 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18 Reported by Giuseppe D'Angelo.
19
20 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21 of a line, was being set incorrectly in cases where .* appeared inside
22 atomic brackets at the start of a pattern, or where there was a subsequent
23 *PRUNE or *SKIP.
24
25 7. Improved instruction cache flush for POWER/PowerPC.
26 Patch by Daniel Richard G.
27
28 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
29 grep:
30
31 (a) There is now no limit to the number of patterns to be matched.
32
33 (b) An error is given if a pattern is too long.
34
35 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
36 are now supported.
37
38 (d) --exclude-from and --include-from (multiple use) have been added.
39
40 (e) Exclusions and inclusions now apply to all files and directories, not
41 just to those obtained from scanning a directory recursively.
42
43 (f) Multiple uses of -f and --file-list are now supported.
44
45 (g) In a Windows environment, the default for -d has been changed from
46 "read" (the GNU grep default) to "skip", because otherwise the presence
47 of a directory in the file list provokes an error.
48
49 (h) The documentation has been revised and clarified in places.
50
51 9. Improve the matching speed of capturing brackets.
52
53
54 Version 8.31 06-July-2012
55 -------------------------
56
57 1. Fixing a wrong JIT test case and some compiler warnings.
58
59 2. Removed a bashism from the RunTest script.
60
61 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
62 to unsigned type, result still unsigned" that was given by an MS compiler
63 on encountering the code "-sizeof(xxx)".
64
65 4. Partial matching support is added to the JIT compiler.
66
67 5. Fixed several bugs concerned with partial matching of items that consist
68 of more than one character:
69
70 (a) /^(..)\1/ did not partially match "aba" because checking references was
71 done on an "all or nothing" basis. This also applied to repeated
72 references.
73
74 (b) \R did not give a hard partial match if \r was found at the end of the
75 subject.
76
77 (c) \X did not give a hard partial match after matching one or more
78 characters at the end of the subject.
79
80 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
81 a partial match for the string "\r".
82
83 (e) When newline was set to CRLF, the metacharacter "." did not recognize
84 a partial match for a CR character at the end of the subject string.
85
86 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
87 running pcretest, the text "(JIT)" added to the output whenever JIT is
88 actually used to run the match.
89
90 7. Individual JIT compile options can be set in pcretest by following -s+[+]
91 or /S+[+] with a digit between 1 and 7.
92
93 8. OP_NOT now supports any UTF character not just single-byte ones.
94
95 9. (*MARK) control verb is now supported by the JIT compiler.
96
97 10. The command "./RunTest list" lists the available tests without actually
98 running any of them. (Because I keep forgetting what they all are.)
99
100 11. Add PCRE_INFO_MAXLOOKBEHIND.
101
102 12. Applied a (slightly modified) user-supplied patch that improves performance
103 when the heap is used for recursion (compiled with --disable-stack-for-
104 recursion). Instead of malloc and free for each heap frame each time a
105 logical recursion happens, frames are retained on a chain and re-used where
106 possible. This sometimes gives as much as 30% improvement.
107
108 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
109 call.
110
111 14. As documented, (*COMMIT) is now confined to within a positive assertion.
112
113 15. It is now possible to link pcretest with libedit as an alternative to
114 libreadline.
115
116 16. (*COMMIT) control verb is now supported by the JIT compiler.
117
118 17. The Unicode data tables have been updated to Unicode 6.1.0.
119
120 18. Added --file-list option to pcregrep.
121
122 19. Added binary file support to pcregrep, including the -a, --binary-files,
123 -I, and --text options.
124
125 20. The madvise function is renamed for posix_madvise for QNX compatibility
126 reasons. Fixed by Giuseppe D'Angelo.
127
128 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
129
130 22. Changed the option for creating symbolic links for 16-bit man pages from
131 -s to -sf so that re-installing does not cause issues.
132
133 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
134
135 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
136 matches in certain environments (the workspace was not being correctly
137 retained). Also added to pcre_dfa_exec() a simple plausibility check on
138 some of the workspace data at the beginning of a restart.
139
140 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
141 was not doing so when it should - probably a typo introduced by SVN 528
142 (change 8.10/14).
143
144 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
145 \w+ when the character tables indicated that \x{c4} was a word character.
146 There were several related cases, all because the tests for doing a table
147 lookup were testing for characters less than 127 instead of 255.
148
149 27. If a pattern contains capturing parentheses that are not used in a match,
150 their slots in the ovector are set to -1. For those that are higher than
151 any matched groups, this happens at the end of processing. In the case when
152 there were back references that the ovector was too small to contain
153 (causing temporary malloc'd memory to be used during matching), and the
154 highest capturing number was not used, memory off the end of the ovector
155 was incorrectly being set to -1. (It was using the size of the temporary
156 memory instead of the true size.)
157
158 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
159 ovector size, it uses memory at the end of the block that it has got.
160
161 29. Check for an overlong MARK name and give an error at compile time. The
162 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
163
164 30. JIT compiler update.
165
166 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
167 Rill for the patch.
168
169 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
170
171 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
172
173 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
174 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
175 was enabled.
176
177 35. Improve JIT code generation for greedy plus quantifier.
178
179 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
180 1 to "aa" instead of to an empty string. The bug affected repeated groups
181 that could potentially match an empty string.
182
183 37. Optimizing single character iterators in JIT.
184
185 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
186 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
187 codepoints that are too big for the mode are faulted, and in a UTF mode,
188 disallowed codepoints are also faulted.
189
190 39. If PCRE was compiled with UTF support, in three places in the DFA
191 matcher there was code that should only have been obeyed in UTF mode, but
192 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
193 processing when bytes with values greater than 127 were present. In 16-bit
194 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
195 both cases the values are those that cannot be the first data item in a UTF
196 character. The three items that might have provoked this were recursions,
197 possessively repeated groups, and atomic groups.
198
199 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
200 and pcregrep, because some OS require shared objects to be explicitly
201 passed to ld, causing the link step to fail if they are not.
202
203 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
204 mode, patterns that started with \h* or \R* might be incorrectly matched.
205
206
207 Version 8.30 04-February-2012
208 -----------------------------
209
210 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
211 name is defined in ctype.h.
212
213 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
214 only in quite long subpatterns.
215
216 3. Removed the function pcre_info(), which has been obsolete and deprecated
217 since it was replaced by pcre_fullinfo() in February 2000.
218
219 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
220 match a (*MARK), and the match failed at the start of the subject, a
221 reference to memory before the start of the subject could occur. This bug
222 was introduced by fix 17 of release 8.21.
223
224 5. A reference to an unset group with zero minimum repetition was giving
225 totally wrong answers (in non-JavaScript-compatibility mode). For example,
226 /(another)?(\1?)test/ matched against "hello world test". This bug was
227 introduced in release 8.13.
228
229 6. Add support for 16-bit character strings (a large amount of work involving
230 many changes and refactorings).
231
232 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
233 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
234 from a file.
235
236 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
237 rounding is not applied in this particular case).
238
239 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
240 if they appear, or are escaped, in patterns.
241
242 10. Get rid of a number of -Wunused-but-set-variable warnings.
243
244 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
245 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
246 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
247 also returns the mark "x". This bug applied to capturing parentheses,
248 non-capturing parentheses, and atomic parentheses. It also applied to some
249 assertions.
250
251 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
252 information out of configure.ac instead of relying on pcre.h.generic, which
253 is not stored in the repository.
254
255 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
256 -lreadline.
257
258 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
259
260 15. Applied Graycode's patch to put the top-level frame on the stack rather
261 than the heap when not using the stack for recursion. This gives a
262 performance improvement in many cases when recursion is not deep.
263
264 16. Experimental code added to "pcretest -C" to output the stack frame size.
265
266
267 Version 8.21 12-Dec-2011
268 ------------------------
269
270 1. Updating the JIT compiler.
271
272 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
273 are added as well.
274
275 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
276 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
277 calling _pcre_jit_exec. Some extra comments are added.
278
279 4. (*MARK) settings inside atomic groups that do not contain any capturing
280 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
281 was introduced by change 18 for 8.20.
282
283 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
284 ECMA-262 standard.
285
286 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
287 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
288 This bug was probably introduced by change 9 of 8.13.
289
290 7. While fixing 6 above, I noticed that a number of other items were being
291 incorrectly rejected as "not fixed length". This arose partly because newer
292 opcodes had not been added to the fixed-length checking code. I have (a)
293 corrected the bug and added tests for these items, and (b) arranged for an
294 error to occur if an unknown opcode is encountered while checking for fixed
295 length instead of just assuming "not fixed length". The items that were
296 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
297 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
298 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
299
300 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
301 being incorrectly compiled and would have given unpredicatble results.
302
303 9. A possessively repeated subpattern with minimum repeat count greater than
304 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
305 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
306 the first (A) could occur when it should not.
307
308 10. Add a cast and remove a redundant test from the code.
309
310 11. JIT should use pcre_malloc/pcre_free for allocation.
311
312 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
313 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
314 is anything other than /usr, -L is still shown).
315
316 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
317
318 14. Perl does not support \N without a following name in a [] class; PCRE now
319 also gives an error.
320
321 15. If a forward reference was repeated with an upper limit of around 2000,
322 it caused the error "internal error: overran compiling workspace". The
323 maximum number of forward references (including repeats) was limited by the
324 internal workspace, and dependent on the LINK_SIZE. The code has been
325 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
326 the default depends on LINK_SIZE. There is a new upper limit (for safety)
327 of around 200,000 forward references. While doing this, I also speeded up
328 the filling in of repeated forward references.
329
330 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
331 incorrectly expecting the subject to contain another "a" after the start.
332
333 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
334 in the match, the SKIP should be ignored. This was not happening; instead
335 the SKIP was being treated as NOMATCH. For patterns such as
336 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
337 tested.
338
339 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
340 now much more compatible with Perl, in particular in cases where the result
341 is a non-match for a non-anchored pattern. For example, if
342 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
343 "m", where previously it did not return a name. A side effect of this
344 change is that for partial matches, the last encountered mark name is
345 returned, as for non matches. A number of tests that were previously not
346 Perl-compatible have been moved into the Perl-compatible test files. The
347 refactoring has had the pleasing side effect of removing one argument from
348 the match() function, thus reducing its stack requirements.
349
350 19. If the /S+ option was used in pcretest to study a pattern using JIT,
351 subsequent uses of /S (without +) incorrectly behaved like /S+.
352
353 21. Retrieve executable code size support for the JIT compiler and fixing
354 some warnings.
355
356 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
357 not work when the shorter character appeared right at the end of the
358 subject string.
359
360 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
361 systems.
362
363 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
364 output it when the /M option is used in pcretest.
365
366 25. The CheckMan script was not being included in the distribution. Also, added
367 an explicit "perl" to run Perl scripts from the PrepareRelease script
368 because this is reportedly needed in Windows.
369
370 26. If study data was being save in a file and studying had not found a set of
371 "starts with" bytes for the pattern, the data written to the file (though
372 never used) was taken from uninitialized memory and so caused valgrind to
373 complain.
374
375 27. Updated RunTest.bat as provided by Sheri Pierce.
376
377 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
378
379 29. Computation of memory usage for the table of capturing group names was
380 giving an unnecessarily large value.
381
382
383 Version 8.20 21-Oct-2011
384 ------------------------
385
386 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
387 a POSIX class. After further experiments with Perl, which convinced me that
388 Perl has bugs and confusions, a closing square bracket is no longer allowed
389 in a POSIX name. This bug also affected patterns with classes that started
390 with full stops.
391
392 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
393 captured substring, but while checking the failing first alternative,
394 substring 1 is temporarily captured. If the output vector supplied to
395 pcre_exec() was not big enough for this capture, the yield of the function
396 was still zero ("insufficient space for captured substrings"). This cannot
397 be totally fixed without adding another stack variable, which seems a lot
398 of expense for a edge case. However, I have improved the situation in cases
399 such as /(a)(b)x|abc/ matched against "abc", where the return code
400 indicates that fewer than the maximum number of slots in the ovector have
401 been set.
402
403 3. Related to (2) above: when there are more back references in a pattern than
404 slots in the output vector, pcre_exec() uses temporary memory during
405 matching, and copies in the captures as far as possible afterwards. It was
406 using the entire output vector, but this conflicts with the specification
407 that only 2/3 is used for passing back captured substrings. Now it uses
408 only the first 2/3, for compatibility. This is, of course, another edge
409 case.
410
411 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
412 main code base, and can be used by building with --enable-jit. When this is
413 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
414 runtime --no-jit option is given.
415
416 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
417 ovector, the return from the function was zero, implying that there were
418 other matches that did not fit. The correct "exactly full" value is now
419 returned.
420
421 6. If a subpattern that was called recursively or as a subroutine contained
422 (*PRUNE) or any other control that caused it to give a non-standard return,
423 invalid errors such as "Error -26 (nested recursion at the same subject
424 position)" or even infinite loops could occur.
425
426 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
427 computing the minimum length on reaching *ACCEPT, and so ended up with the
428 wrong value of 1 rather than 0. Further investigation indicates that
429 computing a minimum subject length in the presence of *ACCEPT is difficult
430 (think back references, subroutine calls), and so I have changed the code
431 so that no minimum is registered for a pattern that contains *ACCEPT.
432
433 8. If (*THEN) was present in the first (true) branch of a conditional group,
434 it was not handled as intended. [But see 16 below.]
435
436 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
437 Sheri Pierce.
438
439 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
440 the first byte in a match must be "a".
441
442 11. Change 17 for 8.13 increased the recursion depth for patterns like
443 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
444 pattern contains any instances of (*THEN). If it does not, the old
445 optimizations are restored. It would be nice to do this on a per-group
446 basis, but at the moment that is not feasible.
447
448 12. In some environments, the output of pcretest -C is CRLF terminated. This
449 broke RunTest's code that checks for the link size. A single white space
450 character after the value is now allowed for.
451
452 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
453 For "fr", it uses the Windows-specific input and output files.
454
455 14. If (*THEN) appeared in a group that was called recursively or as a
456 subroutine, it did not work as intended. [But see next item.]
457
458 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
459 pattern fragments (but not containing any | characters). If A and B are
460 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
461 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
462 D. In other words, Perl considers parentheses that do not contain any |
463 characters to be part of a surrounding alternative, whereas PCRE was
464 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
465 differently. PCRE now behaves in the same way as Perl, except in the case
466 of subroutine/recursion calls such as (?1) which have in any case always
467 been different (but PCRE had them first :-).
468
469 16. Related to 15 above: Perl does not treat the | in a conditional group as
470 creating alternatives. Such a group is treated in the same way as an
471 ordinary group without any | characters when processing (*THEN). PCRE has
472 been changed to match Perl's behaviour.
473
474 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
475 RunGrepTest script failed.
476
477 18. Change 22 for version 13 caused atomic groups to use more stack. This is
478 inevitable for groups that contain captures, but it can lead to a lot of
479 stack use in large patterns. The old behaviour has been restored for atomic
480 groups that do not contain any capturing parentheses.
481
482 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
483 suppress the check for a minimum subject length at run time. (If it was
484 given to pcre_exec() or pcre_dfa_exec() it did work.)
485
486 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
487 fail to work when decoding hex characters in data strings in EBCDIC
488 environments.
489
490 21. It appears that in at least one Mac OS environment, the isxdigit() function
491 is implemented as a macro that evaluates to its argument more than once,
492 contravening the C 90 Standard (I haven't checked a later standard). There
493 was an instance in pcretest which caused it to go wrong when processing
494 \x{...} escapes in subject strings. The has been rewritten to avoid using
495 things like p++ in the argument of isxdigit().
496
497
498 Version 8.13 16-Aug-2011
499 ------------------------
500
501 1. The Unicode data tables have been updated to Unicode 6.0.0.
502
503 2. Two minor typos in pcre_internal.h have been fixed.
504
505 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
506 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
507 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
508
509 4. There were a number of related bugs in the code for matching backrefences
510 caselessly in UTF-8 mode when codes for the characters concerned were
511 different numbers of bytes. For example, U+023A and U+2C65 are an upper
512 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
513 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
514 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
515 2-byte code at the end of the subject (it thought there wasn't enough data
516 left).
517
518 5. Comprehensive information about what went wrong is now returned by
519 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
520 as the output vector has at least 2 elements. The offset of the start of
521 the failing character and a reason code are placed in the vector.
522
523 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
524 now returned is for the first byte of the failing character, instead of the
525 last byte inspected. This is an incompatible change, but I hope it is small
526 enough not to be a problem. It makes the returned offset consistent with
527 pcre_exec() and pcre_dfa_exec().
528
529 7. pcretest now gives a text phrase as well as the error number when
530 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
531 failure, the offset and reason code are output.
532
533 8. When \R was used with a maximizing quantifier it failed to skip backwards
534 over a \r\n pair if the subsequent match failed. Instead, it just skipped
535 back over a single character (\n). This seems wrong (because it treated the
536 two characters as a single entity when going forwards), conflicts with the
537 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
538 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
539 has been changed.
540
541 9. Some internal refactoring has changed the processing so that the handling
542 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
543 time (the PCRE_DOTALL option was changed this way some time ago: version
544 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
545 which was always a bit of a fudge. It also means that there is one less
546 argument for the match() function, which reduces its stack requirements
547 slightly. This change also fixes an incompatibility with Perl: the pattern
548 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
549
550 10. More internal refactoring has drastically reduced the number of recursive
551 calls to match() for possessively repeated groups such as (abc)++ when
552 using pcre_exec().
553
554 11. While implementing 10, a number of bugs in the handling of groups were
555 discovered and fixed:
556
557 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
558 (a|)*(?1) gave a compile-time internal error.
559 ((a|)+)+ did not notice that the outer group could match an empty string.
560 (^a|^)+ was not marked as anchored.
561 (.*a|.*)+ was not marked as matching at start or after a newline.
562
563 12. Yet more internal refactoring has removed another argument from the match()
564 function. Special calls to this function are now indicated by setting a
565 value in a variable in the "match data" data block.
566
567 13. Be more explicit in pcre_study() instead of relying on "default" for
568 opcodes that mean there is no starting character; this means that when new
569 ones are added and accidentally left out of pcre_study(), testing should
570 pick them up.
571
572 14. The -s option of pcretest has been documented for ages as being an old
573 synonym of -m (show memory usage). I have changed it to mean "force study
574 for every regex", that is, assume /S for every regex. This is similar to -i
575 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
576 using it. It makes it easier to run collections of tests with and without
577 study enabled, and thereby test pcre_study() more easily. All the standard
578 tests are now run with and without -s (but some patterns can be marked as
579 "never study" - see 20 below).
580
581 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
582 restoration of the capturing data to the outer values was not happening
583 correctly.
584
585 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
586 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
587 pattern had matched an empty string, and so incorrectly returned a no
588 match.
589
590 17. There was optimizing code for the last branch of non-capturing parentheses,
591 and also for the obeyed branch of a conditional subexpression, which used
592 tail recursion to cut down on stack usage. Unfortunately, now that there is
593 the possibility of (*THEN) occurring in these branches, tail recursion is
594 no longer possible because the return has to be checked for (*THEN). These
595 two optimizations have therefore been removed. [But see 8.20/11 above.]
596
597 18. If a pattern containing \R was studied, it was assumed that \R always
598 matched two bytes, thus causing the minimum subject length to be
599 incorrectly computed because \R can also match just one byte.
600
601 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
602 was incorrectly computed.
603
604 20. If /S is present twice on a test pattern in pcretest input, it now
605 *disables* studying, thereby overriding the use of -s on the command line
606 (see 14 above). This is necessary for one or two tests to keep the output
607 identical in both cases.
608
609 21. When (*ACCEPT) was used in an assertion that matched an empty string and
610 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
611
612 22. When an atomic group that contained a capturing parenthesis was
613 successfully matched, but the branch in which it appeared failed, the
614 capturing was not being forgotten if a higher numbered group was later
615 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
616 group 1 to "a", when in fact it should be unset. This applied to multi-
617 branched capturing and non-capturing groups, repeated or not, and also to
618 positive assertions (capturing in negative assertions does not happen
619 in PCRE) and also to nested atomic groups.
620
621 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
622 subject after a captured substring, to make it easier to tell which of a
623 number of identical substrings has been captured.
624
625 24. The way atomic groups are processed by pcre_exec() has been changed so that
626 if they are repeated, backtracking one repetition now resets captured
627 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
628 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
629 "aaa". Previously, it would have been "a". As part of this code
630 refactoring, the way recursive calls are handled has also been changed.
