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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1022 - (show annotations)
Tue Aug 28 12:28:15 2012 UTC (7 years, 3 months ago) by ph10
File size: 234887 byte(s)
Error occurred while calculating annotation data.
Add support for PCRE_STUDY_EXTRA_NEEDED.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.32
5 ------------
6
7 1. Improved JIT compiler optimizations for first character search and single
8 character iterators.
9
10 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11 Patch by Daniel Richard G.
12
13 3. Single character iterator optimizations in the JIT compiler.
14
15 4. Improved JIT compiler optimizations for character ranges.
16
17 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18 Reported by Giuseppe D'Angelo.
19
20 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21 of a line, was being set incorrectly in cases where .* appeared inside
22 atomic brackets at the start of a pattern, or where there was a subsequent
23 *PRUNE or *SKIP.
24
25 7. Improved instruction cache flush for POWER/PowerPC.
26 Patch by Daniel Richard G.
27
28 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
29 grep:
30
31 (a) There is now no limit to the number of patterns to be matched.
32
33 (b) An error is given if a pattern is too long.
34
35 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
36 are now supported.
37
38 (d) --exclude-from and --include-from (multiple use) have been added.
39
40 (e) Exclusions and inclusions now apply to all files and directories, not
41 just to those obtained from scanning a directory recursively.
42
43 (f) Multiple uses of -f and --file-list are now supported.
44
45 (g) In a Windows environment, the default for -d has been changed from
46 "read" (the GNU grep default) to "skip", because otherwise the presence
47 of a directory in the file list provokes an error.
48
49 (h) The documentation has been revised and clarified in places.
50
51 9. Improve the matching speed of capturing brackets.
52
53 10. Changed the meaning of \X so that it now matches a Unicode extended
54 grapheme cluster.
55
56 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
57 out POSIX threads when JIT support is configured.
58
59 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
60
61
62
63 Version 8.31 06-July-2012
64 -------------------------
65
66 1. Fixing a wrong JIT test case and some compiler warnings.
67
68 2. Removed a bashism from the RunTest script.
69
70 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
71 to unsigned type, result still unsigned" that was given by an MS compiler
72 on encountering the code "-sizeof(xxx)".
73
74 4. Partial matching support is added to the JIT compiler.
75
76 5. Fixed several bugs concerned with partial matching of items that consist
77 of more than one character:
78
79 (a) /^(..)\1/ did not partially match "aba" because checking references was
80 done on an "all or nothing" basis. This also applied to repeated
81 references.
82
83 (b) \R did not give a hard partial match if \r was found at the end of the
84 subject.
85
86 (c) \X did not give a hard partial match after matching one or more
87 characters at the end of the subject.
88
89 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
90 a partial match for the string "\r".
91
92 (e) When newline was set to CRLF, the metacharacter "." did not recognize
93 a partial match for a CR character at the end of the subject string.
94
95 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
96 running pcretest, the text "(JIT)" added to the output whenever JIT is
97 actually used to run the match.
98
99 7. Individual JIT compile options can be set in pcretest by following -s+[+]
100 or /S+[+] with a digit between 1 and 7.
101
102 8. OP_NOT now supports any UTF character not just single-byte ones.
103
104 9. (*MARK) control verb is now supported by the JIT compiler.
105
106 10. The command "./RunTest list" lists the available tests without actually
107 running any of them. (Because I keep forgetting what they all are.)
108
109 11. Add PCRE_INFO_MAXLOOKBEHIND.
110
111 12. Applied a (slightly modified) user-supplied patch that improves performance
112 when the heap is used for recursion (compiled with --disable-stack-for-
113 recursion). Instead of malloc and free for each heap frame each time a
114 logical recursion happens, frames are retained on a chain and re-used where
115 possible. This sometimes gives as much as 30% improvement.
116
117 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
118 call.
119
120 14. As documented, (*COMMIT) is now confined to within a positive assertion.
121
122 15. It is now possible to link pcretest with libedit as an alternative to
123 libreadline.
124
125 16. (*COMMIT) control verb is now supported by the JIT compiler.
126
127 17. The Unicode data tables have been updated to Unicode 6.1.0.
128
129 18. Added --file-list option to pcregrep.
130
131 19. Added binary file support to pcregrep, including the -a, --binary-files,
132 -I, and --text options.
133
134 20. The madvise function is renamed for posix_madvise for QNX compatibility
135 reasons. Fixed by Giuseppe D'Angelo.
136
137 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
138
139 22. Changed the option for creating symbolic links for 16-bit man pages from
140 -s to -sf so that re-installing does not cause issues.
141
142 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
143
144 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
145 matches in certain environments (the workspace was not being correctly
146 retained). Also added to pcre_dfa_exec() a simple plausibility check on
147 some of the workspace data at the beginning of a restart.
148
149 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
150 was not doing so when it should - probably a typo introduced by SVN 528
151 (change 8.10/14).
152
153 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
154 \w+ when the character tables indicated that \x{c4} was a word character.
155 There were several related cases, all because the tests for doing a table
156 lookup were testing for characters less than 127 instead of 255.
157
158 27. If a pattern contains capturing parentheses that are not used in a match,
159 their slots in the ovector are set to -1. For those that are higher than
160 any matched groups, this happens at the end of processing. In the case when
161 there were back references that the ovector was too small to contain
162 (causing temporary malloc'd memory to be used during matching), and the
163 highest capturing number was not used, memory off the end of the ovector
164 was incorrectly being set to -1. (It was using the size of the temporary
165 memory instead of the true size.)
166
167 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
168 ovector size, it uses memory at the end of the block that it has got.
169
170 29. Check for an overlong MARK name and give an error at compile time. The
171 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
172
173 30. JIT compiler update.
174
175 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
176 Rill for the patch.
177
178 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
179
180 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
181
182 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
183 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
184 was enabled.
185
186 35. Improve JIT code generation for greedy plus quantifier.
187
188 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
189 1 to "aa" instead of to an empty string. The bug affected repeated groups
190 that could potentially match an empty string.
191
192 37. Optimizing single character iterators in JIT.
193
194 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
195 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
196 codepoints that are too big for the mode are faulted, and in a UTF mode,
197 disallowed codepoints are also faulted.
198
199 39. If PCRE was compiled with UTF support, in three places in the DFA
200 matcher there was code that should only have been obeyed in UTF mode, but
201 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
202 processing when bytes with values greater than 127 were present. In 16-bit
203 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
204 both cases the values are those that cannot be the first data item in a UTF
205 character. The three items that might have provoked this were recursions,
206 possessively repeated groups, and atomic groups.
207
208 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
209 and pcregrep, because some OS require shared objects to be explicitly
210 passed to ld, causing the link step to fail if they are not.
211
212 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
213 mode, patterns that started with \h* or \R* might be incorrectly matched.
214
215
216 Version 8.30 04-February-2012
217 -----------------------------
218
219 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
220 name is defined in ctype.h.
221
222 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
223 only in quite long subpatterns.
224
225 3. Removed the function pcre_info(), which has been obsolete and deprecated
226 since it was replaced by pcre_fullinfo() in February 2000.
227
228 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
229 match a (*MARK), and the match failed at the start of the subject, a
230 reference to memory before the start of the subject could occur. This bug
231 was introduced by fix 17 of release 8.21.
232
233 5. A reference to an unset group with zero minimum repetition was giving
234 totally wrong answers (in non-JavaScript-compatibility mode). For example,
235 /(another)?(\1?)test/ matched against "hello world test". This bug was
236 introduced in release 8.13.
237
238 6. Add support for 16-bit character strings (a large amount of work involving
239 many changes and refactorings).
240
241 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
242 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
243 from a file.
244
245 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
246 rounding is not applied in this particular case).
247
248 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
249 if they appear, or are escaped, in patterns.
250
251 10. Get rid of a number of -Wunused-but-set-variable warnings.
252
253 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
254 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
255 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
256 also returns the mark "x". This bug applied to capturing parentheses,
257 non-capturing parentheses, and atomic parentheses. It also applied to some
258 assertions.
259
260 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
261 information out of configure.ac instead of relying on pcre.h.generic, which
262 is not stored in the repository.
263
264 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
265 -lreadline.