631
632 25. If an assertion condition captured any substrings, they were not passed
633 back unless some other capturing happened later. For example, if
634 (?(?=(a))a) was matched against "a", no capturing was returned.
635
636 26. When studying a pattern that contained subroutine calls or assertions,
637 the code for finding the minimum length of a possible match was handling
638 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
639 group 1 called group 2 while simultaneously a separate group 2 called group
640 1). A stack overflow occurred in this case. I have fixed this by limiting
641 the recursion depth to 10.
642
643 27. Updated RunTest.bat in the distribution to the version supplied by Tom
644 Fortmann. This supports explicit test numbers on the command line, and has
645 argument validation and error reporting.
646
647 28. An instance of \X with an unlimited repeat could fail if at any point the
648 first character it looked at was a mark character.
649
650 29. Some minor code refactoring concerning Unicode properties and scripts
651 should reduce the stack requirement of match() slightly.
652
653 30. Added the '=' option to pcretest to check the setting of unused capturing
654 slots at the end of the pattern, which are documented as being -1, but are
655 not included in the return count.
656
657 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
658 compiled something random. Now it gives a compile-time error (as does
659 Perl).
660
661 32. A *MARK encountered during the processing of a positive assertion is now
662 recorded and passed back (compatible with Perl).
663
664 33. If --only-matching or --colour was set on a pcregrep call whose pattern
665 had alternative anchored branches, the search for a second match in a line
666 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
667 matched the line "0102" twice. The same bug affected patterns that started
668 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
669 twice.
670
671 34. Previously, PCRE did not allow quantification of assertions. However, Perl
672 does, and because of capturing effects, quantifying parenthesized
673 assertions may at times be useful. Quantifiers are now allowed for
674 parenthesized assertions.
675
676 35. A minor code tidy in pcre_compile() when checking options for \R usage.
677
678 36. \g was being checked for fancy things in a character class, when it should
679 just be a literal "g".
680
681 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
682 appearance of a nested POSIX class supersedes an apparent external class.
683 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
684 unescaped square brackets may also appear as part of class names. For
685 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
686 more like Perl. (But see 8.20/1 above.)
687
688 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
689 was because it thought it was \N{name}, which is not supported).
690
691 39. Add minix to OS list not supporting the -S option in pcretest.
692
693 40. PCRE tries to detect cases of infinite recursion at compile time, but it
694 cannot analyze patterns in sufficient detail to catch mutual recursions
695 such as ((?1))((?2)). There is now a runtime test that gives an error if a
696 subgroup is called recursively as a subpattern for a second time at the
697 same position in the subject string. In previous releases this might have
698 been caught by the recursion limit, or it might have run out of stack.
699
700 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
701 happen only once. PCRE was, however incorrectly giving a compile time error
702 "recursive call could loop indefinitely" because it cannot analyze the
703 pattern in sufficient detail. The compile time test no longer happens when
704 PCRE is compiling a conditional subpattern, but actual runaway loops are
705 now caught at runtime (see 40 above).
706
707 42. It seems that Perl allows any characters other than a closing parenthesis
708 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
709 has been changed to be the same.
710
711 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
712 as not to get warnings when autogen.sh is called. Also changed
713 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
714
715 44. To help people who use pcregrep to scan files containing exceedingly long
716 lines, the following changes have been made:
717
718 (a) The default value of the buffer size parameter has been increased from
719 8K to 20K. (The actual buffer used is three times this size.)
720
721 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
722 PCRE is built.
723
724 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
725 to be set at run time.
726
727 (d) Numerical values in pcregrep options can be followed by K or M, for
728 example --buffer-size=50K.
729
730 (e) If a line being scanned overflows pcregrep's buffer, an error is now
731 given and the return code is set to 2.
732
733 45. Add a pointer to the latest mark to the callout data block.
734
735 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
736 partial match of an empty string instead of no match. This was specific to
737 the use of ".".
738
739 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
740 complete match instead of a partial match. This bug was dependent on both
741 the PCRE_UTF8 and PCRE_DOTALL options being set.
742
743 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
744 starting byte set, because \b was not being ignored.
745
746
747 Version 8.12 15-Jan-2011
748 ------------------------
749
750 1. Fixed some typos in the markup of the man pages, and wrote a script that
751 checks for such things as part of the documentation building process.
752
753 2. On a big-endian 64-bit system, pcregrep did not correctly process the
754 --match-limit and --recursion-limit options (added for 8.11). In
755 particular, this made one of the standard tests fail. (The integer value
756 went into the wrong half of a long int.)
757
758 3. If the --colour option was given to pcregrep with -v (invert match), it
759 did strange things, either producing crazy output, or crashing. It should,
760 of course, ignore a request for colour when reporting lines that do not
761 match.
762
763 4. Another pcregrep bug caused similar problems if --colour was specified with
764 -M (multiline) and the pattern match finished with a line ending.
765
766 5. In pcregrep, when a pattern that ended with a literal newline sequence was
767 matched in multiline mode, the following line was shown as part of the
768 match. This seems wrong, so I have changed it.
769
770 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
771 the check for further matches in the same line (so they could be coloured)
772 to overrun the end of the current line. If another match was found, it was
773 incorrectly shown (and then shown again when found in the next line).
774
775 7. If pcregrep was compiled under Windows, there was a reference to the
776 function pcregrep_exit() before it was defined. I am assuming this was
777 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
778 reported by a user. I've moved the definition above the reference.
779
780
781 Version 8.11 10-Dec-2010
782 ------------------------
783
784 1. (*THEN) was not working properly if there were untried alternatives prior
785 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
786 backtracked to try for "b" instead of moving to the next alternative branch
787 at the same level (in this case, to look for "c"). The Perl documentation
788 is clear that when (*THEN) is backtracked onto, it goes to the "next
789 alternative in the innermost enclosing group".
790
791 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
792 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
793 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
794 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
795 (*THEN).
796
797 3. If \s appeared in a character class, it removed the VT character from
798 the class, even if it had been included by some previous item, for example
799 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
800 of \s, but is part of the POSIX "space" class.)
801
802 4. A partial match never returns an empty string (because you can always
803 match an empty string at the end of the subject); however the checking for
804 an empty string was starting at the "start of match" point. This has been
805 changed to the "earliest inspected character" point, because the returned
806 data for a partial match starts at this character. This means that, for
807 example, /(?<=abc)def/ gives a partial match for the subject "abc"
808 (previously it gave "no match").
809
810 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
811 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
812 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
813 has an implication that the given string is incomplete (because a partial
814 match is preferred over a full match). For this reason, these items now
815 give a partial match in this situation. [Aside: previously, the one case
816 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
817 match rather than a full match, which was wrong by the old rules, but is
818 now correct.]
819
820 6. There was a bug in the handling of #-introduced comments, recognized when
821 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
822 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
823 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
824 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
825 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
826 places in pcre_compile().
827
828 7. Related to (6) above, when pcre_compile() was skipping #-introduced
829 comments when looking ahead for named forward references to subpatterns,
830 the only newline sequence it recognized was NL. It now handles newlines
831 according to the set newline convention.
832
833 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
834 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
835 cater for a lack of strerror(). These oversights have been fixed.
836
837 9. Added --match-limit and --recursion-limit to pcregrep.
838
839 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
840
841 11. When the -o option was used, pcregrep was setting a return code of 1, even
842 when matches were found, and --line-buffered was not being honoured.
843
844 12. Added an optional parentheses number to the -o and --only-matching options
845 of pcregrep.
846
847 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
848 can match an empty string. The code to do it in pcretest and pcredemo
849 needed fixing:
850
851 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
852 only one byte after an empty string match just before CRLF (this case
853 just got forgotten; "any" and "anycrlf" were OK).
854
855 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
856 mode when an empty string match preceded an ASCII character followed by
857 a non-ASCII character. (The code for advancing by one character rather
858 than one byte was nonsense.)
859
860 (c) The pcredemo.c sample program did not have any code at all to handle
861 the cases when CRLF is a valid newline sequence.
862
863 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
864 as a starting offset was within the subject string. There is now a new
865 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
866 negative or greater than the length of the string. In order to test this,
867 pcretest is extended to allow the setting of negative starting offsets.
868
869 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
870 starting offset points to the beginning of a UTF-8 character was
871 unnecessarily clumsy. I tidied it up.
872
873 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
874 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
875
876 17. Nobody had reported that the --include_dir option, which was added in
877 release 7.7 should have been called --include-dir (hyphen, not underscore)
878 for compatibility with GNU grep. I have changed it to --include-dir, but
879 left --include_dir as an undocumented synonym, and the same for
880 --exclude-dir, though that is not available in GNU grep, at least as of
881 release 2.5.4.
882
883 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
884 characters from a string of bytes) have been redefined so as not to use
885 loops, in order to improve performance in some environments. At the same
886 time, I abstracted some of the common code into auxiliary macros to save
887 repetition (this should not affect the compiled code).
888
889 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
890 compile-time error is now given if \c is not followed by an ASCII
891 character, that is, a byte less than 128. (In EBCDIC mode, the code is
892 different, and any byte value is allowed.)
893
894 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
895 START_OPTIMIZE option, which is now allowed at compile time - but just
896 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
897 to pcregrep and other applications that have no direct access to PCRE
898 options. The new /Y option in pcretest sets this option when calling
899 pcre_compile().
900
901 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
902 back references. Groups containing recursive back references were forced to
903 be atomic by that change, but in the case of named groups, the amount of
904 memory required was incorrectly computed, leading to "Failed: internal
905 error: code overflow". This has been fixed.
906
907 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
908 pcretest.c, to avoid build problems in some Borland environments.
909
910
911 Version 8.10 25-Jun-2010
912 ------------------------
913
914 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
915 THEN.
916
917 2. (*ACCEPT) was not working when inside an atomic group.
918
919 3. Inside a character class, \B is treated as a literal by default, but
920 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
921 causes the error). The code is unchanged, but I tidied the documentation.
922
923 4. Inside a character class, PCRE always treated \R and \X as literals,
924 whereas Perl faults them if its -w option is set. I have changed PCRE so
925 that it faults them when PCRE_EXTRA is set.
926
927 5. Added support for \N, which always matches any character other than
928 newline. (It is the same as "." when PCRE_DOTALL is not set.)
929
930 6. When compiling pcregrep with newer versions of gcc which may have
931 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
932 declared with attribute warn_unused_result" were given. Just casting the
933 result to (void) does not stop the warnings; a more elaborate fudge is
934 needed. I've used a macro to implement this.
935
936 7. Minor change to pcretest.c to avoid a compiler warning.
937
938 8. Added four artifical Unicode properties to help with an option to make
939 \s etc use properties (see next item). The new properties are: Xan
940 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
941
942 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
943 use Unicode properties. (*UCP) at the start of a pattern can be used to set
944 this option. Modified pcretest to add /W to test this facility. Added
945 REG_UCP to make it available via the POSIX interface.
946
947 10. Added --line-buffered to pcregrep.
948
949 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
950 studied, and the match started with a letter with a code point greater than
951 127 whose first byte was different to the first byte of the other case of
952 the letter, the other case of this starting letter was not recognized
953 (#976).
954
955 12. If a pattern that was studied started with a repeated Unicode property
956 test, for example, \p{Nd}+, there was the theoretical possibility of
957 setting up an incorrect bitmap of starting bytes, but fortunately it could
958 not have actually happened in practice until change 8 above was made (it
959 added property types that matched character-matching opcodes).
960
961 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
962 possible starting bytes for non-anchored patterns.
963
964 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
965 \R, and also a number of cases that involve Unicode properties, both
966 explicit and implicit when PCRE_UCP is set.
967
968 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
969 input, it could crash or give wrong results if characters with values
970 greater than 0xc0 were present in the subject string. (Detail: it assumed
971 UTF-8 input when processing these items.)
972
973 16. Added a lot of (int) casts to avoid compiler warnings in systems where
974 size_t is 64-bit (#991).
975
976 17. Added a check for running out of memory when PCRE is compiled with
977 --disable-stack-for-recursion (#990).
978
979 18. If the last data line in a file for pcretest does not have a newline on
980 the end, a newline was missing in the output.
981
982 19. The default pcre_chartables.c file recognizes only ASCII characters (values
983 less than 128) in its various bitmaps. However, there is a facility for
984 generating tables according to the current locale when PCRE is compiled. It
985 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
986 characters, are recognized by isspace() and therefore were getting set in
987 these tables, and indeed these tables seem to approximate to ISO 8859. This
988 caused a problem in UTF-8 mode when pcre_study() was used to create a list
989 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
990 which of course cannot start UTF-8 characters. I have changed the code so
991 that only real ASCII characters (less than 128) and the correct starting
992 bytes for UTF-8 encodings are set for characters greater than 127 when in
993 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
994 altogether.)
995
996 20. Added the /T option to pcretest so as to be able to run tests with non-
997 standard character tables, thus making it possible to include the tests
998 used for 19 above in the standard set of tests.
999
1000 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1001 reference to a subpattern the other side of a comment that contains an
1002 opening parenthesis caused either an internal compiling error, or a
1003 reference to the wrong subpattern.
1004
1005
1006 Version 8.02 19-Mar-2010
1007 ------------------------
1008
1009 1. The Unicode data tables have been updated to Unicode 5.2.0.
1010
1011 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1012 configured.
1013
1014 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1015 original author of that file, following a query about its status.
1016
1017 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1018 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1019
1020 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1021 quantifier applied to a forward-referencing subroutine call, could compile
1022 incorrect code or give the error "internal error: previously-checked
1023 referenced subpattern not found".
1024
1025 6. Both MS Visual Studio and Symbian OS have problems with initializing
1026 variables to point to external functions. For these systems, therefore,
1027 pcre_malloc etc. are now initialized to local functions that call the
1028 relevant global functions.
1029
1030 7. There were two entries missing in the vectors called coptable and poptable
1031 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1032 I've fixed the data, and added a kludgy way of testing at compile time that
1033 the lengths are correct (equal to the number of opcodes).
1034
1035 8. Following on from 7, I added a similar kludge to check the length of the
1036 eint vector in pcreposix.c.
1037
1038 9. Error texts for pcre_compile() are held as one long string to avoid too
1039 much relocation at load time. To find a text, the string is searched,
1040 counting zeros. There was no check for running off the end of the string,
1041 which could happen if a new error number was added without updating the
1042 string.
1043
1044 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1045
1046 11. \K was not working if it appeared in an atomic group or in a group that
1047 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1048 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1049 the assertion is positive, but not if it is negative.
1050
1051 12. Change 11 fortuitously reduced the size of the stack frame used in the
1052 "match()" function of pcre_exec.c by one pointer. Forthcoming
1053 implementation of support for (*MARK) will need an extra pointer on the
1054 stack; I have reserved it now, so that the stack frame size does not
1055 decrease.
1056
1057 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1058 item in branch that calls a recursion is a subroutine call - as in the
1059 second branch in the above example - was incorrectly given the compile-
1060 time error "recursive call could loop indefinitely" because pcre_compile()
1061 was not correctly checking the subroutine for matching a non-empty string.
1062
1063 14. The checks for overrunning compiling workspace could trigger after an
1064 overrun had occurred. This is a "should never occur" error, but it can be
1065 triggered by pathological patterns such as hundreds of nested parentheses.
1066 The checks now trigger 100 bytes before the end of the workspace.
1067
1068 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1069
1070
1071 Version 8.01 19-Jan-2010
1072 ------------------------
1073
1074 1. If a pattern contained a conditional subpattern with only one branch (in
1075 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1076 computed the wrong minimum data length (which is of course zero for such
1077 subpatterns). This could cause incorrect "no match" results.
1078
1079 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1080 the pattern is reset in the first branch, pcre_compile() failed with
1081 "internal error: code overflow at offset...". This happened only when
1082 the reset was to the original external option setting. (An optimization
1083 abstracts leading options settings into an external setting, which was the
1084 cause of this.)
1085
1086 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1087 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1088 assertion pattern did not match (meaning that the assertion was true), it
1089 was incorrectly treated as false if the SKIP had been reached during the
1090 matching. This also applied to assertions used as conditions.
1091
1092 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1093 assertion subpattern, including such a pattern used as a condition,
1094 unpredictable results occurred, instead of the error return
1095 PCRE_ERROR_DFA_UITEM.
1096
1097 5. The C++ GlobalReplace function was not working like Perl for the special
1098 situation when an empty string is matched. It now does the fancy magic
1099 stuff that is necessary.
1100
1101 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1102 removed. (These were left over from very, very early versions of PCRE.)
1103
1104 7. Some cosmetic changes to the code to make life easier when compiling it
1105 as part of something else:
1106
1107 (a) Change DEBUG to PCRE_DEBUG.
1108
1109 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1110 called "current" as "current_branch", to prevent a collision with the
1111 Linux macro when compiled as a kernel module.
1112
1113 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1114 prevent a collision with the Linux macro when compiled as a kernel
1115 module.
1116
1117 8. In pcre_compile() there are some checks for integer overflows that used to
1118 cast potentially large values to (double). This has been changed to that
1119 when building, a check for int64_t is made, and if it is found, it is used
1120 instead, thus avoiding the use of floating point arithmetic. (There is no
1121 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1122 double.
1123
1124 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1125 2005 (difference between two addresses compared to an unsigned value).
1126
1127 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1128 custom one, because of the following reported problem in Windows:
1129
1130 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1131 under Win32.
1132 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1133 therefore missing the function definition.
1134 - The compiler thus generates a "C" signature for the test function.
1135 - The linker fails to find the "C" function.
1136 - PCRE fails to configure if asked to do so against libbz2.
1137
1138 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1139 messages were output:
1140
1141 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1142 rerunning libtoolize, to keep the correct libtool macros in-tree.
1143 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1144
1145 I have done both of these things.
1146
1147 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1148 most of the time, it *can* run out if it is given a pattern that contains a
1149 runaway infinite recursion. I updated the discussion in the pcrestack man
1150 page.
1151
1152 13. Now that we have gone to the x.xx style of version numbers, the minor
1153 version may start with zero. Using 08 or 09 is a bad idea because users
1154 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1155 interpreted as invalid octal numbers. I've updated the previous comment in
1156 configure.ac, and also added a check that gives an error if 08 or 09 are
1157 used.
1158
1159 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1160 causing partial matching to fail when the end of the subject matched \W
1161 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1162
1163 15. There were some discrepancies between the declarations in pcre_internal.h
1164 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1165 their definitions. The declarations used "const uschar *" and the
1166 definitions used USPTR. Even though USPTR is normally defined as "const
1167 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1168 reported that: "This difference in casting confuses some C++ compilers, for
1169 example, SunCC recognizes above declarations as different functions and
1170 generates broken code for hbpcre." I have changed the declarations to use
1171 USPTR.
1172
1173 16. GNU libtool is named differently on some systems. The autogen.sh script now
1174 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1175 (FreeBSD).
1176
1177 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1178 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1179 comment: "Figure out how to create a longlong from a string: strtoll and
1180 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1181 instance, but it only takes 2 args instead of 3!"
1182
1183 18. A subtle bug concerned with back references has been fixed by a change of
1184 specification, with a corresponding code fix. A pattern such as
1185 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1186 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1187 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1188 same bug. Such groups have to be quantified to be useful, or contained
1189 inside another quantified group. (If there's no repetition, the reference
1190 can never match.) The problem arises because, having left the group and
1191 moved on to the rest of the pattern, a later failure that backtracks into
1192 the group uses the captured value from the final iteration of the group
1193 rather than the correct earlier one. I have fixed this in PCRE by forcing
1194 any group that contains a reference to itself to be an atomic group; that
1195 is, there cannot be any backtracking into it once it has completed. This is
1196 similar to recursive and subroutine calls.
1197
1198
1199 Version 8.00 19-Oct-09
1200 ----------------------
1201
1202 1. The table for translating pcre_compile() error codes into POSIX error codes
1203 was out-of-date, and there was no check on the pcre_compile() error code
1204 being within the table. This could lead to an OK return being given in
1205 error.
1206
1207 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1208 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1209 in a Windows environment.
1210
1211 3. The pcregrep --count option prints the count for each file even when it is
1212 zero, as does GNU grep. However, pcregrep was also printing all files when
1213 --files-with-matches was added. Now, when both options are given, it prints
1214 counts only for those files that have at least one match. (GNU grep just
1215 prints the file name in this circumstance, but including the count seems
1216 more useful - otherwise, why use --count?) Also ensured that the
1217 combination -clh just lists non-zero counts, with no names.
1218
1219 4. The long form of the pcregrep -F option was incorrectly implemented as
1220 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1221 but it seems right to fix it, and I didn't think it was worth preserving
1222 the old behaviour.
1223
1224 5. The command line items --regex=pattern and --regexp=pattern were not
1225 recognized by pcregrep, which required --regex pattern or --regexp pattern
1226 (with a space rather than an '='). The man page documented the '=' forms,
1227 which are compatible with GNU grep; these now work.
1228
1229 6. No libpcreposix.pc file was created for pkg-config; there was just
1230 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1231
1232 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1233 when UCP support is not needed, by modifying the Python script that
1234 generates it from Unicode data files. This should not matter if the module
1235 is correctly used as a library, but I received one complaint about 50K of
1236 unwanted data. My guess is that the person linked everything into his
1237 program rather than using a library. Anyway, it does no harm.
1238
1239 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1240 was a minimum greater than 1 for a wide character in a possessive
1241 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1242 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1243 character. Chaos in the form of incorrect output or a compiling loop could
1244 result.
1245
1246 9. The restrictions on what a pattern can contain when partial matching is
1247 requested for pcre_exec() have been removed. All patterns can now be
1248 partially matched by this function. In addition, if there are at least two
1249 slots in the offset vector, the offset of the earliest inspected character
1250 for the match and the offset of the end of the subject are set in them when
1251 PCRE_ERROR_PARTIAL is returned.
1252
1253 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1254 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1255 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1256 and may be more useful for multi-segment matching.
1257
1258 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1259 used to be given if ever the end of the subject was reached; now it is
1260 given only if matching could not proceed because another character was
1261 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1262 string "Z", which now yields "no match" instead of "partial match". In the
1263 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1264 final character ended with (*FAIL).
1265
1266 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1267 if the pattern had a "must contain" character that was already found in the
1268 earlier partial match, unless partial matching was again requested. For
1269 example, with the pattern /dog.(body)?/, the "must contain" character is
1270 "g". If the first part-match was for the string "dog", restarting with
1271 "sbody" failed. This bug has been fixed.
1272
1273 13. The string returned by pcre_dfa_exec() after a partial match has been
1274 changed so that it starts at the first inspected character rather than the
1275 first character of the match. This makes a difference only if the pattern
1276 starts with a lookbehind assertion or \b or \B (\K is not supported by
1277 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1278 matching functions compatible, and I think it's the right thing to do.
1279
1280 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1281 so that the demonstration program is easily available in environments where
1282 PCRE has not been installed from source.
1283
1284 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1285 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1286 library.
1287
1288 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1289 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1290 is not the first non-POSIX option to be added. Clearly some people find
1291 these options useful.
1292
1293 17. If a caller to the POSIX matching function regexec() passes a non-zero
1294 value for nmatch with a NULL value for pmatch, the value of
1295 nmatch is forced to zero.
1296
1297 18. RunGrepTest did not have a test for the availability of the -u option of
1298 the diff command, as RunTest does. It now checks in the same way as
1299 RunTest, and also checks for the -b option.
1300
1301 19. If an odd number of negated classes containing just a single character
1302 interposed, within parentheses, between a forward reference to a named
1303 subpattern and the definition of the subpattern, compilation crashed with
1304 an internal error, complaining that it could not find the referenced
1305 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1306 [The bug was that it was starting one character too far in when skipping
1307 over the character class, thus treating the ] as data rather than
1308 terminating the class. This meant it could skip too much.]
1309
1310 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1311 /g option in pcretest when the pattern contains \K, which makes it possible
1312 to have an empty string match not at the start, even when the pattern is
1313 anchored. Updated pcretest and pcredemo to use this option.
1314
1315 21. If the maximum number of capturing subpatterns in a recursion was greater
1316 than the maximum at the outer level, the higher number was returned, but
1317 with unset values at the outer level. The correct (outer level) value is
1318 now given.
1319
1320 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1321 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1322 make it do so. The string so far is captured, making this feature
1323 compatible with Perl.
1324
1325 23. The tests have been re-organized, adding tests 11 and 12, to make it
1326 possible to check the Perl 5.10 features against Perl 5.10.
1327
1328 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1329 pattern matches a fixed length string. PCRE did not allow this; now it
1330 does. Neither allows recursion.
1331
1332 25. I finally figured out how to implement a request to provide the minimum
1333 length of subject string that was needed in order to match a given pattern.
1334 (It was back references and recursion that I had previously got hung up
1335 on.) This code has now been added to pcre_study(); it finds a lower bound
1336 to the length of subject needed. It is not necessarily the greatest lower
1337 bound, but using it to avoid searching strings that are too short does give
1338 some useful speed-ups. The value is available to calling programs via
1339 pcre_fullinfo().
1340
1341 26. While implementing 25, I discovered to my embarrassment that pcretest had
1342 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1343 study optimizations had never been tested with that matching function.
1344 Oops. What is worse, even when it was passed study data, there was a bug in
1345 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1346 were also very few tests of studied patterns with pcre_dfa_exec().
1347
1348 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1349 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1350 on the other side of the coin, they are no longer allowed to have different
1351 names, because these cannot be distinguished in PCRE, and this has caused
1352 confusion. (This is a difference from Perl.)
1353
1354 28. When duplicate subpattern names are present (necessarily with different
1355 numbers, as required by 27 above), and a test is made by name in a
1356 conditional pattern, either for a subpattern having been matched, or for
1357 recursion in such a pattern, all the associated numbered subpatterns are
1358 tested, and the overall condition is true if the condition is true for any
1359 one of them. This is the way Perl works, and is also more like the way
1360 testing by number works.
1361
1362
1363 Version 7.9 11-Apr-09
1364 ---------------------
1365
1366 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1367 (pcretest), all targets were linked against these libraries. This included
1368 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1369 libraries. This caused unwanted dependencies to be created. This problem
1370 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1371 pcretest is linked with readline.
1372
1373 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1374 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1375 moved outside it again, because FALSE and TRUE are already defined in AIX,
1376 but BOOL is not.
1377
1378 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1379 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1380
1381 4. The pcregrep documentation said spaces were inserted as well as colons (or
1382 hyphens) following file names and line numbers when outputting matching
1383 lines. This is not true; no spaces are inserted. I have also clarified the
1384 wording for the --colour (or --color) option.
1385
1386 5. In pcregrep, when --colour was used with -o, the list of matching strings
1387 was not coloured; this is different to GNU grep, so I have changed it to be
1388 the same.
1389
1390 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1391 each matching line was coloured. Now it goes on to look for further matches
1392 of any of the test patterns, which is the same behaviour as GNU grep.
1393
1394 7. A pattern that could match an empty string could cause pcregrep to loop; it
1395 doesn't make sense to accept an empty string match in pcregrep, so I have
1396 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1397 seems to be how GNU grep behaves.
1398
1399 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1400 start or after a newline", because the conditional assertion was not being
1401 correctly handled. The rule now is that both the assertion and what follows
1402 in the first alternative must satisfy the test.
1403
1404 9. If auto-callout was enabled in a pattern with a conditional group whose
1405 condition was an assertion, PCRE could crash during matching, both with
1406 pcre_exec() and pcre_dfa_exec().
1407
1408 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1409 used for matching.
1410
1411 11. Unicode property support in character classes was not working for
1412 characters (bytes) greater than 127 when not in UTF-8 mode.
1413
1414 12. Added the -M command line option to pcretest.
1415
1416 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1417
1418 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1419
1420 16. Added comments and documentation about mis-use of no_arg in the C++
1421 wrapper.
1422
1423 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1424 from Martin Jerabek that uses macro names for all relevant character and
1425 string constants.
1426
1427 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1428 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1429 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1430 these, but not everybody uses configure.
1431
1432 19. A conditional group that had only one branch was not being correctly
1433 recognized as an item that could match an empty string. This meant that an
1434 enclosing group might also not be so recognized, causing infinite looping
1435 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1436 with the subject "ab", where knowledge that the repeated group can match
1437 nothing is needed in order to break the loop.
1438
1439 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1440 exec(), but without supplying a callout function, matching went wrong.
1441
1442 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1443 leak if the size of the offset vector was greater than 30. When the vector
1444 is smaller, the saved offsets during recursion go onto a local stack
1445 vector, but for larger vectors malloc() is used. It was failing to free
1446 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1447 error, in fact).
1448
1449 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1450 heapframe that is used only when UTF-8 support is enabled. This caused no
1451 problem, but was untidy.
1452
1453 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1454 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1455 included within another project.
1456
1457 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1458 slightly modified by me:
1459
1460 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1461 not building pcregrep.
1462
1463 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1464 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1465
1466 25. Forward references, both numeric and by name, in patterns that made use of
1467 duplicate group numbers, could behave incorrectly or give incorrect errors,
1468 because when scanning forward to find the reference group, PCRE was not
1469 taking into account the duplicate group numbers. A pattern such as
1470 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1471
1472 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1473 the feature of a custom pointer more persuasive (as requested by a user).
1474
1475 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1476 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1477 pre-defined.
1478
1479 28. Added support for (*UTF8) at the start of a pattern.
1480
1481 29. Arrange for flags added by the "release type" setting in CMake to be shown
1482 in the configuration summary.
1483
1484
1485 Version 7.8 05-Sep-08
1486 ---------------------
1487
1488 1. Replaced UCP searching code with optimized version as implemented for Ad
1489 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1490 stage table and inline lookup instead of a function, giving speed ups of 2
1491 to 5 times on some simple patterns that I tested. Permission was given to
1492 distribute the MultiStage2.py script that generates the tables (it's not in
1493 the tarball, but is in the Subversion repository).
1494
1495 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1496 scripts.
1497
1498 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1499 a group with a zero qualifier. The result of the study could be incorrect,
1500 or the function might crash, depending on the pattern.
1501
1502 4. Caseless matching was not working for non-ASCII characters in back
1503 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1504 It now works when Unicode Property Support is available.
1505
1506 5. In pcretest, an escape such as \x{de} in the data was always generating
1507 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1508 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1509 truncation.
1510
1511 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1512
1513 7. Added two (int) casts to pcregrep when printing the difference of two
1514 pointers, in case they are 64-bit values.
1515
1516 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1517 test 2 if it fails.
1518
1519 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1520 and a #define of that name to empty if it is not externally set. This is to
1521 allow users of MSVC to set it if necessary.
1522
1523 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1524 the convenience functions in the pcre_get.c source file.
1525
1526 11. An option change at the start of a pattern that had top-level alternatives
1527 could cause overwriting and/or a crash. This command provoked a crash in
1528 some environments:
1529
1530 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1531
1532 This potential security problem was recorded as CVE-2008-2371.
1533
1534 12. For a pattern where the match had to start at the beginning or immediately
1535 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1536 pcre_dfa_exec() could read past the end of the passed subject if there was
1537 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1538 pcretest so that it places the subject at the end of its malloc-ed buffer.
1539
1540 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1541 exec() might read past the end of the data buffer in UTF-8 mode.
1542
1543 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1544 the data contained the byte 0x85 as part of a UTF-8 character within its
1545 first line. This applied both to normal and DFA matching.
1546
1547 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1548 /^[^d]*?$/8 failed to match "abc".
1549
1550 16. Added a missing copyright notice to pcrecpp_internal.h.
1551
1552 17. Make it more clear in the documentation that values returned from
1553 pcre_exec() in ovector are byte offsets, not character counts.
1554
1555 18. Tidied a few places to stop certain compilers from issuing warnings.
1556
1557 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1558 supplied by Stefan Weber. I made a further small update for 7.8 because
1559 there is a change of source arrangements: the pcre_searchfuncs.c module is
1560 replaced by pcre_ucd.c.
1561
1562
1563 Version 7.7 07-May-08
1564 ---------------------
1565
1566 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1567 a string to a long long, pretend we don't even have a long long." This is
1568 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1569
1570 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1571 pre-7.6 versions, which defined a global no_arg variable instead of putting
1572 it in the RE class. (See also #8 below.)
1573
1574 3. Remove a line of dead code, identified by coverity and reported by Nuno
1575 Lopes.
1576
1577 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1578
1579 (1) The include/exclude patterns were being applied to the whole pathnames
1580 of files, instead of just to the final components.
1581
1582 (2) If there was more than one level of directory, the subdirectories were
1583 skipped unless they satisfied the include/exclude conditions. This is
1584 inconsistent with GNU grep (and could even be seen as contrary to the
1585 pcregrep specification - which I improved to make it absolutely clear).
1586 The action now is always to scan all levels of directory, and just
1587 apply the include/exclude patterns to regular files.
1588
1589 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1590 --exclude_dir in the tests to avoid scanning .svn directories.
1591
1592 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1593 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1594 doesn't support NULs in patterns.
1595
1596 7. Added some missing "const"s to declarations of static tables in
1597 pcre_compile.c and pcre_dfa_exec.c.
1598
1599 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1600 caused by fix #2 above. (Subsequently also a second patch to fix the
1601 first patch. And a third patch - this was a messy problem.)
1602
1603 9. Applied Craig's patch to remove the use of push_back().
1604
1605 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1606 matching function regexec().
1607
1608 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1609 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1610 references. PCRE supports relative numbers with this syntax (I don't think
1611 Oniguruma does).
1612
1613 12. Previously, a group with a zero repeat such as (...){0} was completely
1614 omitted from the compiled regex. However, this means that if the group
1615 was called as a subroutine from elsewhere in the pattern, things went wrong
1616 (an internal error was given). Such groups are now left in the compiled
1617 pattern, with a new opcode that causes them to be skipped at execution
1618 time.
1619
1620 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1621 to the way PCRE behaves:
1622
1623 (a) A lone ] character is dis-allowed (Perl treats it as data).
1624
1625 (b) A back reference to an unmatched subpattern matches an empty string
1626 (Perl fails the current match path).
1627
1628 (c) A data ] in a character class must be notated as \] because if the
1629 first data character in a class is ], it defines an empty class. (In
1630 Perl it is not possible to have an empty class.) The empty class []
1631 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1632 The negative empty class [^] matches any one character, independently
1633 of the DOTALL setting.
1634
1635 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1636 non-existent subpattern following a character class starting with ']' and
1637 containing () gave an internal compiling error instead of "reference to
1638 non-existent subpattern". Fortunately, when the pattern did exist, the
1639 compiled code was correct. (When scanning forwards to check for the
1640 existencd of the subpattern, it was treating the data ']' as terminating
1641 the class, so got the count wrong. When actually compiling, the reference
1642 was subsequently set up correctly.)
1643
1644 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1645 it was being rejected as not supported by pcre_dfa_exec(), even though
1646 other assertions are supported. I have made pcre_dfa_exec() support
1647 (*FAIL).
1648
1649 16. The implementation of 13c above involved the invention of a new opcode,
1650 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1651 cannot be changed at match time, I realized I could make a small
1652 improvement to matching performance by compiling OP_ALLANY instead of
1653 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1654 on the OP_ANY path.
1655
1656 17. Compiling pcretest on Windows with readline support failed without the
1657 following two fixes: (1) Make the unistd.h include conditional on
1658 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1659
1660 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1661 ncurses library to be included for pcretest when ReadLine support is
1662 requested, but also to allow for it to be overridden. This patch came from
1663 Daniel Bergström.
1664
1665 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1666 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1667 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1668 spotting this.
1669
1670
1671 Version 7.6 28-Jan-08
1672 ---------------------
1673
1674 1. A character class containing a very large number of characters with
1675 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1676 overflow.
1677
1678 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1679 HAVE_LONG_LONG is not defined.
1680
1681 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1682 bring them up to date and include new features. This patch includes:
1683
1684 - Fixed PH's badly added libz and libbz2 support.
1685 - Fixed a problem with static linking.
1686 - Added pcredemo. [But later removed - see 7 below.]
1687 - Fixed dftables problem and added an option.
1688 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1689 HAVE_LONG_LONG.
1690 - Added readline support for pcretest.
1691 - Added an listing of the option settings after cmake has run.
1692
1693 4. A user submitted a patch to Makefile that makes it easy to create
1694 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1695 Makefile.am that cause it to include this special target, without
1696 affecting anything else. Note that the same mingw target plus all
1697 the other distribution libraries and programs are now supported
1698 when configuring with CMake (see 6 below) instead of with
1699 Configure/Make.
1700
1701 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1702 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1703 exported in the Windows port". It has not yet been confirmed that the patch
1704 solves the problem, but it does no harm.
1705
1706 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1707 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1708 with CMake, and also correct the comment about stack recursion.
1709
1710 7. Remove the automatic building of pcredemo from the ./configure system and
1711 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1712 of a program that users should build themselves after PCRE is installed, so
1713 building it automatically is not really right. What is more, it gave
1714 trouble in some build environments.
1715
1716 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1717
1718
1719 Version 7.5 10-Jan-08
1720 ---------------------
1721
1722 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1723 values in parens when parsing an RE using the C++ wrapper."
1724
1725 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1726 Characters greater than 255 were excluded from the class instead of being
1727 included.
1728
1729 3. The same bug as (2) above applied to negated POSIX classes such as
1730 [:^space:].
1731
1732 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1733 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1734 I have changed it.
1735
1736 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1737 first named subpattern) and a construct such as (?&a) would reference the
1738 first named subpattern whose name started with "a" (in other words, the
1739 length check was missing). Both these problems are fixed. "Subpattern name
1740 expected" is now given for (?&) (a zero-length name), and this patch also
1741 makes it give the same error for \k'' (previously it complained that that
1742 was a reference to a non-existent subpattern).
1743
1744 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1745 this is right because (?- can be followed by option settings as well as by
1746 digits. I have, however, made the messages clearer.
1747
1748 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1749 than the number used in the conditional) now cause a compile-time error.
1750 This is actually not compatible with Perl, which accepts such patterns, but
1751 treats the conditional as always being FALSE (as PCRE used to), but it
1752 seems to me that giving a diagnostic is better.
1753
1754 8. Change "alphameric" to the more common word "alphanumeric" in comments
1755 and messages.
1756
1757 9. Fix two occurrences of "backslash" in comments that should have been
1758 "backspace".
1759
1760 10. Remove two redundant lines of code that can never be obeyed (their function
1761 was moved elsewhere).
1762
1763 11. The program that makes PCRE's Unicode character property table had a bug
1764 which caused it to generate incorrect table entries for sequences of
1765 characters that have the same character type, but are in different scripts.
1766 It amalgamated them into a single range, with the script of the first of
1767 them. In other words, some characters were in the wrong script. There were
1768 thirteen such cases, affecting characters in the following ranges:
1769
1770 U+002b0 - U+002c1
1771 U+0060c - U+0060d
1772 U+0061e - U+00612
1773 U+0064b - U+0065e
1774 U+0074d - U+0076d
1775 U+01800 - U+01805
1776 U+01d00 - U+01d77
1777 U+01d9b - U+01dbf
1778 U+0200b - U+0200f
1779 U+030fc - U+030fe
1780 U+03260 - U+0327f
1781 U+0fb46 - U+0fbb1
1782 U+10450 - U+1049d
1783
1784 12. The -o option (show only the matching part of a line) for pcregrep was not
1785 compatible with GNU grep in that, if there was more than one match in a
1786 line, it showed only the first of them. It now behaves in the same way as
1787 GNU grep.
1788
1789 13. If the -o and -v options were combined for pcregrep, it printed a blank
1790 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1791 does the same. The return code can be used to tell if there were any
1792 non-matching lines.
1793
1794 14. Added --file-offsets and --line-offsets to pcregrep.
1795
1796 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1797 infinitely looping recursion. The bug was that positive lookaheads were not
1798 being skipped when checking for a possible empty match (negative lookaheads
1799 and both kinds of lookbehind were skipped).
1800
1801 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1802 inclusion of <windows.h> to before rather than after the definition of
1803 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1804
1805 17. Specifying a possessive quantifier with a specific limit for a Unicode
1806 character property caused pcre_compile() to compile bad code, which led at
1807 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1808 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1809 caused the error; without that there was no problem.
1810
1811 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1812
1813 19. Added --enable-pcretest-libreadline.
1814
1815 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1816 RE::GlobalReplace(). As a result, the number of replacements returned was
1817 double what it should be. I removed one of the increments, but Craig sent a
1818 later patch that removed the other one (the right fix) and added unit tests
1819 that check the return values (which was not done before).
1820
1821 21. Several CMake things:
1822
1823 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1824 the names libpcre and libpcreposix, not just pcre and pcreposix.
1825
1826 (2) The above change means that pcretest and pcregrep are now correctly
1827 linked with the newly-built libraries, not previously installed ones.
1828
1829 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1830
1831 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1832 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1833 UTF-8 newline character). The key issue is that the pattern starts .*;
1834 this means that the match must be either at the beginning, or after a
1835 newline. The bug was in the code for advancing after a failed match and
1836 checking that the new position followed a newline. It was not taking
1837 account of UTF-8 characters correctly.