266
267 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
268
269 15. Applied Graycode's patch to put the top-level frame on the stack rather
270 than the heap when not using the stack for recursion. This gives a
271 performance improvement in many cases when recursion is not deep.
272
273 16. Experimental code added to "pcretest -C" to output the stack frame size.
274
275
276 Version 8.21 12-Dec-2011
277 ------------------------
278
279 1. Updating the JIT compiler.
280
281 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
282 are added as well.
283
284 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
285 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
286 calling _pcre_jit_exec. Some extra comments are added.
287
288 4. (*MARK) settings inside atomic groups that do not contain any capturing
289 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
290 was introduced by change 18 for 8.20.
291
292 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
293 ECMA-262 standard.
294
295 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
296 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
297 This bug was probably introduced by change 9 of 8.13.
298
299 7. While fixing 6 above, I noticed that a number of other items were being
300 incorrectly rejected as "not fixed length". This arose partly because newer
301 opcodes had not been added to the fixed-length checking code. I have (a)
302 corrected the bug and added tests for these items, and (b) arranged for an
303 error to occur if an unknown opcode is encountered while checking for fixed
304 length instead of just assuming "not fixed length". The items that were
305 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
306 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
307 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
308
309 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
310 being incorrectly compiled and would have given unpredicatble results.
311
312 9. A possessively repeated subpattern with minimum repeat count greater than
313 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
314 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
315 the first (A) could occur when it should not.
316
317 10. Add a cast and remove a redundant test from the code.
318
319 11. JIT should use pcre_malloc/pcre_free for allocation.
320
321 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
322 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
323 is anything other than /usr, -L is still shown).
324
325 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
326
327 14. Perl does not support \N without a following name in a [] class; PCRE now
328 also gives an error.
329
330 15. If a forward reference was repeated with an upper limit of around 2000,
331 it caused the error "internal error: overran compiling workspace". The
332 maximum number of forward references (including repeats) was limited by the
333 internal workspace, and dependent on the LINK_SIZE. The code has been
334 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
335 the default depends on LINK_SIZE. There is a new upper limit (for safety)
336 of around 200,000 forward references. While doing this, I also speeded up
337 the filling in of repeated forward references.
338
339 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
340 incorrectly expecting the subject to contain another "a" after the start.
341
342 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
343 in the match, the SKIP should be ignored. This was not happening; instead
344 the SKIP was being treated as NOMATCH. For patterns such as
345 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
346 tested.
347
348 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
349 now much more compatible with Perl, in particular in cases where the result
350 is a non-match for a non-anchored pattern. For example, if
351 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
352 "m", where previously it did not return a name. A side effect of this
353 change is that for partial matches, the last encountered mark name is
354 returned, as for non matches. A number of tests that were previously not
355 Perl-compatible have been moved into the Perl-compatible test files. The
356 refactoring has had the pleasing side effect of removing one argument from
357 the match() function, thus reducing its stack requirements.
358
359 19. If the /S+ option was used in pcretest to study a pattern using JIT,
360 subsequent uses of /S (without +) incorrectly behaved like /S+.
361
362 21. Retrieve executable code size support for the JIT compiler and fixing
363 some warnings.
364
365 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
366 not work when the shorter character appeared right at the end of the
367 subject string.
368
369 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
370 systems.
371
372 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
373 output it when the /M option is used in pcretest.
374
375 25. The CheckMan script was not being included in the distribution. Also, added
376 an explicit "perl" to run Perl scripts from the PrepareRelease script
377 because this is reportedly needed in Windows.
378
379 26. If study data was being save in a file and studying had not found a set of
380 "starts with" bytes for the pattern, the data written to the file (though
381 never used) was taken from uninitialized memory and so caused valgrind to
382 complain.
383
384 27. Updated RunTest.bat as provided by Sheri Pierce.
385
386 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
387
388 29. Computation of memory usage for the table of capturing group names was
389 giving an unnecessarily large value.
390
391
392 Version 8.20 21-Oct-2011
393 ------------------------
394
395 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
396 a POSIX class. After further experiments with Perl, which convinced me that
397 Perl has bugs and confusions, a closing square bracket is no longer allowed
398 in a POSIX name. This bug also affected patterns with classes that started
399 with full stops.