1838
1839 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1840 character classes. PCRE was not treating the sequence [:...:] as a
1841 character class unless the ... were all letters. Perl, however, seems to
1842 allow any characters between [: and :], though of course it rejects as
1843 unknown any "names" that contain non-letters, because all the known class
1844 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1845 for example, whereas PCRE did not - it did not recognize a POSIX character
1846 class. This seemed a bit dangerous, so the code has been changed to be
1847 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1848 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1849 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1850 Perl does, and where it didn't before.
1851
1852 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1853 Windows environments %n is disabled by default.
1854
1855
1856 Version 7.4 21-Sep-07
1857 ---------------------
1858
1859 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1860 means that a class such as [\s] counted as "explicit reference to CR or
1861 LF". That isn't really right - the whole point of the change was to try to
1862 help when there was an actual mention of one of the two characters. So now
1863 the change happens only if \r or \n (or a literal CR or LF) character is
1864 encountered.
1865
1866 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1867 of both had grown to the point where there were only 3 bits left.
1868 Fortunately, there was spare space in the data structure, and so I have
1869 moved the internal flags into a new 16-bit field to free up more option
1870 bits.
1871
1872 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1873 but did not set the internal JCHANGED flag - either of these is enough to
1874 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1875 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1876 start sets both bits.
1877
1878 4. Added options (at build time, compile time, exec time) to change \R from
1879 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1880
1881 5. doc/pcresyntax.html was missing from the distribution.
1882
1883 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1884 compatibility, even though it is no longer used.
1885
1886 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1887 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1888 windows.h file is present (where different names are used). [This was
1889 reversed later after testing - see 16 below.]
1890
1891 8. Changed all #include <config.h> to #include "config.h". There were also
1892 some further <pcre.h> cases that I changed to "pcre.h".
1893
1894 9. When pcregrep was used with the --colour option, it missed the line ending
1895 sequence off the lines that it output.
1896
1897 10. It was pointed out to me that arrays of string pointers cause lots of
1898 relocations when a shared library is dynamically loaded. A technique of
1899 using a single long string with a table of offsets can drastically reduce
1900 these. I have refactored PCRE in four places to do this. The result is
1901 dramatic:
1902
1903 Originally: 290
1904 After changing UCP table: 187
1905 After changing error message table: 43
1906 After changing table of "verbs" 36
1907 After changing table of Posix names 22
1908
1909 Thanks to the folks working on Gregex for glib for this insight.
1910
1911 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1912 unicode-properties was also set.
1913
1914 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1915
1916 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1917 checked only for CRLF.
1918
1919 14. Added casts to pcretest.c to avoid compiler warnings.
1920
1921 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1922
1923 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1924 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1925 entirely. This removes changes made in 7 above.
1926
1927 17. The CMake files have been updated, and there is now more information about
1928 building with CMake in the NON-UNIX-USE document.
1929
1930
1931 Version 7.3 28-Aug-07
1932 ---------------------
1933
1934 1. In the rejigging of the build system that eventually resulted in 7.1, the
1935 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1936 brackets there is not right, since it causes compilers to look for an
1937 installed pcre.h, not the version that is in the source that is being
1938 compiled (which of course may be different). I have changed it back to:
1939
1940 #include "pcre.h"
1941
1942 I have a vague recollection that the change was concerned with compiling in
1943 different directories, but in the new build system, that is taken care of
1944 by the VPATH setting the Makefile.
1945
1946 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1947 when the subject happened to end in the byte 0x85 (e.g. if the last
1948 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1949 characters but of course it shouldn't be taken as a newline when it is part
1950 of another character. The bug was that, for an unlimited repeat of . in
1951 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1952 characters when looking for a newline.
1953
1954 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1955
1956 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1957 in debug output.
1958
1959 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1960 long printing in the pcrecpp unittest when running under MinGW.
1961
1962 6. ESC_K was left out of the EBCDIC table.
1963
1964 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1965 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1966 limit also applies to "virtual nesting" when a pattern is recursive, and in
1967 this case 1000 isn't so big. I have been able to remove this limit at the
1968 expense of backing off one optimization in certain circumstances. Normally,
1969 when pcre_exec() would call its internal match() function recursively and
1970 immediately return the result unconditionally, it uses a "tail recursion"
1971 feature to save stack. However, when a subpattern that can match an empty
1972 string has an unlimited repetition quantifier, it no longer makes this
1973 optimization. That gives it a stack frame in which to save the data for
1974 checking that an empty string has been matched. Previously this was taken
1975 from the 1000-entry workspace that had been reserved. So now there is no
1976 explicit limit, but more stack is used.
1977
1978 8. Applied Daniel's patches to solve problems with the import/export magic
1979 syntax that is required for Windows, and which was going wrong for the
1980 pcreposix and pcrecpp parts of the library. These were overlooked when this
1981 problem was solved for the main library.
1982
1983 9. There were some crude static tests to avoid integer overflow when computing
1984 the size of patterns that contain repeated groups with explicit upper
1985 limits. As the maximum quantifier is 65535, the maximum group length was
1986 set at 30,000 so that the product of these two numbers did not overflow a
1987 32-bit integer. However, it turns out that people want to use groups that
1988 are longer than 30,000 bytes (though not repeat them that many times).
1989 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1990 made it possible to implement the integer overflow checks in a much more
1991 dynamic way, which I have now done. The artificial limitation on group
1992 length has been removed - we now have only the limit on the total length of
1993 the compiled pattern, which depends on the LINK_SIZE setting.
1994
1995 10. Fixed a bug in the documentation for get/copy named substring when
1996 duplicate names are permitted. If none of the named substrings are set, the
1997 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1998 empty string.
1999
2000 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2001 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2002 because the ] is interpreted as the first data character and the
2003 terminating ] is not found. PCRE has been made compatible with Perl in this
2004 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2005 cause memory overwriting.
2006
2007 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2008 string has been matched (to stop an infinite loop). It was not recognizing
2009 a conditional subpattern that could match an empty string if that
2010 subpattern was within another subpattern. For example, it looped when
2011 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2012 condition was not nested. This bug has been fixed.
2013
2014 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2015 past the start of the subject in the presence of bytes with the top bit
2016 set, for example "\x8aBCD".
2017
2018 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2019 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2020
2021 14. Optimized (?!) to (*FAIL).
2022
2023 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2024 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2025 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2026 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2027 does: it's just the validity check that is more restrictive.
2028
2029 16. Inserted checks for integer overflows during escape sequence (backslash)
2030 processing, and also fixed erroneous offset values for syntax errors during
2031 backslash processing.
2032
2033 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2034 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2035
2036 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2037 caused an overrun.
2038
2039 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2040 something other than just ASCII characters) inside a group that had an
2041 unlimited repeat caused a loop at compile time (while checking to see
2042 whether the group could match an empty string).
2043
2044 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2045 [\P{Any}] did so. (Error in the code for printing property names.)
2046
2047 21. An orphan \E inside a character class could cause a crash.
2048
2049 22. A repeated capturing bracket such as (A)? could cause a wild memory
2050 reference during compilation.
2051
2052 23. There are several functions in pcre_compile() that scan along a compiled
2053 expression for various reasons (e.g. to see if it's fixed length for look
2054 behind). There were bugs in these functions when a repeated \p or \P was
2055 present in the pattern. These operators have additional parameters compared
2056 with \d, etc, and these were not being taken into account when moving along
2057 the compiled data. Specifically:
2058
2059 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2060 length.
2061
2062 (b) An item such as \pL+ within a repeated group could cause crashes or
2063 loops.
2064
2065 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2066 "reference to non-existent subpattern" error.
2067
2068 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2069
2070 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2071 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2072
2073 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2074
2075 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2076 character were causing crashes (broken optimization).
2077
2078 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2079 \p or \P) caused a compile-time loop.
2080
2081 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2082 break. For example, the unstudied pattern [\r\n]A does not match the string
2083 "\r\nA" because change 7.0/46 below moves the current point on by two
2084 characters after failing to match at the start. However, the pattern \nA
2085 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2086 the same is true. There doesn't seem any very clean way out of this, but
2087 what I have chosen to do makes the common cases work: PCRE now takes note
2088 of whether there can be an explicit match for \r or \n anywhere in the
2089 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2090 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2091 pattern has explicit CR or LF references.
2092
2093 29. Added (*CR) etc for changing newline setting at start of pattern.
2094
2095
2096 Version 7.2 19-Jun-07
2097 ---------------------
2098
2099 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2100 which is apparently normally available under Windows.
2101
2102 2. Re-jig the pcregrep tests with different newline settings in an attempt
2103 to make them independent of the local environment's newline setting.
2104
2105 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2106
2107 4. Some of the "internals" tests were previously cut out when the link size
2108 was not 2, because the output contained actual offsets. The recent new
2109 "Z" feature of pcretest means that these can be cut out, making the tests
2110 usable with all link sizes.
2111
2112 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2113 stack recursion. This gives a massive performance boost under BSD, but just
2114 a small improvement under Linux. However, it saves one field in the frame
2115 in all cases.
2116
2117 6. Added more features from the forthcoming Perl 5.10:
2118
2119 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2120 recursion call. It refers to the nth most recently opened parentheses.
2121
2122 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2123 to be opened parentheses.
2124
2125 (c) Conditions that refer to capturing parentheses can be specified
2126 relatively, for example, (?(-2)... or (?(+3)...
2127
2128 (d) \K resets the start of the current match so that everything before
2129 is not part of it.
2130
2131 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2132
2133 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2134 reference syntax.
2135
2136 (g) (?| introduces a group in which the numbering of parentheses in each
2137 alternative starts with the same number.
2138
2139 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2140
2141 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2142 PCRE_INFO_JCHANGED.
2143
2144 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2145 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2146 for detecting groups that can match an empty string.
2147
2148 9. A pattern with a very large number of alternatives (more than several
2149 hundred) was running out of internal workspace during the pre-compile
2150 phase, where pcre_compile() figures out how much memory will be needed. A
2151 bit of new cunning has reduced the workspace needed for groups with
2152 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2153 workspace instead of running out of the 4096 that are available.
2154
2155 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2156
2157 11. Applied patch from Google to remove an optimization that didn't quite work.
2158 The report of the bug said:
2159
2160 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2161 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2162 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2163
2164 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2165 it matched the wrong number of bytes.
2166
2167
2168 Version 7.1 24-Apr-07
2169 ---------------------
2170
2171 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2172 that is more "standard", making use of automake and other Autotools. There
2173 is some re-arrangement of the files and adjustment of comments consequent
2174 on this.
2175
2176 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2177 for recursive directory scanning broke on some systems because the files
2178 are not scanned in any specific order and on different systems the order
2179 was different. A call to "sort" has been inserted into RunGrepTest for the
2180 approprate test as a short-term fix. In the longer term there may be an
2181 alternative.
2182
2183 3. I had an email from Eric Raymond about problems translating some of PCRE's
2184 man pages to HTML (despite the fact that I distribute HTML pages, some
2185 people do their own conversions for various reasons). The problems
2186 concerned the use of low-level troff macros .br and .in. I have therefore
2187 removed all such uses from the man pages (some were redundant, some could
2188 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2189 HTML has been updated to handle .nf/.fi and to complain if it encounters
2190 .br or .in.
2191
2192 4. Updated comments in configure.ac that get placed in config.h.in and also
2193 arranged for config.h to be included in the distribution, with the name
2194 config.h.generic, for the benefit of those who have to compile without
2195 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2196
2197 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2198 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2199 makevp.bat for the current PCRE, using the additional files
2200 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2201
2202 6. A Windows user reported a minor discrepancy with test 2, which turned out
2203 to be caused by a trailing space on an input line that had got lost in his
2204 copy. The trailing space was an accident, so I've just removed it.
2205
2206 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2207 that is needed.
2208
2209 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2210 as "const" (a) because they are and (b) because it helps the PHP
2211 maintainers who have recently made a script to detect big data structures
2212 in the php code that should be moved to the .rodata section. I remembered
2213 to update Builducptable as well, so it won't revert if ucptable.h is ever
2214 re-created.
2215
2216 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2217 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2218 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2219 support is not required. This saves 1.5-2K of code, which is important in
2220 some applications.
2221
2222 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2223 so as not to refer to the tables, even though these functions will never be
2224 called when UTF-8 support is disabled. Otherwise there are problems with a
2225 shared library.
2226
2227 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2228
2229 (a) It was defining its arguments as char * instead of void *.
2230
2231 (b) It was assuming that all moves were upwards in memory; this was true
2232 a long time ago when I wrote it, but is no longer the case.
2233
2234 The emulated memove() is provided for those environments that have neither
2235 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2236 is clearly not the case, as these two bugs were recently reported.
2237
2238 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2239 and Detrail to create the HTML documentation, the .txt form of the man
2240 pages, and it removes trailing spaces from listed files. It also creates
2241 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2242 case, it wraps all the #defines with #ifndefs. This script should be run
2243 before "make dist".
2244
2245 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2246 with Unicode property support.
2247
2248 (a) For a maximizing quantifier, if the two different cases of the
2249 character were of different lengths in their UTF-8 codings (there are
2250 some cases like this - I found 11), and the matching function had to
2251 back up over a mixture of the two cases, it incorrectly assumed they
2252 were both the same length.
2253
2254 (b) When PCRE was configured to use the heap rather than the stack for
2255 recursion during matching, it was not correctly preserving the data for
2256 the other case of a UTF-8 character when checking ahead for a match
2257 while processing a minimizing repeat. If the check also involved
2258 matching a wide character, but failed, corruption could cause an
2259 erroneous result when trying to check for a repeat of the original
2260 character.
2261
2262 13. Some tidying changes to the testing mechanism:
2263
2264 (a) The RunTest script now detects the internal link size and whether there
2265 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2266 values substituted by "configure". (The RunGrepTest script already did
2267 this for UTF-8.) The configure.ac script no longer substitutes the
2268 relevant variables.
2269
2270 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2271 with length and offset values. This means that the output is different
2272 for different internal link sizes. Test 2 is skipped for link sizes
2273 other than 2 because of this, bypassing the problem. Unfortunately,
2274 there was also a test in test 3 (the locale tests) that used /B and
2275 failed for link sizes other than 2. Rather than cut the whole test out,
2276 I have added a new /Z option to pcretest that replaces the length and
2277 offset values with spaces. This is now used to make test 3 independent
2278 of link size. (Test 2 will be tidied up later.)
2279
2280 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2281 segmentation fault instead of returning the appropriate error message.
2282
2283 15. In multiline mode when the newline sequence was set to "any", the pattern
2284 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2285 This doesn't seem right; it now treats the CRLF combination as the line
2286 ending, and so does not match in that case. It's only a pattern such as ^$
2287 that would hit this one: something like ^ABC$ would have failed after \r
2288 and then tried again after \r\n.
2289
2290 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2291 in an attempt to make files that differ only in their line terminators
2292 compare equal. This works on Linux.
2293
2294 17. Under certain error circumstances pcregrep might try to free random memory
2295 as it exited. This is now fixed, thanks to valgrind.
2296
2297 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2298 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2299 was because its rules for how to advance for /g after matching an empty
2300 string at the end of a line did not allow for this case. They now check for
2301 it specially.
2302
2303 20. pcretest is supposed to handle patterns and data of any length, by
2304 extending its buffers when necessary. It was getting this wrong when the
2305 buffer for a data line had to be extended.
2306
2307 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2308 CRLF as a newline sequence.
2309
2310 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2311 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2312 I have nevertheless tidied it up.
2313
2314 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2315
2316 24. Added a man page for pcre-config.
2317
2318
2319 Version 7.0 19-Dec-06
2320 ---------------------
2321
2322 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2323 moving to gcc 4.1.1.
2324
2325 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2326 sys/time.h, which is documented as needed for this function. It doesn't
2327 seem to matter on Linux, but it showed up on some releases of OS X.
2328
2329 3. It seems that there are systems where bytes whose values are greater than
2330 127 match isprint() in the "C" locale. The "C" locale should be the
2331 default when a C program starts up. In most systems, only ASCII printing
2332 characters match isprint(). This difference caused the output from pcretest
2333 to vary, making some of the tests fail. I have changed pcretest so that:
2334
2335 (a) When it is outputting text in the compiled version of a pattern, bytes
2336 other than 32-126 are always shown as hex escapes.
2337
2338 (b) When it is outputting text that is a matched part of a subject string,
2339 it does the same, unless a different locale has been set for the match
2340 (using the /L modifier). In this case, it uses isprint() to decide.
2341
2342 4. Fixed a major bug that caused incorrect computation of the amount of memory
2343 required for a compiled pattern when options that changed within the
2344 pattern affected the logic of the preliminary scan that determines the
2345 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2346 that the computed length was too small. The symptoms of this bug were
2347 either the PCRE error "internal error: code overflow" from pcre_compile(),
2348 or a glibc crash with a message such as "pcretest: free(): invalid next
2349 size (fast)". Examples of patterns that provoked this bug (shown in
2350 pcretest format) are:
2351
2352 /(?-x: )/x
2353 /(?x)(?-x: \s*#\s*)/
2354 /((?i)[\x{c0}])/8
2355 /(?i:[\x{c0}])/8
2356
2357 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2358 is now done differently.
2359
2360 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2361 wrapper classes; (b) implement a new function in the C++ scanner that is
2362 more efficient than the old way of doing things because it avoids levels of
2363 recursion in the regex matching; (c) add a paragraph to the documentation
2364 for the FullMatch() function.
2365
2366 6. The escape sequence \n was being treated as whatever was defined as
2367 "newline". Not only was this contrary to the documentation, which states
2368 that \n is character 10 (hex 0A), but it also went horribly wrong when
2369 "newline" was defined as CRLF. This has been fixed.
2370
2371 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2372 was being set to -1 for the "end of line" case (supposedly a value that no
2373 character can have). Though this value is never used (the check for end of
2374 line is "zero bytes in current character"), it caused compiler complaints.
2375 I've changed it to 0xffffffff.
2376
2377 8. In pcre_version.c, the version string was being built by a sequence of
2378 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2379 string (as it is for production releases) called a macro with an empty
2380 argument. The C standard says the result of this is undefined. The gcc
2381 compiler treats it as an empty string (which was what was wanted) but it is
2382 reported that Visual C gives an error. The source has been hacked around to
2383 avoid this problem.
2384
2385 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2386 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2387 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2388 of them did).
2389
2390 10. Originally, pcretest opened its input and output without "b"; then I was
2391 told that "b" was needed in some environments, so it was added for release
2392 5.0 to both the input and output. (It makes no difference on Unix-like
2393 systems.) Later I was told that it is wrong for the input on Windows. I've
2394 now abstracted the modes into two macros, to make it easier to fiddle with
2395 them, and removed "b" from the input mode under Windows.
2396
2397 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2398
2399 12. Added -help and --help to pcretest as an official way of being reminded
2400 of the options.
2401
2402 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2403 and pcrecpp.cc because they annoy compilers at high warning levels.
2404
2405 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2406
2407 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2408 scripts are not C programs :-) and which was not noticed because it works
2409 on Linux.
2410
2411 16. pcretest is supposed to handle any length of pattern and data line (as one
2412 line or as a continued sequence of lines) by extending its input buffer if
2413 necessary. This feature was broken for very long pattern lines, leading to
2414 a string of junk being passed to pcre_compile() if the pattern was longer
2415 than about 50K.
2416
2417 17. I have done a major re-factoring of the way pcre_compile() computes the
2418 amount of memory needed for a compiled pattern. Previously, there was code
2419 that made a preliminary scan of the pattern in order to do this. That was
2420 OK when PCRE was new, but as the facilities have expanded, it has become
2421 harder and harder to keep it in step with the real compile phase, and there
2422 have been a number of bugs (see for example, 4 above). I have now found a
2423 cunning way of running the real compile function in a "fake" mode that
2424 enables it to compute how much memory it would need, while actually only
2425 ever using a few hundred bytes of working memory and without too many
2426 tests of the mode. This should make future maintenance and development
2427 easier. A side effect of this work is that the limit of 200 on the nesting
2428 depth of parentheses has been removed (though this was never a serious
2429 limitation, I suspect). However, there is a downside: pcre_compile() now
2430 runs more slowly than before (30% or more, depending on the pattern). I
2431 hope this isn't a big issue. There is no effect on runtime performance.
2432
2433 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2434 newline (only possible for the last line of a file) and it was a
2435 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2436
2437 19. Added additional timing features to pcretest. (1) The -tm option now times
2438 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2439 separate command line item, by a number that specifies the number of
2440 repeats to use when timing. The default is 50000; this gives better
2441 precision, but takes uncomfortably long for very large patterns.