400
401 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
402 captured substring, but while checking the failing first alternative,
403 substring 1 is temporarily captured. If the output vector supplied to
404 pcre_exec() was not big enough for this capture, the yield of the function
405 was still zero ("insufficient space for captured substrings"). This cannot
406 be totally fixed without adding another stack variable, which seems a lot
407 of expense for a edge case. However, I have improved the situation in cases
408 such as /(a)(b)x|abc/ matched against "abc", where the return code
409 indicates that fewer than the maximum number of slots in the ovector have
410 been set.
411
412 3. Related to (2) above: when there are more back references in a pattern than
413 slots in the output vector, pcre_exec() uses temporary memory during
414 matching, and copies in the captures as far as possible afterwards. It was
415 using the entire output vector, but this conflicts with the specification
416 that only 2/3 is used for passing back captured substrings. Now it uses
417 only the first 2/3, for compatibility. This is, of course, another edge
418 case.
419
420 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
421 main code base, and can be used by building with --enable-jit. When this is
422 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
423 runtime --no-jit option is given.
424
425 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
426 ovector, the return from the function was zero, implying that there were
427 other matches that did not fit. The correct "exactly full" value is now
428 returned.
429
430 6. If a subpattern that was called recursively or as a subroutine contained
431 (*PRUNE) or any other control that caused it to give a non-standard return,
432 invalid errors such as "Error -26 (nested recursion at the same subject
433 position)" or even infinite loops could occur.
434
435 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
436 computing the minimum length on reaching *ACCEPT, and so ended up with the
437 wrong value of 1 rather than 0. Further investigation indicates that
438 computing a minimum subject length in the presence of *ACCEPT is difficult
439 (think back references, subroutine calls), and so I have changed the code
440 so that no minimum is registered for a pattern that contains *ACCEPT.
441
442 8. If (*THEN) was present in the first (true) branch of a conditional group,
443 it was not handled as intended. [But see 16 below.]
444
445 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
446 Sheri Pierce.
447
448 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
449 the first byte in a match must be "a".
450
451 11. Change 17 for 8.13 increased the recursion depth for patterns like
452 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
453 pattern contains any instances of (*THEN). If it does not, the old
454 optimizations are restored. It would be nice to do this on a per-group
455 basis, but at the moment that is not feasible.
456
457 12. In some environments, the output of pcretest -C is CRLF terminated. This
458 broke RunTest's code that checks for the link size. A single white space
459 character after the value is now allowed for.
460
461 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
462 For "fr", it uses the Windows-specific input and output files.
463
464 14. If (*THEN) appeared in a group that was called recursively or as a
465 subroutine, it did not work as intended. [But see next item.]
466
467 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
468 pattern fragments (but not containing any | characters). If A and B are
469 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
470 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
471 D. In other words, Perl considers parentheses that do not contain any |
472 characters to be part of a surrounding alternative, whereas PCRE was
473 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
474 differently. PCRE now behaves in the same way as Perl, except in the case
475 of subroutine/recursion calls such as (?1) which have in any case always
476 been different (but PCRE had them first :-).
477
478 16. Related to 15 above: Perl does not treat the | in a conditional group as
479 creating alternatives. Such a group is treated in the same way as an
480 ordinary group without any | characters when processing (*THEN). PCRE has
481 been changed to match Perl's behaviour.
482
483 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
484 RunGrepTest script failed.
485
486 18. Change 22 for version 13 caused atomic groups to use more stack. This is
487 inevitable for groups that contain captures, but it can lead to a lot of
488 stack use in large patterns. The old behaviour has been restored for atomic
489 groups that do not contain any capturing parentheses.
490
491 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
492 suppress the check for a minimum subject length at run time. (If it was
493 given to pcre_exec() or pcre_dfa_exec() it did work.)
494
495 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
496 fail to work when decoding hex characters in data strings in EBCDIC
497 environments.
498
499 21. It appears that in at least one Mac OS environment, the isxdigit() function
500 is implemented as a macro that evaluates to its argument more than once,
501 contravening the C 90 Standard (I haven't checked a later standard). There
502 was an instance in pcretest which caused it to go wrong when processing
503 \x{...} escapes in subject strings. The has been rewritten to avoid using
504 things like p++ in the argument of isxdigit().