2442
2443 20. Extended pcre_study() to be more clever in cases where a branch of a
2444 subpattern has no definite first character. For example, (a*|b*)[cd] would
2445 previously give no result from pcre_study(). Now it recognizes that the
2446 first character must be a, b, c, or d.
2447
2448 21. There was an incorrect error "recursive call could loop indefinitely" if
2449 a subpattern (or the entire pattern) that was being tested for matching an
2450 empty string contained only one non-empty item after a nested subpattern.
2451 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2452 incorrectly, because the \d was being skipped in the check.
2453
2454 22. The pcretest program now has a new pattern option /B and a command line
2455 option -b, which is equivalent to adding /B to every pattern. This causes
2456 it to show the compiled bytecode, without the additional information that
2457 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2458 is the same as /B/I).
2459
2460 23. A new optimization is now able automatically to treat some sequences such
2461 as a*b as a*+b. More specifically, if something simple (such as a character
2462 or a simple class like \d) has an unlimited quantifier, and is followed by
2463 something that cannot possibly match the quantified thing, the quantifier
2464 is automatically "possessified".
2465
2466 24. A recursive reference to a subpattern whose number was greater than 39
2467 went wrong under certain circumstances in UTF-8 mode. This bug could also
2468 have affected the operation of pcre_study().
2469
2470 25. Realized that a little bit of performance could be had by replacing
2471 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2472
2473 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2474
2475 27. Possessive quantifiers such as a++ were previously implemented by turning
2476 them into atomic groups such as ($>a+). Now they have their own opcodes,
2477 which improves performance. This includes the automatically created ones
2478 from 23 above.
2479
2480 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2481 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2482 the first matched character to be a colon. This applied both to named and
2483 numbered groups.
2484
2485 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2486
2487 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2488 building PCRE on Windows easier, so I have included it in the distribution.
2489
2490 31. There is now a check in pcretest against a ridiculously large number being
2491 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2492 loop, the loop is abandoned.
2493
2494 32. Forward references to subpatterns in conditions such as (?(2)...) where
2495 subpattern 2 is defined later cause pcre_compile() to search forwards in
2496 the pattern for the relevant set of parentheses. This search went wrong
2497 when there were unescaped parentheses in a character class, parentheses
2498 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2499
2500 33. "Subroutine" calls and backreferences were previously restricted to
2501 referencing subpatterns earlier in the regex. This restriction has now
2502 been removed.
2503
2504 34. Added a number of extra features that are going to be in Perl 5.10. On the
2505 whole, these are just syntactic alternatives for features that PCRE had
2506 previously implemented using the Python syntax or my own invention. The
2507 other formats are all retained for compatibility.
2508
2509 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2510 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2511 also .NET compatible.
2512
2513 (b) A recursion or subroutine call to a named group can now be defined as
2514 (?&name) as well as (?P>name).
2515
2516 (c) A backreference to a named group can now be defined as \k<name> or
2517 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2518 5.10, are also .NET compatible.
2519
2520 (d) A conditional reference to a named group can now use the syntax
2521 (?(<name>) or (?('name') as well as (?(name).
2522
2523 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2524 groups (named and numbered) that are never evaluated inline, but can be
2525 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2526 is always false. There may be only one alternative in such a group.
2527
2528 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2529 as the simple (?(R). The condition is true only if the most recent
2530 recursion is that of the given number or name. It does not search out
2531 through the entire recursion stack.
2532
2533 (g) The escape \gN or \g{N} has been added, where N is a positive or
2534 negative number, specifying an absolute or relative reference.
2535
2536 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2537 some "unreachable code" warnings.
2538
2539 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2540 things, this adds five new scripts.
2541
2542 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2543 There were also incompatibilities regarding the handling of \Q..\E inside
2544 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2545 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2546
2547 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2548 matches an empty string, and forcibly breaks the loop. There were bugs in
2549 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2550 against aaaa the result was just "a" rather than "aaaa", for example. Two
2551 separate and independent bugs (that affected different cases) have been
2552 fixed.
2553
2554 39. Refactored the code to abolish the use of different opcodes for small
2555 capturing bracket numbers. This is a tidy that I avoided doing when I
2556 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2557 The new approach is not only tidier, it makes it possible to reduce the
2558 memory needed to fix the previous bug (38).
2559
2560 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2561 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2562 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2563 mode.
2564
2565 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2566 report.
2567
2568 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2569 copy construction and assignment in the C++ wrapper.
2570
2571 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2572 couple of bugs that could have given wrong results in the "--newline=crlf"
2573 case.
2574
2575 44. Added a number of casts and did some reorganization of signed/unsigned int
2576 variables following suggestions from Dair Grant. Also renamed the variable
2577 "this" as "item" because it is a C++ keyword.
2578
2579 45. Arranged for dftables to add
2580
2581 #include "pcre_internal.h"
2582
2583 to pcre_chartables.c because without it, gcc 4.x may remove the array
2584 definition from the final binary if PCRE is built into a static library and
2585 dead code stripping is activated.
2586
2587 46. For an unanchored pattern, if a match attempt fails at the start of a
2588 newline sequence, and the newline setting is CRLF or ANY, and the next two
2589 characters are CRLF, advance by two characters instead of one.
2590
2591
2592 Version 6.7 04-Jul-06
2593 ---------------------
2594
2595 1. In order to handle tests when input lines are enormously long, pcretest has
2596 been re-factored so that it automatically extends its buffers when
2597 necessary. The code is crude, but this _is_ just a test program. The
2598 default size has been increased from 32K to 50K.
2599
2600 2. The code in pcre_study() was using the value of the re argument before
2601 testing it for NULL. (Of course, in any sensible call of the function, it
2602 won't be NULL.)
2603
2604 3. The memmove() emulation function in pcre_internal.h, which is used on
2605 systems that lack both memmove() and bcopy() - that is, hardly ever -
2606 was missing a "static" storage class specifier.
2607
2608 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2609 containing an extended class (one that cannot be represented by a bitmap
2610 because it contains high-valued characters or Unicode property items, e.g.
2611 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2612 pattern, but PCRE should not loop if you do not (it no longer does).
2613 [Detail: two cases were found: (a) a repeated subpattern containing an
2614 extended class; (b) a recursive reference to a subpattern that followed a
2615 previous extended class. It wasn't skipping over the extended class
2616 correctly when UTF-8 mode was not set.]
2617
2618 5. A negated single-character class was not being recognized as fixed-length
2619 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2620 compile error "lookbehind assertion is not fixed length".
2621
2622 6. The RunPerlTest auxiliary script was showing an unexpected difference
2623 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2624 write a Perl script that can interpret lines of an input file either as
2625 byte characters or as UTF-8, which is what "perltest" was being required to
2626 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2627 can't do is switch easily at run time between having the "use utf8;" pragma
2628 or not. In the end, I fudged it by using the RunPerlTest script to insert
2629 "use utf8;" explicitly for the UTF-8 tests.
2630
2631 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2632 the end of the subject string, contrary to the documentation and to what
2633 Perl does. This was true of both matching functions. Now it matches only at
2634 the start of the subject and immediately after *internal* newlines.
2635
2636 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2637 a pointer to an int instead of a pointer to an unsigned long int. This
2638 caused problems on 64-bit systems.
2639
2640 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2641 instance of the 'standard' template library not being so standard".
2642
2643 10. There was no check on the number of named subpatterns nor the maximum
2644 length of a subpattern name. The product of these values is used to compute
2645 the size of the memory block for a compiled pattern. By supplying a very
2646 long subpattern name and a large number of named subpatterns, the size
2647 computation could be caused to overflow. This is now prevented by limiting
2648 the length of names to 32 characters, and the number of named subpatterns
2649 to 10,000.
2650
2651 11. Subpatterns that are repeated with specific counts have to be replicated in
2652 the compiled pattern. The size of memory for this was computed from the
2653 length of the subpattern and the repeat count. The latter is limited to
2654 65535, but there was no limit on the former, meaning that integer overflow
2655 could in principle occur. The compiled length of a repeated subpattern is
2656 now limited to 30,000 bytes in order to prevent this.
2657
2658 12. Added the optional facility to have named substrings with the same name.
2659
2660 13. Added the ability to use a named substring as a condition, using the
2661 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2662 are numbers (not recommended). Forward references are permitted.
2663
2664 14. Added forward references in named backreferences (if you see what I mean).
2665
2666 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2667 pattern could run off the end of the subject. For example, the pattern
2668 "(?s)(.{1,5})"8 did this with the subject "ab".
2669
2670 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2671 PCRE_CASELESS was set when matching characters that were quantified with ?
2672 or *.
2673
2674 17. A character class other than a single negated character that had a minimum
2675 but no maximum quantifier - for example [ab]{6,} - was not handled
2676 correctly by pce_dfa_exec(). It would match only one character.
2677
2678 18. A valid (though odd) pattern that looked like a POSIX character
2679 class but used an invalid character after [ (for example [[,abc,]]) caused
2680 pcre_compile() to give the error "Failed: internal error: code overflow" or
2681 in some cases to crash with a glibc free() error. This could even happen if
2682 the pattern terminated after [[ but there just happened to be a sequence of
2683 letters, a binary zero, and a closing ] in the memory that followed.
2684
2685 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2686 over the years. Originally (before any Unicode support), just the bottom 8
2687 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2688 output from "man perlunicode" includes this:
2689
2690 The regular expression compiler produces polymorphic opcodes. That
2691 is, the pattern adapts to the data and automatically switches to
2692 the Unicode character scheme when presented with Unicode data--or
2693 instead uses a traditional byte scheme when presented with byte
2694 data.
2695
2696 Sadly, a wide octal escape does not cause a switch, and in a string with
2697 no other multibyte characters, these octal escapes are treated as before.
2698 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2699 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2700 Unicode string.
2701
2702 I have not perpetrated such confusion in PCRE. Up till now, it took just
2703 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2704 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2705 translate to the appropriate multibyte character.
2706
2707 29. Applied some refactoring to reduce the number of warnings from Microsoft
2708 and Borland compilers. This has included removing the fudge introduced
2709 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2710 a warning about an unused variable.
2711
2712 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2713 characters since release 4.0, because Perl (from release 5.004) does not.
2714 [Or at least, is documented not to: some releases seem to be in conflict
2715 with the documentation.] However, when a pattern was studied with
2716 pcre_study() and all its branches started with \s, PCRE still included VT
2717 as a possible starting character. Of course, this did no harm; it just
2718 caused an unnecessary match attempt.
2719
2720 22. Removed a now-redundant internal flag bit that recorded the fact that case
2721 dependency changed within the pattern. This was once needed for "required
2722 byte" processing, but is no longer used. This recovers a now-scarce options
2723 bit. Also moved the least significant internal flag bit to the most-
2724 significant bit of the word, which was not previously used (hangover from
2725 the days when it was an int rather than a uint) to free up another bit for
2726 the future.
2727
2728 23. Added support for CRLF line endings as well as CR and LF. As well as the
2729 default being selectable at build time, it can now be changed at runtime
2730 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2731 specify that it is scanning data with non-default line endings.
2732
2733 24. Changed the definition of CXXLINK to make it agree with the definition of
2734 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2735
2736 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2737 recursions. This makes a big different to stack usage for some patterns.
2738
2739 26. If a subpattern containing a named recursion or subroutine reference such
2740 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2741 the space required for the compiled pattern went wrong and gave too small a
2742 value. Depending on the environment, this could lead to "Failed: internal
2743 error: code overflow at offset 49" or "glibc detected double free or
2744 corruption" errors.
2745
2746 27. Applied patches from Google (a) to support the new newline modes and (b) to
2747 advance over multibyte UTF-8 characters in GlobalReplace.
2748
2749 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2750 difference for some implementation of PCRE in some Windows version.
2751
2752 29. Added some extra testing facilities to pcretest:
2753
2754 \q<number> in a data line sets the "match limit" value
2755 \Q<number> in a data line sets the "match recursion limt" value
2756 -S <number> sets the stack size, where <number> is in megabytes
2757
2758 The -S option isn't available for Windows.
2759
2760
2761 Version 6.6 06-Feb-06
2762 ---------------------
2763
2764 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2765 in pcreposix.h. I have copied the definition from pcre.h.
2766
2767 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2768 because pcre.h is no longer a built file.
2769
2770 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2771 not normally included in the compiled code.
2772
2773
2774 Version 6.5 01-Feb-06
2775 ---------------------
2776
2777 1. When using the partial match feature with pcre_dfa_exec(), it was not
2778 anchoring the second and subsequent partial matches at the new starting
2779 point. This could lead to incorrect results. For example, with the pattern
2780 /1234/, partially matching against "123" and then "a4" gave a match.
2781
2782 2. Changes to pcregrep:
2783
2784 (a) All non-match returns from pcre_exec() were being treated as failures
2785 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2786 error message is output. Some extra information is given for the
2787 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2788 probably the only errors that are likely to be caused by users (by
2789 specifying a regex that has nested indefinite repeats, for instance).
2790 If there are more than 20 of these errors, pcregrep is abandoned.
2791
2792 (b) A binary zero was treated as data while matching, but terminated the
2793 output line if it was written out. This has been fixed: binary zeroes
2794 are now no different to any other data bytes.
2795
2796 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2797 used to set a locale for matching. The --locale=xxxx long option has
2798 been added (no short equivalent) to specify a locale explicitly on the
2799 pcregrep command, overriding the environment variables.
2800
2801 (d) When -B was used with -n, some line numbers in the output were one less
2802 than they should have been.
2803
2804 (e) Added the -o (--only-matching) option.
2805
2806 (f) If -A or -C was used with -c (count only), some lines of context were
2807 accidentally printed for the final match.
2808
2809 (g) Added the -H (--with-filename) option.
2810
2811 (h) The combination of options -rh failed to suppress file names for files
2812 that were found from directory arguments.
2813
2814 (i) Added the -D (--devices) and -d (--directories) options.
2815
2816 (j) Added the -F (--fixed-strings) option.
2817
2818 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2819
2820 (l) Added the --colo(u)r option.
2821
2822 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2823 is not present by default.
2824
2825 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2826 items such as (?R) or (?1), when the recursion could match a number of
2827 alternatives. If it matched one of the alternatives, but subsequently,
2828 outside the recursion, there was a failure, the code tried to back up into
2829 the recursion. However, because of the way PCRE is implemented, this is not
2830 possible, and the result was an incorrect result from the match.
2831
2832 In order to prevent this happening, the specification of recursion has
2833 been changed so that all such subpatterns are automatically treated as
2834 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2835
2836 4. I had overlooked the fact that, in some locales, there are characters for
2837 which isalpha() is true but neither isupper() nor islower() are true. In
2838 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2839 and ordfeminine) are like this. This affected the treatment of \w and \W
2840 when they appeared in character classes, but not when they appeared outside
2841 a character class. The bit map for "word" characters is now created
2842 separately from the results of isalnum() instead of just taking it from the
2843 upper, lower, and digit maps. (Plus the underscore character, of course.)
2844
2845 5. The above bug also affected the handling of POSIX character classes such as
2846 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2847 permanent tables. Instead, the bit maps for such a class were previously
2848 created as the appropriate unions of the upper, lower, and digit bitmaps.
2849 Now they are created by subtraction from the [[:word:]] class, which has
2850 its own bitmap.
2851
2852 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2853 It is created by subtracting the vertical space characters (\x09, \x0a,
2854 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2855 subtraction was done in the overall bitmap for a character class, meaning
2856 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2857 be recognized. This bug has been fixed.
2858
2859 7. Patches from the folks at Google:
2860
2861 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2862 real life, but is still worth protecting against".
2863
2864 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2865 regular expressions".
2866
2867 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2868 have it.
2869
2870 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2871 "configure" and the latter not, in order to fix a problem somebody had
2872 with compiling the Arg class on HP-UX.
2873
2874 (e) Improve the error-handling of the C++ wrapper a little bit.
2875
2876 (f) New tests for checking recursion limiting.
2877
2878 8. The pcre_memmove() function, which is used only if the environment does not
2879 have a standard memmove() function (and is therefore rarely compiled),
2880 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2881 returning a result (though PCRE never actually uses the result).
2882
2883 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2884 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2885 returned instead of calling malloc() with an overflowing number that would
2886 most likely cause subsequent chaos.
2887
2888 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2889
2890 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2891 with this option is matched, the nmatch and pmatch options of regexec() are
2892 ignored.
2893
2894 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2895 provided in case anyone wants to the the POSIX interface with UTF-8
2896 strings.
2897
2898 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2899 C++ linking (needed for some HP-UX environments).
2900
2901 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2902 (unused parameter) and in the pcre_printint() function (omitted "default"
2903 switch label when the default is to do nothing).
2904
2905 15. Added some code to make it possible, when PCRE is compiled as a C++
2906 library, to replace subject pointers for pcre_exec() with a smart pointer
2907 class, thus making it possible to process discontinuous strings.
2908
2909 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2910 much the same function. They were added by different people who were trying
2911 to make PCRE easy to compile on non-Unix systems. It has been suggested
2912 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2913 for compiling on Windows systems. I have therefore replaced it with
2914 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2915 defaults to "extern" for C or "extern C" for C++, which works fine on
2916 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2917 SCOPE with something explicit in config.h. In addition:
2918
2919 (a) pcreposix.h still had just "extern" instead of either of these macros;
2920 I have replaced it with PCRE_DATA_SCOPE.
2921
2922 (b) Functions such as _pcre_xclass(), which are internal to the library,
2923 but external in the C sense, all had PCRE_EXPORT in their definitions.
2924 This is apparently wrong for the Windows case, so I have removed it.
2925 (It makes no difference on Unix-like systems.)
2926
2927 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2928 of recursive calls to match(). This is different to MATCH_LIMIT because
2929 that limits the total number of calls to match(), not all of which increase
2930 the depth of recursion. Limiting the recursion depth limits the amount of
2931 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2932 when PCRE is compiled, and changed at run time. A patch from Google adds
2933 this functionality to the C++ interface.
2934
2935 18. Changes to the handling of Unicode character properties:
2936
2937 (a) Updated the table to Unicode 4.1.0.
2938
2939 (b) Recognize characters that are not in the table as "Cn" (undefined).
2940
2941 (c) I revised the way the table is implemented to a much improved format
2942 which includes recognition of ranges. It now supports the ranges that
2943 are defined in UnicodeData.txt, and it also amalgamates other
2944 characters into ranges. This has reduced the number of entries in the
2945 table from around 16,000 to around 3,000, thus reducing its size
2946 considerably. I realized I did not need to use a tree structure after
2947 all - a binary chop search is just as efficient. Having reduced the
2948 number of entries, I extended their size from 6 bytes to 8 bytes to
2949 allow for more data.
2950
2951 (d) Added support for Unicode script names via properties such as \p{Han}.
2952
2953 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2954 matching that character.
2955
2956 20. When matching a repeated Unicode property with a minimum greater than zero,
2957 (for example \pL{2,}), PCRE could look past the end of the subject if it
2958 reached it while seeking the minimum number of characters. This could
2959 happen only if some of the characters were more than one byte long, because
2960 there is a check for at least the minimum number of bytes.
2961
2962 21. Refactored the implementation of \p and \P so as to be more general, to
2963 allow for more different types of property in future. This has changed the
2964 compiled form incompatibly. Anybody with saved compiled patterns that use
2965 \p or \P will have to recompile them.
2966
2967 22. Added "Any" and "L&" to the supported property types.
2968
2969 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2970 but give a compile time error if the value is greater than 0xff.
2971
2972 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2973 accidentally not being installed or uninstalled.
2974
2975 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2976 made were to insert the current release number. This seemed silly, because
2977 it made things harder for people building PCRE on systems that don't run
2978 "configure". I have turned pcre.h into a distributed file, no longer built
2979 by "configure", with the version identification directly included. There is
2980 no longer a pcre.h.in file.
2981
2982 However, this change necessitated a change to the pcre-config script as
2983 well. It is built from pcre-config.in, and one of the substitutions was the
2984 release number. I have updated configure.ac so that ./configure now finds
2985 the release number by grepping pcre.h.
2986
2987 26. Added the ability to run the tests under valgrind.
2988
2989
2990 Version 6.4 05-Sep-05
2991 ---------------------
2992
2993 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2994 "--" to be printed when multiple files were scanned, even when none of the
2995 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2996 consider it to be a bug, and have restored the previous behaviour.
2997
2998 2. A couple of code tidies to get rid of compiler warnings.
2999
3000 3. The pcretest program used to cheat by referring to symbols in the library
3001 whose names begin with _pcre_. These are internal symbols that are not
3002 really supposed to be visible externally, and in some environments it is
3003 possible to suppress them. The cheating is now confined to including
3004 certain files from the library's source, which is a bit cleaner.
3005
3006 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3007 file's purpose clearer.
3008
3009 5. Reorganized pcre_ucp_findchar().
3010
3011
3012 Version 6.3 15-Aug-05
3013 ---------------------
3014
3015 1. The file libpcre.pc.in did not have general read permission in the tarball.
3016
3017 2. There were some problems when building without C++ support:
3018
3019 (a) If C++ support was not built, "make install" and "make test" still
3020 tried to test it.
3021
3022 (b) There were problems when the value of CXX was explicitly set. Some
3023 changes have been made to try to fix these, and ...
3024
3025 (c) --disable-cpp can now be used to explicitly disable C++ support.
3026
3027 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3028 backslash in a target when C++ was disabled. This confuses some
3029 versions of "make", apparently. Using an intermediate variable solves
3030 this. (Same for CPP_LOBJ.)
3031
3032 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3033 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3034 necessary on certain architectures.
3035
3036 4. Added a setting of -export-symbols-regex to the link command to remove
3037 those symbols that are exported in the C sense, but actually are local
3038 within the library, and not documented. Their names all begin with
3039 "_pcre_". This is not a perfect job, because (a) we have to except some
3040 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3041 available (and never for static libraries). I have made a note to try to
3042 find a way round (a) in the future.
3043
3044
3045 Version 6.2 01-Aug-05
3046 ---------------------
3047
3048 1. There was no test for integer overflow of quantifier values. A construction
3049 such as {1111111111111111} would give undefined results. What is worse, if
3050 a minimum quantifier for a parenthesized subpattern overflowed and became
3051 negative, the calculation of the memory size went wrong. This could have
3052 led to memory overwriting.
3053
3054 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3055
3056 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3057 operating environments where this matters.
3058
3059 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3060 PCRE options from within the C++ wrapper.
3061
3062 5. Named capturing subpatterns were not being correctly counted when a pattern
3063 was compiled. This caused two problems: (a) If there were more than 100
3064 such subpatterns, the calculation of the memory needed for the whole
3065 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3066 back references of the form \12, where the number was greater than 9, were
3067 not recognized as back references, even though there were sufficient
3068 previous subpatterns.
3069
3070 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3071 versions of gcc, e.g. 2.95.4.
3072
3073
3074 Version 6.1 21-Jun-05
3075 ---------------------
3076
3077 1. There was one reference to the variable "posix" in pcretest.c that was not
3078 surrounded by "#if !defined NOPOSIX".
3079
3080 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3081 the cross-check on the old pcre_info() function, for the benefit of the
3082 cut-down version of PCRE that is currently imported into Exim.
3083
3084 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3085 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3086 patterns that start (?i) but I don't think that matters. The use of (?i) is
3087 just an example; this all applies to the other options as well.
3088
3089 4. Since libtool seems to echo the compile commands it is issuing, the output
3090 from "make" can be reduced a bit by putting "@" in front of each libtool
3091 compile command.
3092
3093 5. Patch from the folks at Google for configure.in to be a bit more thorough
3094 in checking for a suitable C++ installation before trying to compile the
3095 C++ stuff. This should fix a reported problem when a compiler was present,
3096 but no suitable headers.
3097
3098 6. The man pages all had just "PCRE" as their title. I have changed them to
3099 be the relevant file name. I have also arranged that these names are
3100 retained in the file doc/pcre.txt, which is a concatenation in text format
3101 of all the man pages except the little individual ones for each function.
3102
3103 7. The NON-UNIX-USE file had not been updated for the different set of source
3104 files that come with release 6. I also added a few comments about the C++
3105 wrapper.
3106
3107
3108 Version 6.0 07-Jun-05
3109 ---------------------
3110
3111 1. Some minor internal re-organization to help with my DFA experiments.
3112
3113 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3114 didn't matter for the library itself when fully configured, but did matter
3115 when compiling without UCP support, or within Exim, where the ucp files are
3116 not imported.
3117
3118 3. Refactoring of the library code to split up the various functions into
3119 different source modules. The addition of the new DFA matching code (see
3120 below) to a single monolithic source would have made it really too
3121 unwieldy, quite apart from causing all the code to be include in a
3122 statically linked application, when only some functions are used. This is
3123 relevant even without the DFA addition now that patterns can be compiled in
3124 one application and matched in another.
3125
3126 The downside of splitting up is that there have to be some external
3127 functions and data tables that are used internally in different modules of
3128 the library but which are not part of the API. These have all had their
3129 names changed to start with "_pcre_" so that they are unlikely to clash
3130 with other external names.
3131
3132 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3133 a different (DFA) algorithm. Although it is slower than the original
3134 function, it does have some advantages for certain types of matching
3135 problem.
3136
3137 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3138 including restarting after a partial match.
3139
3140 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3141 defined when compiling for Windows was sent to me. I have put it into the
3142 code, though I have no means of testing or verifying it.
3143
3144 7. Added the pcre_refcount() auxiliary function.
3145
3146 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3147 match before or at the first newline in the subject string. In pcretest,
3148 the /f option on a pattern can be used to set this.
3149
3150 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3151 would behave wrongly. This has been present in PCRE since release 4.0.
3152
3153 10. A number of changes to the pcregrep command:
3154
3155 (a) Refactored how -x works; insert ^(...)$ instead of setting
3156 PCRE_ANCHORED and checking the length, in preparation for adding
3157 something similar for -w.
3158
3159 (b) Added the -w (match as a word) option.
3160
3161 (c) Refactored the way lines are read and buffered so as to have more
3162 than one at a time available.
3163
3164 (d) Implemented a pcregrep test script.
3165
3166 (e) Added the -M (multiline match) option. This allows patterns to match
3167 over several lines of the subject. The buffering ensures that at least
3168 8K, or the rest of the document (whichever is the shorter) is available
3169 for matching (and similarly the previous 8K for lookbehind assertions).
3170
3171 (f) Changed the --help output so that it now says
3172
3173 -w, --word-regex(p)
3174
3175 instead of two lines, one with "regex" and the other with "regexp"
3176 because that confused at least one person since the short forms are the
3177 same. (This required a bit of code, as the output is generated
3178 automatically from a table. It wasn't just a text change.)
3179
3180 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3181 option but starts with a hyphen. Could be a pattern or a path name
3182 starting with a hyphen, for instance.
3183
3184 (h) "-" can be given as a file name to represent stdin.
3185
3186 (i) When file names are being printed, "(standard input)" is used for
3187 the standard input, for compatibility with GNU grep. Previously
3188 "<stdin>" was used.
3189
3190 (j) The option --label=xxx can be used to supply a name to be used for
3191 stdin when file names are being printed. There is no short form.
3192
3193 (k) Re-factored the options decoding logic because we are going to add
3194 two more options that take data. Such options can now be given in four
3195 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3196
3197 (l) Added the -A, -B, and -C options for requesting that lines of context
3198 around matches be printed.
3199
3200 (m) Added the -L option to print the names of files that do not contain
3201 any matching lines, that is, the complement of -l.
3202
3203 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3204 continue to scan other files.
3205
3206 (o) The -s option was incorrectly implemented. For compatibility with other
3207 greps, it now suppresses the error message for a non-existent or non-
3208 accessible file (but not the return code). There is a new option called
3209 -q that suppresses the output of matching lines, which was what -s was
3210 previously doing.
3211
3212 (p) Added --include and --exclude options to specify files for inclusion
3213 and exclusion when recursing.
3214
3215 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3216 Hopefully, it now does.
3217
3218 12. Missing cast in pcre_study().
3219
3220 13. Added an "uninstall" target to the makefile.
3221
3222 14. Replaced "extern" in the function prototypes in Makefile.in with
3223 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3224 world, but is set differently for Windows.
3225
3226 15. Added a second compiling function called pcre_compile2(). The only
3227 difference is that it has an extra argument, which is a pointer to an
3228 integer error code. When there is a compile-time failure, this is set
3229 non-zero, in addition to the error test pointer being set to point to an
3230 error message. The new argument may be NULL if no error number is required
3231 (but then you may as well call pcre_compile(), which is now just a
3232 wrapper). This facility is provided because some applications need a
3233 numeric error indication, but it has also enabled me to tidy up the way
3234 compile-time errors are handled in the POSIX wrapper.
3235
3236 16. Added VPATH=.libs to the makefile; this should help when building with one
3237 prefix path and installing with another. (Or so I'm told by someone who
3238 knows more about this stuff than I do.)
3239
3240 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3241 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3242 match everything, including newlines. This is not POSIX-compatible, but
3243 somebody wanted the feature. From pcretest it can be activated by using
3244 both the P and the s flags.
3245
3246 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3247
3248 19. libpcre.pc was being incorrectly installed as executable.
3249
3250 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3251 it now also looks for '\r' so that it will work unmodified on Windows.
3252
3253 21. Added Google's contributed C++ wrapper to the distribution.
3254
3255 22. Added some untidy missing memory free() calls in pcretest, to keep
3256 Electric Fence happy when testing.
3257
3258
3259
3260 Version 5.0 13-Sep-04
3261 ---------------------
3262
3263 1. Internal change: literal characters are no longer packed up into items
3264 containing multiple characters in a single byte-string. Each character
3265 is now matched using a separate opcode. However, there may be more than one
3266 byte in the character in UTF-8 mode.
3267
3268 2. The pcre_callout_block structure has two new fields: pattern_position and
3269 next_item_length. These contain the offset in the pattern to the next match
3270 item, and its length, respectively.
3271
3272 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3273 insertion of callouts before each pattern item. Added the /C option to
3274 pcretest to make use of this.
3275
3276 4. On the advice of a Windows user, the lines
3277
3278 #if defined(_WIN32) || defined(WIN32)
3279 _setmode( _fileno( stdout ), 0x8000 );
3280 #endif /* defined(_WIN32) || defined(WIN32) */
3281
3282 have been added to the source of pcretest. This apparently does useful
3283 magic in relation to line terminators.
3284
3285 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3286 for the benefit of those environments where the "b" makes a difference.
3287
3288 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3289 to know about it. I have put a hack into configure.in that adds in code
3290 to set GCC=yes if CC=icc. This seems to end up at a point in the
3291 generated configure script that is early enough to affect the setting of
3292 compiler options, which is what is needed, but I have no means of testing
3293 whether it really works. (The user who reported this had patched the
3294 generated configure script, which of course I cannot do.)
3295
3296 LATER: After change 22 below (new libtool files), the configure script
3297 seems to know about icc (and also ecc). Therefore, I have commented out
3298 this hack in configure.in.
3299
3300 7. Added support for pkg-config (2 patches were sent in).
3301
3302 8. Negated POSIX character classes that used a combination of internal tables
3303 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3304 [[:^ascii]]. Typically, they would match almost any characters. The other
3305 POSIX classes were not broken in this way.
3306
3307 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3308 to find the match, as PCRE was deluded into thinking that the match had to
3309 start at the start point or following a newline. The same bug applied to
3310 patterns with negative forward assertions or any backward assertions
3311 preceding ".*" at the start, unless the pattern required a fixed first
3312 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3313
3314 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3315 starting at the last subject character, bytes beyond the end of the subject
3316 string were read.
3317
3318 11. Renamed the variable "class" as "classbits" to make life easier for C++
3319 users. (Previously there was a macro definition, but it apparently wasn't
3320 enough.)
3321
3322 12. Added the new field "tables" to the extra data so that tables can be passed
3323 in at exec time, or the internal tables can be re-selected. This allows
3324 a compiled regex to be saved and re-used at a later time by a different
3325 program that might have everything at different addresses.
3326
3327 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3328 -R library as well as a -L library.
3329
3330 14. The debugging options of pcretest (-d on the command line or D on a
3331 pattern) showed incorrect output for anything following an extended class
3332 that contained multibyte characters and which was followed by a quantifier.
3333
3334 15. Added optional support for general category Unicode character properties
3335 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3336 support. It adds about 90K to the size of the library. The meanings of the
3337 inbuilt class escapes such as \d and \s have NOT been changed.
3338
3339 16. Updated pcredemo.c to include calls to free() to release the memory for the
3340 compiled pattern.
3341
3342 17. The generated file chartables.c was being created in the source directory
3343 instead of in the building directory. This caused the build to fail if the
3344 source directory was different from the building directory, and was
3345 read-only.
3346
3347 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3348 file. No doubt somebody will tell me if they don't make sense... Also added
3349 Dan Mooney's comments about building on OpenVMS.
3350
3351 19. Added support for partial matching via the PCRE_PARTIAL option for
3352 pcre_exec() and the \P data escape in pcretest.
3353
3354 20. Extended pcretest with 3 new pattern features:
3355
3356 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3357 write the compiled pattern to the file whose name is "rest-of-line".
3358 This is a straight binary dump of the data, with the saved pointer to
3359 the character tables forced to be NULL. The study data, if any, is
3360 written too. After writing, pcretest reads a new pattern.
3361
3362 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3363 compiled pattern from the given file. There must not be any
3364 occurrences of "<" in the file name (pretty unlikely); if there are,
3365 pcretest will instead treat the initial "<" as a pattern delimiter.
3366 After reading in the pattern, pcretest goes on to read data lines as
3367 usual.
3368
3369 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3370 and 16-bit fields in a compiled pattern, to simulate a pattern that
3371 was compiled on a host of opposite endianness.
3372
3373 21. The pcre-exec() function can now cope with patterns that were compiled on
3374 hosts of opposite endianness, with this restriction:
3375
3376 As for any compiled expression that is saved and used later, the tables
3377 pointer field cannot be preserved; the extra_data field in the arguments
3378 to pcre_exec() should be used to pass in a tables address if a value
3379 other than the default internal tables were used at compile time.
3380
3381 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3382 now diagnosed as an error. Previously, most of the time, a negative number
3383 would have been treated as zero, but if in addition "ovector" was passed as
3384 NULL, a crash could occur.
3385
3386 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3387 new versions from the libtool 1.5 distribution (the last one is a copy of
3388 a file called libtool.m4). This seems to have fixed the need to patch
3389 "configure" to support Darwin 1.3 (which I used to do). However, I still
3390 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3391 workstation).
3392
3393 24. Changed the PCRE licence to be the more standard "BSD" licence.
3394
3395
3396 Version 4.5 01-Dec-03
3397 ---------------------
3398
3399 1. There has been some re-arrangement of the code for the match() function so
3400 that it can be compiled in a version that does not call itself recursively.
3401 Instead, it keeps those local variables that need separate instances for
3402 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3403 needs to "recurse". Keeping track of where control must go is done by means
3404 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3405 hide most of the details from the main code, and operates only if
3406 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3407 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3408 operating.
3409
3410 To make it easier for callers to provide specially tailored get/free
3411 functions for this usage, two new functions, pcre_stack_malloc, and
3412 pcre_stack_free, are used. They are always called in strict stacking order,
3413 and the size of block requested is always the same.
3414
3415 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3416 PCRE has been compiled to use the stack or the heap for recursion. The
3417 -C option of pcretest uses this to show which version is compiled.
3418
3419 A new data escape \S, is added to pcretest; it causes the amounts of store
3420 obtained and freed by both kinds of malloc/free at match time to be added
3421 to the output.
3422
3423 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3424 what's available on my current Linux desktop machine.
3425
3426 3. When matching a UTF-8 string, the test for a valid string at the start has
3427 been extended. If start_offset is not zero, PCRE now checks that it points
3428 to a byte that is the start of a UTF-8 character. If not, it returns
3429 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3430 this is necessary because there may be backward assertions in the pattern.
3431 When matching the same subject several times, it may save resources to use
3432 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3433
3434 4. The code for checking the validity of UTF-8 strings has been tightened so
3435 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3436 containing "overlong sequences".
3437
3438 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3439 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3440 should have been "&", but it just so happened that all the cases this let
3441 through by mistake were picked up later in the function.
3442
3443 6. I had used a variable called "isblank" - this is a C99 function, causing
3444 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3445
3446 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3447 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3448 so the version is shown; (c) stop "make test" echoing "./RunTest".
3449
3450 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3451
3452 9. The prototype for memmove() for systems that don't have it was using
3453 size_t, but the inclusion of the header that defines size_t was later. I've
3454 moved the #includes for the C headers earlier to avoid this.
3455
3456 10. Added some adjustments to the code to make it easier to compiler on certain
3457 special systems:
3458
3459 (a) Some "const" qualifiers were missing.
3460 (b) Added the macro EXPORT before all exported functions; by default this
3461 is defined to be empty.
3462 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3463 that it reads its output file name as an argument instead of writing
3464 to the standard output and assuming this can be redirected.
3465
3466 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3467 class containing characters with values greater than 255, PCRE compilation
3468 went into a loop.
3469
3470 12. A recursive reference to a subpattern that was within another subpattern
3471 that had a minimum quantifier of zero caused PCRE to crash. For example,
3472 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3473 recursion. If the recursively-called subpattern itself had a zero repeat,
3474 that was OK.
3475
3476 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3477 buffer into which it was copied (for escape processing) was still set at
3478 1024, so long lines caused crashes.
3479
3480 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3481 "internal error: code overflow...". This applied to any character class
3482 that was followed by a possessive quantifier.
3483
3484 15. Modified the Makefile to add libpcre.la as a prerequisite for
3485 libpcreposix.la because I was told this is needed for a parallel build to
3486 work.
3487
3488 16. If a pattern that contained .* following optional items at the start was
3489 studied, the wrong optimizing data was generated, leading to matching
3490 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3491 matching string must start with a or b or c. The correct conclusion for
3492 this pattern is that a match can start with any character.
3493
3494
3495 Version 4.4 13-Aug-03
3496 ---------------------
3497
3498 1. In UTF-8 mode, a character class containing characters with values between
3499 127 and 255 was not handled correctly if the compiled pattern was studied.
3500 In fixing this, I have also improved the studying algorithm for such
3501 classes (slightly).
3502
3503 2. Three internal functions had redundant arguments passed to them. Removal
3504 might give a very teeny performance improvement.
3505
3506 3. Documentation bug: the value of the capture_top field in a callout is *one
3507 more than* the number of the hightest numbered captured substring.
3508
3509 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3510 in incorrectly linking with a previously installed version. They now link
3511 explicitly with libpcre.la.
3512
3513 5. configure.in no longer needs to recognize Cygwin specially.
3514
3515 6. A problem in pcre.in for Windows platforms is fixed.
3516
3517 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3518 pcretest, it used to include the size of the study block as part of its
3519 output. Unfortunately, the structure contains a field that has a different
3520 size on different hardware architectures. This meant that the tests that
3521 showed this size failed. As the block is currently always of a fixed size,
3522 this information isn't actually particularly useful in pcretest output, so
3523 I have just removed it.
3524
3525 8. Three pre-processor statements accidentally did not start in column 1.
3526 Sadly, there are *still* compilers around that complain, even though
3527 standard C has not required this for well over a decade. Sigh.
3528
3529 9. In pcretest, the code for checking callouts passed small integers in the
3530 callout_data field, which is a void * field. However, some picky compilers
3531 complained about the casts involved for this on 64-bit systems. Now
3532 pcretest passes the address of the small integer instead, which should get
3533 rid of the warnings.
3534
3535 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3536 both compile and run time, and gives an error if an invalid UTF-8 sequence
3537 is found. There is a option for disabling this check in cases where the
3538 string is known to be correct and/or the maximum performance is wanted.
3539
3540 11. In response to a bug report, I changed one line in Makefile.in from
3541
3542 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3543 to
3544 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3545
3546 to look similar to other lines, but I have no way of telling whether this
3547 is the right thing to do, as I do not use Windows. No doubt I'll get told
3548 if it's wrong...
3549
3550
3551 Version 4.3 21-May-03
3552 ---------------------
3553
3554 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3555 Makefile.
3556
3557 2. Some refactoring to improve the quality of the code:
3558
3559 (i) The utf8_table... variables are now declared "const".
3560
3561 (ii) The code for \cx, which used the "case flipping" table to upper case
3562 lower case letters, now just substracts 32. This is ASCII-specific,
3563 but the whole concept of \cx is ASCII-specific, so it seems
3564 reasonable.
3565
3566 (iii) PCRE was using its character types table to recognize decimal and
3567 hexadecimal digits in the pattern. This is silly, because it handles
3568 only 0-9, a-f, and A-F, but the character types table is locale-
3569 specific, which means strange things might happen. A private
3570 table is now used for this - though it costs 256 bytes, a table is
3571 much faster than multiple explicit tests. Of course, the standard
3572 character types table is still used for matching digits in subject
3573 strings against \d.