505
506
507 Version 8.13 16-Aug-2011
508 ------------------------
509
510 1. The Unicode data tables have been updated to Unicode 6.0.0.
511
512 2. Two minor typos in pcre_internal.h have been fixed.
513
514 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
515 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
516 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
517
518 4. There were a number of related bugs in the code for matching backrefences
519 caselessly in UTF-8 mode when codes for the characters concerned were
520 different numbers of bytes. For example, U+023A and U+2C65 are an upper
521 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
522 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
523 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
524 2-byte code at the end of the subject (it thought there wasn't enough data
525 left).
526
527 5. Comprehensive information about what went wrong is now returned by
528 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
529 as the output vector has at least 2 elements. The offset of the start of
530 the failing character and a reason code are placed in the vector.
531
532 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
533 now returned is for the first byte of the failing character, instead of the
534 last byte inspected. This is an incompatible change, but I hope it is small
535 enough not to be a problem. It makes the returned offset consistent with
536 pcre_exec() and pcre_dfa_exec().
537
538 7. pcretest now gives a text phrase as well as the error number when
539 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
540 failure, the offset and reason code are output.
541
542 8. When \R was used with a maximizing quantifier it failed to skip backwards
543 over a \r\n pair if the subsequent match failed. Instead, it just skipped
544 back over a single character (\n). This seems wrong (because it treated the
545 two characters as a single entity when going forwards), conflicts with the
546 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
547 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
548 has been changed.
549
550 9. Some internal refactoring has changed the processing so that the handling
551 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
552 time (the PCRE_DOTALL option was changed this way some time ago: version
553 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
554 which was always a bit of a fudge. It also means that there is one less
555 argument for the match() function, which reduces its stack requirements
556 slightly. This change also fixes an incompatibility with Perl: the pattern
557 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
558
559 10. More internal refactoring has drastically reduced the number of recursive
560 calls to match() for possessively repeated groups such as (abc)++ when
561 using pcre_exec().
562
563 11. While implementing 10, a number of bugs in the handling of groups were
564 discovered and fixed:
565
566 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
567 (a|)*(?1) gave a compile-time internal error.
568 ((a|)+)+ did not notice that the outer group could match an empty string.
569 (^a|^)+ was not marked as anchored.
570 (.*a|.*)+ was not marked as matching at start or after a newline.
571
572 12. Yet more internal refactoring has removed another argument from the match()
573 function. Special calls to this function are now indicated by setting a
574 value in a variable in the "match data" data block.
575
576 13. Be more explicit in pcre_study() instead of relying on "default" for
577 opcodes that mean there is no starting character; this means that when new
578 ones are added and accidentally left out of pcre_study(), testing should
579 pick them up.
580
581 14. The -s option of pcretest has been documented for ages as being an old
582 synonym of -m (show memory usage). I have changed it to mean "force study
583 for every regex", that is, assume /S for every regex. This is similar to -i
584 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
585 using it. It makes it easier to run collections of tests with and without
586 study enabled, and thereby test pcre_study() more easily. All the standard
587 tests are now run with and without -s (but some patterns can be marked as
588 "never study" - see 20 below).
589
590 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
591 restoration of the capturing data to the outer values was not happening
592 correctly.
593
594 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
595 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
596 pattern had matched an empty string, and so incorrectly returned a no
597 match.
598
599 17. There was optimizing code for the last branch of non-capturing parentheses,
600 and also for the obeyed branch of a conditional subexpression, which used
601 tail recursion to cut down on stack usage. Unfortunately, now that there is
602 the possibility of (*THEN) occurring in these branches, tail recursion is
603 no longer possible because the return has to be checked for (*THEN). These
604 two optimizations have therefore been removed. [But see 8.20/11 above.]
605
606 18. If a pattern containing \R was studied, it was assumed that \R always
607 matched two bytes, thus causing the minimum subject length to be
608 incorrectly computed because \R can also match just one byte.
609
610 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
611 was incorrectly computed.