3574
3575 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3576 ending in _t are). So I've renamed it as ESC_tee.
3577
3578 3. The first argument for regexec() in the POSIX wrapper should have been
3579 defined as "const".
3580
3581 4. Changed pcretest to use malloc() for its buffers so that they can be
3582 Electric Fenced for debugging.
3583
3584 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3585 to read one or more bytes before the start of the subject string. Often this
3586 had no effect on PCRE's behaviour, but in some circumstances it could
3587 provoke a segmentation fault.
3588
3589 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3590 to try to read one or more bytes before the start of the subject string.
3591
3592 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3593 UTF-8 support could misbehave in various ways if the subject string
3594 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3595 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3596 back over UTF-8 characters.)
3597
3598
3599 Version 4.2 14-Apr-03
3600 ---------------------
3601
3602 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3603
3604 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3605 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3606 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3607 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3608 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3609 and BUILD_EXEEXT
3610 Note: automatic setting of the BUILD variables is not yet working
3611 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3612 compile-time but not at link-time
3613 [LINK]: use for linking executables only
3614 make different versions for Windows and non-Windows
3615 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3616 libraries
3617 [LINK_FOR_BUILD]: new variable
3618 [OBJEXT]: use throughout
3619 [EXEEXT]: use throughout
3620 <winshared>: new target
3621 <wininstall>: new target
3622 <dftables.o>: use native compiler
3623 <dftables>: use native linker
3624 <install>: handle Windows platform correctly
3625 <clean>: ditto
3626 <check>: ditto
3627 copy DLL to top builddir before testing
3628
3629 As part of these changes, -no-undefined was removed again. This was reported
3630 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3631 in any case.
3632
3633 3. Some tidies to get rid of compiler warnings:
3634
3635 . In the match_data structure, match_limit was an unsigned long int, whereas
3636 match_call_count was an int. I've made them both unsigned long ints.
3637
3638 . In pcretest the fact that a const uschar * doesn't automatically cast to
3639 a void * provoked a warning.
3640
3641 . Turning on some more compiler warnings threw up some "shadow" variables
3642 and a few more missing casts.
3643
3644 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3645 option, a class that contained a single character with a value between 128
3646 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3647
3648 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3649 option, a class that contained several characters, but with at least one
3650 whose value was between 128 and 255 caused PCRE to crash.
3651
3652
3653 Version 4.1 12-Mar-03
3654 ---------------------
3655
3656 1. Compiling with gcc -pedantic found a couple of places where casts were
3657 needed, and a string in dftables.c that was longer than standard compilers are
3658 required to support.
3659
3660 2. Compiling with Sun's compiler found a few more places where the code could
3661 be tidied up in order to avoid warnings.
3662
3663 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3664 first of these names is deprecated in the latest Autoconf in favour of the name
3665 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3666 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3667 analogy I have changed it to CFLAGS_FOR_BUILD.
3668
3669 4. Added -no-undefined to the linking command in the Makefile, because this is
3670 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3671 linking step for the pcreposix library.
3672
3673 5. PCRE was failing to diagnose the case of two named groups with the same
3674 name.
3675
3676 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3677 literal character that is needed in the subject for a match, and scans along to
3678 ensure that it is present before embarking on the full matching process. This
3679 saves time in cases of nested unlimited repeats that are never going to match.
3680 Problem: the scan can take a lot of time if the subject is very long (e.g.
3681 megabytes), thus penalizing straightforward matches. It is now done only if the
3682 amount of subject to be scanned is less than 1000 bytes.
3683
3684 7. A lesser problem with the same optimization is that it was recording the
3685 first character of an anchored pattern as "needed", thus provoking a search
3686 right along the subject, even when the first match of the pattern was going to
3687 fail. The "needed" character is now not set for anchored patterns, unless it
3688 follows something in the pattern that is of non-fixed length. Thus, it still
3689 fulfils its original purpose of finding quick non-matches in cases of nested
3690 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3691
3692
3693 Version 4.0 17-Feb-03
3694 ---------------------
3695
3696 1. If a comment in an extended regex that started immediately after a meta-item
3697 extended to the end of string, PCRE compiled incorrect data. This could lead to
3698 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3699
3700 2. Moved to autoconf 2.53 and libtool 1.4.2.
3701
3702 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3703 the special perltest8 script is no longer needed - all the tests can be run
3704 from a single perltest script.
3705
3706 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3707 by \s. It has now been removed in PCRE. This means it isn't recognized as
3708 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3709 class [:space:] *does* include VT, thereby creating a mess.
3710
3711 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3712 space and tab.
3713
3714 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3715 its new features into the main test script, reducing the number of scripts.
3716
3717 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3718 were backward compatible, and made the (?i) apply to the whole pattern, as if
3719 /i were given. Now it behaves more logically, and applies the option setting
3720 only to what follows. PCRE has been changed to follow suit. However, if it
3721 finds options settings right at the start of the pattern, it extracts them into
3722 the global options, as before. Thus, they show up in the info data.
3723
3724 8. Added support for the \Q...\E escape sequence. Characters in between are
3725 treated as literals. This is slightly different from Perl in that $ and @ are
3726 also handled as literals inside the quotes. In Perl, they will cause variable
3727 interpolation. Note the following examples:
3728
3729 Pattern PCRE matches Perl matches
3730
3731 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3732 \Qabc\$xyz\E abc\$xyz abc\$xyz
3733 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3734
3735 For compatibility with Perl, \Q...\E sequences are recognized inside character
3736 classes as well as outside them.
3737
3738 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3739 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3740 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3741 signed/unsigned warnings.
3742
3743 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3744 option for pcretest, so I've replaced it by a simple function that does just
3745 that job.
3746
3747 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3748 "pcregrep -".
3749
3750 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3751 Java package. This provides some syntactic sugar for simple cases of what my
3752 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3753 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3754 item, you can use this simplified notation. Note that only makes sense with
3755 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3756 greediness, whatever the setting of the PCRE_UNGREEDY option.
3757
3758 13. A change of greediness default within a pattern was not taking effect at
3759 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3760 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3761 was abstracted outside.
3762
3763 14. PCRE now supports the \G assertion. It is true when the current matching
3764 position is at the start point of the match. This differs from \A when the
3765 starting offset is non-zero. Used with the /g option of pcretest (or similar
3766 code), it works in the same way as it does for Perl's /g option. If all
3767 alternatives of a regex begin with \G, the expression is anchored to the start
3768 match position, and the "anchored" flag is set in the compiled expression.
3769
3770 15. Some bugs concerning the handling of certain option changes within patterns
3771 have been fixed. These applied to options other than (?ims). For example,
3772 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3773 been the other way round. Some of this was related to change 7 above.
3774
3775 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3776 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3777 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3778 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3779
3780 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3781 mode. Unlike ".", it always matches newline, whatever the setting of
3782 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3783 assertions. Perl allows it, but it doesn't (in general) work because it can't
3784 calculate the length of the lookbehind. At least, that's the case for Perl
3785 5.8.0 - I've been told they are going to document that it doesn't work in
3786 future.
3787
3788 18. Added an error diagnosis for escapes that PCRE does not support: these are
3789 \L, \l, \N, \P, \p, \U, \u, and \X.
3790
3791 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3792 reading past the end of the pattern in cases such as /[abcd/.
3793
3794 20. PCRE was getting more memory than necessary for patterns with classes that
3795 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3796
3797 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3798 compiling PCRE for use with Virtual Pascal.
3799
3800 22. Small fix to the Makefile to make it work properly if the build is done
3801 outside the source tree.
3802
3803 23. Added a new extension: a condition to go with recursion. If a conditional
3804 subpattern starts with (?(R) the "true" branch is used if recursion has
3805 happened, whereas the "false" branch is used only at the top level.
3806
3807 24. When there was a very long string of literal characters (over 255 bytes
3808 without UTF support, over 250 bytes with UTF support), the computation of how
3809 much memory was required could be incorrect, leading to segfaults or other
3810 strange effects.
3811
3812 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3813 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3814 there was a subsequent back reference to those brackets. This meant that, for
3815 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3816 possible to check for precisely this case. All we can do is abandon the
3817 optimization if .* occurs inside capturing brackets when there are any back
3818 references whatsoever. (See below for a better fix that came later.)
3819
3820 26. The handling of the optimization for finding the first character of a
3821 non-anchored pattern, and for finding a character that is required later in the
3822 match were failing in some cases. This didn't break the matching; it just
3823 failed to optimize when it could. The way this is done has been re-implemented.
3824
3825 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3826
3827 28. Added a new feature that provides some of the functionality that Perl
3828 provides with (?{...}). The facility is termed a "callout". The way it is done
3829 in PCRE is for the caller to provide an optional function, by setting
3830 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3831 global variable. By default it is unset, which disables all calling out. To get
3832 the function called, the regex must include (?C) at appropriate points. This
3833 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3834 This provides a means of identifying different callout points. When PCRE
3835 reaches such a point in the regex, if pcre_callout has been set, the external
3836 function is called. It is provided with data in a structure called
3837 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3838 matching continues; if it returns a non-zero value, the match at the current
3839 point fails. However, backtracking will occur if possible. [This was changed
3840 later and other features added - see item 49 below.]
3841
3842 29. pcretest is upgraded to test the callout functionality. It provides a
3843 callout function that displays information. By default, it shows the start of
3844 the match and the current position in the text. There are some new data escapes
3845 to vary what happens:
3846
3847 \C+ in addition, show current contents of captured substrings
3848 \C- do not supply a callout function
3849 \C!n return 1 when callout number n is reached
3850 \C!n!m return 1 when callout number n is reached for the mth time
3851
3852 30. If pcregrep was called with the -l option and just a single file name, it
3853 output "<stdin>" if a match was found, instead of the file name.
3854
3855 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3856 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3857 pcre_exec(). This saves a malloc/free per call. The default value of
3858 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3859 when configuring.
3860
3861 32. The default maximum size of a compiled pattern is 64K. There have been a
3862 few cases of people hitting this limit. The code now uses macros to handle the
3863 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3864 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3865 configuring. Tests 2 and 5 work only with 2-byte links because they output
3866 debugging information about compiled patterns.
3867
3868 33. Internal code re-arrangements:
3869
3870 (a) Moved the debugging function for printing out a compiled regex into
3871 its own source file (printint.c) and used #include to pull it into
3872 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3873 separate copies.
3874
3875 (b) Defined the list of op-code names for debugging as a macro in
3876 internal.h so that it is next to the definition of the opcodes.
3877
3878 (c) Defined a table of op-code lengths for simpler skipping along compiled
3879 code. This is again a macro in internal.h so that it is next to the
3880 definition of the opcodes.
3881
3882 34. Added support for recursive calls to individual subpatterns, along the
3883 lines of Robin Houston's patch (but implemented somewhat differently).
3884
3885 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3886 allow it to read and process whole directories in Win32. This code was
3887 contributed by Lionel Fourquaux; it has not been tested by me.
3888
3889 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3890 used to name a group. Names consist of alphanumerics and underscores, and must
3891 be unique. Back references use the syntax (?P=name) and recursive calls use
3892 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3893 numbers. The function pcre_fullinfo() can be used after compilation to extract
3894 a name/number map. There are three relevant calls:
3895
3896 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3897 PCRE_INFO_NAMECOUNT yields the number of entries
3898 PCRE_INFO_NAMETABLE yields a pointer to the map.
3899
3900 The map is a vector of fixed-size entries. The size of each entry depends on
3901 the length of the longest name used. The first two bytes of each entry are the
3902 group number, most significant byte first. There follows the corresponding
3903 name, zero terminated. The names are in alphabetical order.
3904
3905 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3906 case instead of 255. Making it the same both with and without UTF-8 support
3907 means that the same test output works with both.
3908
3909 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3910 calling malloc() with a zero argument.
3911
3912 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3913 optimization. I've improved things by keeping a bitmap of backreferences with
3914 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3915 fact referenced, the optimization can be applied. It is unlikely that a
3916 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3917 the match to follow \n) will appear inside brackets with a number greater than
3918 31, but if it does, any back reference > 31 suppresses the optimization.
3919
3920 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3921 of disabling numbered capturing parentheses. Any opening parenthesis that is
3922 not followed by ? behaves as if it were followed by ?: but named parentheses
3923 can still be used for capturing (and they will acquire numbers in the usual
3924 way).
3925
3926 41. Redesigned the return codes from the match() function into yes/no/error so
3927 that errors can be passed back from deep inside the nested calls. A malloc
3928 failure while inside a recursive subpattern call now causes the
3929 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3930
3931 42. It is now possible to set a limit on the number of times the match()
3932 function is called in a call to pcre_exec(). This facility makes it possible to
3933 limit the amount of recursion and backtracking, though not in a directly
3934 obvious way, because the match() function is used in a number of different
3935 circumstances. The count starts from zero for each position in the subject
3936 string (for non-anchored patterns). The default limit is, for compatibility, a
3937 large number, namely 10 000 000. You can change this in two ways:
3938
3939 (a) When configuring PCRE before making, you can use --with-match-limit=n
3940 to set a default value for the compiled library.
3941
3942 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3943 a different value is set. See 45 below.
3944
3945 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3946
3947 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3948 of things that can be changed at compile time. The first argument specifies
3949 what is wanted and the second points to where the information is to be placed.
3950 The current list of available information is:
3951
3952 PCRE_CONFIG_UTF8
3953
3954 The output is an integer that is set to one if UTF-8 support is available;
3955 otherwise it is set to zero.
3956
3957 PCRE_CONFIG_NEWLINE
3958
3959 The output is an integer that it set to the value of the code that is used for
3960 newline. It is either LF (10) or CR (13).
3961
3962 PCRE_CONFIG_LINK_SIZE
3963
3964 The output is an integer that contains the number of bytes used for internal
3965 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3966
3967 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3968
3969 The output is an integer that contains the threshold above which the POSIX
3970 interface uses malloc() for output vectors. See item 31 above.
3971
3972 PCRE_CONFIG_MATCH_LIMIT
3973
3974 The output is an unsigned integer that contains the default limit of the number
3975 of match() calls in a pcre_exec() execution. See 42 above.
3976
3977 44. pcretest has been upgraded by the addition of the -C option. This causes it
3978 to extract all the available output from the new pcre_config() function, and to
3979 output it. The program then exits immediately.
3980
3981 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3982 order to support additional features. One way would have been to define
3983 pcre_exec2() (for example) with extra arguments, but this would not have been
3984 extensible, and would also have required all calls to the original function to
3985 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3986 is used for passing in "extra" data from pcre_study().
3987
3988 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3989 contains the following fields:
3990
3991 flags a bitmap indicating which of the following fields are set
3992 study_data opaque data from pcre_study()
3993 match_limit a way of specifying a limit on match() calls for a specific
3994 call to pcre_exec()
3995 callout_data data for callouts (see 49 below)
3996
3997 The flag bits are also defined in pcre.h, and are
3998
3999 PCRE_EXTRA_STUDY_DATA
4000 PCRE_EXTRA_MATCH_LIMIT
4001 PCRE_EXTRA_CALLOUT_DATA
4002
4003 The pcre_study() function now returns one of these new pcre_extra blocks, with
4004 the actual study data pointed to by the study_data field, and the
4005 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4006 before. That is, this change is entirely upwards-compatible and requires no
4007 change to existing code.
4008
4009 If you want to pass in additional data to pcre_exec(), you can either place it
4010 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4011 block.
4012
4013 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4014 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4015 times with different match limits, until it finds the minimum value needed for
4016 pcre_exec() to complete. The value is then output. This can be instructive; for
4017 most simple matches the number is quite small, but for pathological cases it
4018 gets very large very quickly.
4019
4020 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4021 returns the size of the data block pointed to by the study_data field in a
4022 pcre_extra block, that is, the value that was passed as the argument to
4023 pcre_malloc() when PCRE was getting memory in which to place the information
4024 created by pcre_study(). The fourth argument should point to a size_t variable.
4025 pcretest has been extended so that this information is shown after a successful
4026 pcre_study() call when information about the compiled regex is being displayed.
4027
4028 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4029 because what follows is always an absolute path. (Later: it turns out that this
4030 is more than cosmetic for MinGW, because it doesn't like empty path
4031 components.)
4032
4033 49. Some changes have been made to the callout feature (see 28 above):
4034
4035 (i) A callout function now has three choices for what it returns:
4036
4037 0 => success, carry on matching
4038 > 0 => failure at this point, but backtrack if possible
4039 < 0 => serious error, return this value from pcre_exec()
4040
4041 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4042 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4043 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4044 use by callout functions. It will never be used by PCRE itself.
4045
4046 (ii) The pcre_extra structure (see 45 above) has a void * field called
4047 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4048 pcre_callout_block structure has a field of the same name. The contents of
4049 the field passed in the pcre_extra structure are passed to the callout
4050 function in the corresponding field in the callout block. This makes it
4051 easier to use the same callout-containing regex from multiple threads. For
4052 testing, the pcretest program has a new data escape
4053
4054 \C*n pass the number n (may be negative) as callout_data
4055
4056 If the callout function in pcretest receives a non-zero value as
4057 callout_data, it returns that value.
4058
4059 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4060 there were some redundant $(CFLAGS) in commands that are now specified as
4061 $(LINK), which already includes $(CFLAGS).
4062
4063 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4064 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4065 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4066 one-byte characters throughout. Note that case-insensitive matching applies
4067 only to characters whose values are less than 256. PCRE doesn't support the
4068 notion of cases for higher-valued characters.
4069
4070 (i) A character class whose characters are all within 0-255 is handled as
4071 a bit map, and the map is inverted for negative classes. Previously, a
4072 character > 255 always failed to match such a class; however it should
4073 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4074
4075 (ii) A negated character class with a single character < 255 is coded as
4076 "not this character" (OP_NOT). This wasn't working properly when the test
4077 character was multibyte, either singly or repeated.
4078
4079 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4080 mode, for example: \x{100}{2,3}.
4081
4082 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4083 singly or repeated) now correctly test multibyte characters. However,
4084 PCRE doesn't recognize any characters with values greater than 255 as
4085 digits, spaces, or word characters. Such characters always match \D, \S,
4086 and \W, and never match \d, \s, or \w.
4087
4088 (v) Classes may now contain characters and character ranges with values
4089 greater than 255. For example: [ab\x{100}-\x{400}].
4090
4091 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4092 PCRE in UTF-8 mode.
4093
4094 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4095 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4096 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4097 value.)
4098
4099 53. The single man page has become too large. I have therefore split it up into
4100 a number of separate man pages. These also give rise to individual HTML pages;
4101 these are now put in a separate directory, and there is an index.html page that
4102 lists them all. Some hyperlinking between the pages has been installed.
4103
4104 54. Added convenience functions for handling named capturing parentheses.
4105
4106 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4107 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4108 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4109 are faulted.
4110
4111 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4112 calling configure. These values are used when compiling the dftables.c program
4113 which is run to generate the source of the default character tables. They
4114 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4115 you will need to set these values.
4116
4117 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4118
4119
4120 Version 3.9 02-Jan-02
4121 ---------------------
4122
4123 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4124
4125 2. If --disable-static was given, the building process failed when trying to
4126 build pcretest and pcregrep. (For some reason it was using libtool to compile
4127 them, which is not right, as they aren't part of the library.)
4128
4129
4130 Version 3.8 18-Dec-01
4131 ---------------------
4132
4133 1. The experimental UTF-8 code was completely screwed up. It was packing the
4134 bytes in the wrong order. How dumb can you get?
4135
4136
4137 Version 3.7 29-Oct-01
4138 ---------------------
4139
4140 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4141 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4142 this didn't happen under Solaris 8, where I normally test things.
4143
4144 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4145 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4146 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4147 this fix an hour or so after the initial 3.7 release.)
4148
4149
4150 Version 3.6 23-Oct-01
4151 ---------------------
4152
4153 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4154 offsets passed as NULL with zero offset count.
4155
4156 2. The config.guess and config.sub files had not been updated when I moved to
4157 the latest autoconf.
4158
4159
4160 Version 3.5 15-Aug-01
4161 ---------------------
4162
4163 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4164 had been forgotten.
4165
4166 2. By using declared but undefined structures, we can avoid using "void"
4167 definitions in pcre.h while keeping the internal definitions of the structures
4168 private.
4169
4170 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4171 user point of view, this means that both static and shared libraries are built
4172 by default, but this can be individually controlled. More of the work of
4173 handling this static/shared cases is now inside libtool instead of PCRE's make
4174 file.