612
613 20. If /S is present twice on a test pattern in pcretest input, it now
614 *disables* studying, thereby overriding the use of -s on the command line
615 (see 14 above). This is necessary for one or two tests to keep the output
616 identical in both cases.
617
618 21. When (*ACCEPT) was used in an assertion that matched an empty string and
619 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
620
621 22. When an atomic group that contained a capturing parenthesis was
622 successfully matched, but the branch in which it appeared failed, the
623 capturing was not being forgotten if a higher numbered group was later
624 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
625 group 1 to "a", when in fact it should be unset. This applied to multi-
626 branched capturing and non-capturing groups, repeated or not, and also to
627 positive assertions (capturing in negative assertions does not happen
628 in PCRE) and also to nested atomic groups.
629
630 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
631 subject after a captured substring, to make it easier to tell which of a
632 number of identical substrings has been captured.
633
634 24. The way atomic groups are processed by pcre_exec() has been changed so that
635 if they are repeated, backtracking one repetition now resets captured
636 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
637 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
638 "aaa". Previously, it would have been "a". As part of this code
639 refactoring, the way recursive calls are handled has also been changed.
640
641 25. If an assertion condition captured any substrings, they were not passed
642 back unless some other capturing happened later. For example, if
643 (?(?=(a))a) was matched against "a", no capturing was returned.
644
645 26. When studying a pattern that contained subroutine calls or assertions,
646 the code for finding the minimum length of a possible match was handling
647 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
648 group 1 called group 2 while simultaneously a separate group 2 called group
649 1). A stack overflow occurred in this case. I have fixed this by limiting
650 the recursion depth to 10.
651
652 27. Updated RunTest.bat in the distribution to the version supplied by Tom
653 Fortmann. This supports explicit test numbers on the command line, and has
654 argument validation and error reporting.
655
656 28. An instance of \X with an unlimited repeat could fail if at any point the
657 first character it looked at was a mark character.
658
659 29. Some minor code refactoring concerning Unicode properties and scripts
660 should reduce the stack requirement of match() slightly.
661
662 30. Added the '=' option to pcretest to check the setting of unused capturing
663 slots at the end of the pattern, which are documented as being -1, but are
664 not included in the return count.
665
666 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
667 compiled something random. Now it gives a compile-time error (as does
668 Perl).
669
670 32. A *MARK encountered during the processing of a positive assertion is now
671 recorded and passed back (compatible with Perl).
672
673 33. If --only-matching or --colour was set on a pcregrep call whose pattern
674 had alternative anchored branches, the search for a second match in a line
675 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
676 matched the line "0102" twice. The same bug affected patterns that started
677 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
678 twice.
679
680 34. Previously, PCRE did not allow quantification of assertions. However, Perl
681 does, and because of capturing effects, quantifying parenthesized
682 assertions may at times be useful. Quantifiers are now allowed for
683 parenthesized assertions.
684
685 35. A minor code tidy in pcre_compile() when checking options for \R usage.
686
687 36. \g was being checked for fancy things in a character class, when it should
688 just be a literal "g".
689
690 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
691 appearance of a nested POSIX class supersedes an apparent external class.
692 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
693 unescaped square brackets may also appear as part of class names. For
694 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
695 more like Perl. (But see 8.20/1 above.)
696
697 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
698 was because it thought it was \N{name}, which is not supported).
699
700 39. Add minix to OS list not supporting the -S option in pcretest.
701
702 40. PCRE tries to detect cases of infinite recursion at compile time, but it
703 cannot analyze patterns in sufficient detail to catch mutual recursions
704 such as ((?1))((?2)). There is now a runtime test that gives an error if a
705 subgroup is called recursively as a subpattern for a second time at the
706 same position in the subject string. In previous releases this might have
707 been caught by the recursion limit, or it might have run out of stack.
708
709 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
710 happen only once. PCRE was, however incorrectly giving a compile time error
711 "recursive call could loop indefinitely" because it cannot analyze the
712 pattern in sufficient detail. The compile time test no longer happens when
713 PCRE is compiling a conditional subpattern, but actual runaway loops are
714 now caught at runtime (see 40 above).
715
716 42. It seems that Perl allows any characters other than a closing parenthesis
717 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
718 has been changed to be the same.