4175
4176 4. The pcretest utility is now installed along with pcregrep because it is
4177 useful for users (to test regexs) and by doing this, it automatically gets
4178 relinked by libtool. The documentation has been turned into a man page, so
4179 there are now .1, .txt, and .html versions in /doc.
4180
4181 5. Upgrades to pcregrep:
4182 (i) Added long-form option names like gnu grep.
4183 (ii) Added --help to list all options with an explanatory phrase.
4184 (iii) Added -r, --recursive to recurse into sub-directories.
4185 (iv) Added -f, --file to read patterns from a file.
4186
4187 6. pcre_exec() was referring to its "code" argument before testing that
4188 argument for NULL (and giving an error if it was NULL).
4189
4190 7. Upgraded Makefile.in to allow for compiling in a different directory from
4191 the source directory.
4192
4193 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4194 options bits, the pointer it was passed was to an int instead of to an unsigned
4195 long int. This mattered only on 64-bit systems.
4196
4197 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4198 generated) instead of pcre.in, which it its source. Also made the same change
4199 in several of the .c files.
4200
4201 10. A new release of gcc defines printf() as a macro, which broke pcretest
4202 because it had an ifdef in the middle of a string argument for printf(). Fixed
4203 by using separate calls to printf().
4204
4205 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4206 script, to force use of CR or LF instead of \n in the source. On non-Unix
4207 systems, the value can be set in config.h.
4208
4209 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4210 absolute limit. Changed the text of the error message to make this clear, and
4211 likewise updated the man page.
4212
4213 13. The limit of 99 on the number of capturing subpatterns has been removed.
4214 The new limit is 65535, which I hope will not be a "real" limit.
4215
4216
4217 Version 3.4 22-Aug-00
4218 ---------------------
4219
4220 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4221
4222 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4223
4224
4225 Version 3.3 01-Aug-00
4226 ---------------------
4227
4228 1. If an octal character was given, but the value was greater than \377, it
4229 was not getting masked to the least significant bits, as documented. This could
4230 lead to crashes in some systems.
4231
4232 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4233 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4234
4235 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4236 These just pass their arguments on to (pcre_free)(), but they are provided
4237 because some uses of PCRE bind it to non-C systems that can call its functions,
4238 but cannot call free() or pcre_free() directly.
4239
4240 4. Add "make test" as a synonym for "make check". Corrected some comments in
4241 the Makefile.
4242
4243 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4244 Makefile.
4245
4246 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4247 command called pgrep for grepping around the active processes.
4248
4249 7. Added the beginnings of support for UTF-8 character strings.
4250
4251 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4252 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4253 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4254 out for the ar command.)
4255
4256
4257 Version 3.2 12-May-00
4258 ---------------------
4259
4260 This is purely a bug fixing release.
4261
4262 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4263 of ZA. This was just one example of several cases that could provoke this bug,
4264 which was introduced by change 9 of version 2.00. The code for breaking
4265 infinite loops after an iteration that matches an empty string was't working
4266 correctly.
4267
4268 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4269 when matched against abbab (for example). After matching an empty string, it
4270 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4271 caused it to match further down the string than it should.
4272
4273 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4274 was there because it doesn't seem to be needed, and it causes trouble on some
4275 systems, as it is not a Standard C header. It has been removed.
4276
4277 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4278 were reported on the Macintosh. The changes were from
4279
4280 while ((c = *(++ptr)) != 0 && c != '\n');
4281 to
4282 while ((c = *(++ptr)) != 0 && c != '\n') ;
4283
4284 Totally extraordinary, but if that's what it takes...
4285
4286 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4287 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4288 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4289 assumes the way PCRE uses memmove() (always moving upwards).
4290
4291 6. PCRE is being used in one environment where strchr() is not available. There
4292 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4293 faster code anyway.
4294
4295
4296 Version 3.1 09-Feb-00
4297 ---------------------
4298
4299 The only change in this release is the fixing of some bugs in Makefile.in for
4300 the "install" target:
4301
4302 (1) It was failing to install pcreposix.h.
4303
4304 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4305
4306
4307 Version 3.0 01-Feb-00
4308 ---------------------
4309
4310 1. Add support for the /+ modifier to perltest (to output $` like it does in
4311 pcretest).
4312
4313 2. Add support for the /g modifier to perltest.
4314
4315 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4316 matches null strings.
4317
4318 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4319 pattern. Perl treats empty patterns specially - it reuses the most recent
4320 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4321 effect.
4322
4323 5. The POSIX interface was broken in that it was just handing over the POSIX
4324 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4325 required a bigger vector, with some working space on the end. This means that
4326 the POSIX wrapper now has to get and free some memory, and copy the results.
4327
4328 6. Added some simple autoconf support, placing the test data and the
4329 documentation in separate directories, re-organizing some of the
4330 information files, and making it build pcre-config (a GNU standard). Also added
4331 libtool support for building PCRE as a shared library, which is now the
4332 default.
4333
4334 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4335 09 are not valid octal constants. Single digits will be used for minor values
4336 less than 10.
4337
4338 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4339 existing programs that set these in the POSIX interface can use PCRE without
4340 modification.
4341
4342 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4343 return all that pcre_info() returns, plus additional data. The pcre_info()
4344 function is retained for compatibility, but is considered to be obsolete.
4345
4346 10. Added experimental recursion feature (?R) to handle one common case that
4347 Perl 5.6 will be able to do with (?p{...}).
4348
4349 11. Added support for POSIX character classes like [:alpha:], which Perl is
4350 adopting.
4351
4352
4353 Version 2.08 31-Aug-99
4354 ----------------------
4355
4356 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4357 trying to match at the startoffset position, but instead was moving forward to
4358 the next newline as if a previous match had failed.
4359
4360 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4361 and could get into a loop if a null string was matched other than at the start
4362 of the subject.
4363
4364 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4365 be distinguished at compile time, and for completeness also added PCRE_DATE.
4366
4367 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4368 in GnuWin32 environments.
4369
4370
4371 Version 2.07 29-Jul-99
4372 ----------------------
4373
4374 1. The documentation is now supplied in plain text form and HTML as well as in
4375 the form of man page sources.
4376
4377 2. C++ compilers don't like assigning (void *) values to other pointer types.
4378 In particular this affects malloc(). Although there is no problem in Standard
4379 C, I've put in casts to keep C++ compilers happy.
4380
4381 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4382 should be (const char *).
4383
4384 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4385 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4386 However, I haven't made this a standard facility. The documentation doesn't
4387 mention it, and the Makefile doesn't support it.
4388
4389 5. The Makefile now contains an "install" target, with editable destinations at
4390 the top of the file. The pcretest program is not installed.
4391
4392 6. pgrep -V now gives the PCRE version number and date.
4393
4394 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4395 causing the entire string to be ignored, instead of just the last character.
4396
4397 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4398 non-matching string, it can take a very, very long time, even for strings of
4399 quite modest length, because of the nested recursion. PCRE now does better in
4400 some of these cases. It does this by remembering the last required literal
4401 character in the pattern, and pre-searching the subject to ensure it is present
4402 before running the real match. In other words, it applies a heuristic to detect
4403 some types of certain failure quickly, and in the above example, if presented
4404 with a string that has no trailing " it gives "no match" very quickly.
4405
4406 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4407 other alternatives are tried instead.
4408
4409
4410 Version 2.06 09-Jun-99
4411 ----------------------
4412
4413 1. Change pcretest's output for amount of store used to show just the code
4414 space, because the remainder (the data block) varies in size between 32-bit and
4415 64-bit systems.
4416
4417 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4418 start matching at. This allows lookbehinds to work when searching for multiple
4419 occurrences in a string.
4420
4421 3. Added additional options to pcretest for testing multiple occurrences:
4422
4423 /+ outputs the rest of the string that follows a match
4424 /g loops for multiple occurrences, using the new startoffset argument
4425 /G loops for multiple occurrences by passing an incremented pointer
4426
4427 4. PCRE wasn't doing the "first character" optimization for patterns starting
4428 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4429 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4430 the letter 'x'. On long subject strings, this gives a significant speed-up.
4431
4432
4433 Version 2.05 21-Apr-99
4434 ----------------------
4435
4436 1. Changed the type of magic_number from int to long int so that it works
4437 properly on 16-bit systems.
4438
4439 2. Fixed a bug which caused patterns starting with .* not to work correctly
4440 when the subject string contained newline characters. PCRE was assuming
4441 anchoring for such patterns in all cases, which is not correct because .* will
4442 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4443 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4444 must be retried after every newline in the subject.
4445
4446
4447 Version 2.04 18-Feb-99
4448 ----------------------
4449
4450 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4451 computation of the store needed to hold the pattern was incorrect (too large).
4452 If such patterns were nested a few deep, this could multiply and become a real
4453 problem.
4454
4455 2. Added /M option to pcretest to show the memory requirement of a specific
4456 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4457
4458 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4459 compiled in such a way that the backtracking after subsequent failure was
4460 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4461 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4462
4463
4464 Version 2.03 02-Feb-99
4465 ----------------------
4466
4467 1. Fixed typo and small mistake in man page.
4468
4469 2. Added 4th condition (GPL supersedes if conflict) and created separate
4470 LICENCE file containing the conditions.
4471
4472 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4473 Perl, that is the internal \ allows the delimiter to be included in the
4474 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4475 the final delimiter, add \ to the end of the pattern (to test the error).
4476
4477 4. Added the convenience functions for extracting substrings after a successful
4478 match. Updated pcretest to make it able to test these functions.
4479
4480
4481 Version 2.02 14-Jan-99
4482 ----------------------
4483
4484 1. Initialized the working variables associated with each extraction so that
4485 their saving and restoring doesn't refer to uninitialized store.
4486
4487 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4488 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4489 fix the problem.
4490
4491 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4492 calls, and wasn't printing the correct value for compiling calls. Increased the
4493 default value of LOOPREPEAT, and the number of significant figures in the
4494 times.
4495
4496 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4497
4498 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4499 a building problem on Windows NT with a FAT file system.
4500
4501
4502 Version 2.01 21-Oct-98
4503 ----------------------
4504
4505 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4506 to character tables built by pcre_maketables() in the current locale. If NULL
4507 is passed, the default tables are used.
4508
4509
4510 Version 2.00 24-Sep-98
4511 ----------------------
4512
4513 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4514 it any more.
4515
4516 2. Allow quantification of (?>) groups, and make it work correctly.
4517
4518 3. The first character computation wasn't working for (?>) groups.
4519
4520 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4521 end of the subject) and add 5.005's \z, which really does match only at the
4522 very end of the subject.
4523
4524 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4525
4526 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4527 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4528 localized options. All options to pcre_study() were also removed.
4529
4530 7. Add other new features from 5.005:
4531
4532 $(?<= positive lookbehind
4533 $(?<! negative lookbehind
4534 (?imsx-imsx) added the unsetting capability
4535 such a setting is global if at outer level; local otherwise
4536 (?imsx-imsx:) non-capturing groups with option setting
4537 (?(cond)re|re) conditional pattern matching
4538
4539 A backreference to itself in a repeated group matches the previous
4540 captured string.
4541
4542 8. General tidying up of studying (both automatic and via "study")
4543 consequential on the addition of new assertions.
4544
4545 9. As in 5.005, unlimited repeated groups that could match an empty substring
4546 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4547 runtime if any iteration does actually match an empty substring.
4548
4549 10. Include the RunTest script in the distribution.
4550
4551 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4552 discrepancies, some of which were old and were also with respect to 5.004. They
4553 have now been fixed.
4554
4555
4556 Version 1.09 28-Apr-98
4557 ----------------------
4558
4559 1. A negated single character class followed by a quantifier with a minimum
4560 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4561 program crashes, or just wrong answers. This did not apply to negated classes
4562 containing more than one character, or to minima other than one.
4563
4564
4565 Version 1.08 27-Mar-98
4566 ----------------------
4567
4568 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4569
4570 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4571 latter must appear before anything that relies on it in the pattern.
4572
4573
4574 Version 1.07 16-Feb-98
4575 ----------------------
4576
4577 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4578 repeat of a potentially empty string).
4579
4580
4581 Version 1.06 23-Jan-98
4582 ----------------------
4583
4584 1. Added Markus Oberhumer's little patches for C++.
4585
4586 2. Literal strings longer than 255 characters were broken.
4587
4588
4589 Version 1.05 23-Dec-97
4590 ----------------------
4591
4592 1. Negated character classes containing more than one character were failing if
4593 PCRE_CASELESS was set at run time.
4594
4595
4596 Version 1.04 19-Dec-97
4597 ----------------------
4598
4599 1. Corrected the man page, where some "const" qualifiers had been omitted.
4600
4601 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4602 input syntax.
4603
4604 3. Fixed memory leak which occurred when a regex with back references was
4605 matched with an offsets vector that wasn't big enough. The temporary memory
4606 that is used in this case wasn't being freed if the match failed.
4607
4608 4. Tidied pcretest to ensure it frees memory that it gets.
4609
4610 5. Temporary memory was being obtained in the case where the passed offsets
4611 vector was exactly big enough.
4612
4613 6. Corrected definition of offsetof() from change 5 below.
4614
4615 7. I had screwed up change 6 below and broken the rules for the use of
4616 setjmp(). Now fixed.
4617
4618
4619 Version 1.03 18-Dec-97
4620 ----------------------
4621
4622 1. A erroneous regex with a missing opening parenthesis was correctly
4623 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4624 on some systems.
4625
4626 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4627 it was reported that one broken compiler failed on the former because "code" is
4628 also an independent variable.
4629
4630 3. The erroneous regex a[]b caused an array overrun reference.
4631
4632 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4633 fail on data ending with that character. (It was going on too far, and checking
4634 the next character, typically a binary zero.) This was specific to the
4635 optimized code for single-character negative classes.
4636
4637 5. Added a contributed patch from the TIN world which does the following:
4638
4639 + Add an undef for memmove, in case the the system defines a macro for it.
4640
4641 + Add a definition of offsetof(), in case there isn't one. (I don't know
4642 the reason behind this - offsetof() is part of the ANSI standard - but
4643 it does no harm).
4644
4645 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4646 most of the places where whitespace preceded '#'. I have given up and
4647 allowed the remaining 2 cases to be at the margin.
4648
4649 + Rename some variables in pcre to eliminate shadowing. This seems very
4650 pedantic, but does no harm, of course.
4651
4652 6. Moved the call to setjmp() into its own function, to get rid of warnings
4653 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4654
4655 7. Constructs such as \d{8,} were compiling into the equivalent of
4656 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4657 outcome, but in this particular case used more store than had been allocated,
4658 which caused the bug to be discovered because it threw up an internal error.
4659
4660 8. The debugging code in both pcre and pcretest for outputting the compiled
4661 form of a regex was going wrong in the case of back references followed by
4662 curly-bracketed repeats.
4663
4664
4665 Version 1.02 12-Dec-97
4666 ----------------------
4667
4668 1. Typos in pcre.3 and comments in the source fixed.
4669
4670 2. Applied a contributed patch to get rid of places where it used to remove
4671 'const' from variables, and fixed some signed/unsigned and uninitialized
4672 variable warnings.
4673
4674 3. Added the "runtest" target to Makefile.
4675
4676 4. Set default compiler flag to -O2 rather than just -O.
4677
4678
4679 Version 1.01 19-Nov-97
4680 ----------------------
4681
4682 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4683 like /([ab]*)*/, that is, for classes with more than one character in them.
4684
4685 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4686 as /((?>a*))*/ (a PCRE_EXTRA facility).
4687
4688
4689 Version 1.00 18-Nov-97
4690 ----------------------
4691
4692 1. Added compile-time macros to support systems such as SunOS4 which don't have
4693 memmove() or strerror() but have other things that can be used instead.
4694
4695 2. Arranged that "make clean" removes the executables.
4696
4697
4698 Version 0.99 27-Oct-97
4699 ----------------------
4700
4701 1. Fixed bug in code for optimizing classes with only one character. It was
4702 initializing a 32-byte map regardless, which could cause it to run off the end
4703 of the memory it had got.
4704
4705 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4706
4707
4708 Version 0.98 22-Oct-97
4709 ----------------------
4710
4711 1. Fixed bug in code for handling temporary memory usage when there are more
4712 back references than supplied space in the ovector. This could cause segfaults.
4713
4714
4715 Version 0.97 21-Oct-97
4716 ----------------------
4717
4718 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4719
4720 2. Optimized negated single characters not to use a bit map.
4721
4722 3. Brought error texts together as macro definitions; clarified some of them;
4723 fixed one that was wrong - it said "range out of order" when it meant "invalid
4724 escape sequence".
4725
4726 4. Changed some char * arguments to const char *.
4727
4728 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4729
4730 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4731 pcretest.
4732
4733
4734 Version 0.96 16-Oct-97
4735 ----------------------
4736
4737 1. Added a simple "pgrep" utility to the distribution.
4738
4739 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4740 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4741 where "ddd" means "one or more decimal digits".
4742
4743 3. Fixed serious bug. If a pattern had a back reference, but the call to
4744 pcre_exec() didn't supply a large enough ovector to record the related
4745 identifying subpattern, the match always failed. PCRE now remembers the number
4746 of the largest back reference, and gets some temporary memory in which to save
4747 the offsets during matching if necessary, in order to ensure that
4748 backreferences always work.
4749
4750 4. Increased the compatibility with Perl in a number of ways:
4751
4752 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4753 to request this handling. The option can be set at compile or exec time.
4754
4755 (b) $ matches before a terminating newline by default; an option
4756 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4757 mode). The option can be set at compile or exec time.
4758
4759 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4760 the same as Perl's. If the decimal number it represents is less than 10
4761 or there aren't that many previous left capturing parentheses, an octal
4762 escape is read. Inside a character class, it's always an octal escape,
4763 even if it is a single digit.
4764
4765 (d) An escaped but undefined alphabetic character is taken as a literal,
4766 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4767 escapes.
4768
4769 (e) {0} is now permitted. (The previous item is removed from the compiled
4770 pattern).
4771
4772 5. Changed all the names of code files so that the basic parts are no longer
4773 than 10 characters, and abolished the teeny "globals.c" file.
4774
4775 6. Changed the handling of character classes; they are now done with a 32-byte
4776 bit map always.
4777
4778 7. Added the -d and /D options to pcretest to make it possible to look at the
4779 internals of compilation without having to recompile pcre.
4780
4781
4782 Version 0.95 23-Sep-97
4783 ----------------------
4784
4785 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4786 \x20 at the start of a run of normal characters. These were being treated as
4787 real characters, instead of the source characters being re-checked.
4788
4789
4790 Version 0.94 18-Sep-97
4791 ----------------------
4792
4793 1. The functions are now thread-safe, with the caveat that the global variables
4794 containing pointers to malloc() and free() or alternative functions are the
4795 same for all threads.
4796
4797 2. Get pcre_study() to generate a bitmap of initial characters for non-
4798 anchored patterns when this is possible, and use it if passed to pcre_exec().
4799
4800
4801 Version 0.93 15-Sep-97
4802 ----------------------
4803
4804 1. /(b)|(:+)/ was computing an incorrect first character.
4805
4806 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4807 but not actually doing anything yet.
4808
4809 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4810 as Perl does (e.g. [-az] or [az-]).
4811
4812 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4813 all possible positions.
4814
4815 5. Split up into different modules to avoid including unneeded functions in a
4816 compiled binary. However, compile and exec are still in one module. The "study"
4817 function is split off.
4818
4819 6. The character tables are now in a separate module whose source is generated
4820 by an auxiliary program - but can then be edited by hand if required. There are
4821 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4822 toupper() in the code.
4823
4824 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4825 make them global. Abolish the function for setting them, as the caller can now
4826 set them directly.
4827
4828
4829 Version 0.92 11-Sep-97
4830 ----------------------
4831
4832 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4833 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4834
4835 2. Caseless matching was not working in character classes if the characters in
4836 the pattern were in upper case.
4837
4838 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4839
4840 4. Make PCRE_ANCHORED public and accept as a compile option.
4841
4842 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4843 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4844 pass them.
4845
4846 6. Give an error if bad option bits passed at compile or run time.
4847
4848 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4849 pcretest to cause it to pass that flag.
4850
4851 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4852 options, and the first character, if set.
4853
4854 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4855
4856
4857 Version 0.91 10-Sep-97
4858 ----------------------
4859
4860 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4861 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4862
4863 2. PCRE was looping on encountering an indefinitely repeated back reference to
4864 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4865 Perl does - treats the match as successful.
4866
4867 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5