719
720 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
721 as not to get warnings when autogen.sh is called. Also changed
722 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
723
724 44. To help people who use pcregrep to scan files containing exceedingly long
725 lines, the following changes have been made:
726
727 (a) The default value of the buffer size parameter has been increased from
728 8K to 20K. (The actual buffer used is three times this size.)
729
730 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
731 PCRE is built.
732
733 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
734 to be set at run time.
735
736 (d) Numerical values in pcregrep options can be followed by K or M, for
737 example --buffer-size=50K.
738
739 (e) If a line being scanned overflows pcregrep's buffer, an error is now
740 given and the return code is set to 2.
741
742 45. Add a pointer to the latest mark to the callout data block.
743
744 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
745 partial match of an empty string instead of no match. This was specific to
746 the use of ".".
747
748 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
749 complete match instead of a partial match. This bug was dependent on both
750 the PCRE_UTF8 and PCRE_DOTALL options being set.
751
752 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
753 starting byte set, because \b was not being ignored.
754
755
756 Version 8.12 15-Jan-2011
757 ------------------------
758
759 1. Fixed some typos in the markup of the man pages, and wrote a script that
760 checks for such things as part of the documentation building process.
761
762 2. On a big-endian 64-bit system, pcregrep did not correctly process the
763 --match-limit and --recursion-limit options (added for 8.11). In
764 particular, this made one of the standard tests fail. (The integer value
765 went into the wrong half of a long int.)
766
767 3. If the --colour option was given to pcregrep with -v (invert match), it
768 did strange things, either producing crazy output, or crashing. It should,
769 of course, ignore a request for colour when reporting lines that do not
770 match.
771
772 4. Another pcregrep bug caused similar problems if --colour was specified with
773 -M (multiline) and the pattern match finished with a line ending.
774
775 5. In pcregrep, when a pattern that ended with a literal newline sequence was
776 matched in multiline mode, the following line was shown as part of the
777 match. This seems wrong, so I have changed it.
778
779 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
780 the check for further matches in the same line (so they could be coloured)
781 to overrun the end of the current line. If another match was found, it was
782 incorrectly shown (and then shown again when found in the next line).
783
784 7. If pcregrep was compiled under Windows, there was a reference to the
785 function pcregrep_exit() before it was defined. I am assuming this was
786 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
787 reported by a user. I've moved the definition above the reference.
788
789
790 Version 8.11 10-Dec-2010
791 ------------------------
792
793 1. (*THEN) was not working properly if there were untried alternatives prior
794 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
795 backtracked to try for "b" instead of moving to the next alternative branch
796 at the same level (in this case, to look for "c"). The Perl documentation
797 is clear that when (*THEN) is backtracked onto, it goes to the "next
798 alternative in the innermost enclosing group".
799
800 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
801 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
802 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
803 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
804 (*THEN).
805
806 3. If \s appeared in a character class, it removed the VT character from
807 the class, even if it had been included by some previous item, for example
808 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
809 of \s, but is part of the POSIX "space" class.)
810
811 4. A partial match never returns an empty string (because you can always
812 match an empty string at the end of the subject); however the checking for
813 an empty string was starting at the "start of match" point. This has been
814 changed to the "earliest inspected character" point, because the returned
815 data for a partial match starts at this character. This means that, for
816 example, /(?<=abc)def/ gives a partial match for the subject "abc"
817 (previously it gave "no match").
818
819 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
820 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
821 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
822 has an implication that the given string is incomplete (because a partial
823 match is preferred over a full match). For this reason, these items now
824 give a partial match in this situation. [Aside: previously, the one case
825 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
826 match rather than a full match, which was wrong by the old rules, but is
827 now correct.]
828
829 6. There was a bug in the handling of #-introduced comments, recognized when
830 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
831 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
832 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
833 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
834 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
835 places in pcre_compile().
836
837 7. Related to (6) above, when pcre_compile() was skipping #-introduced
838 comments when looking ahead for named forward references to subpatterns,
839 the only newline sequence it recognized was NL. It now handles newlines
840 according to the set newline convention.
841
842 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
843 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
844 cater for a lack of strerror(). These oversights have been fixed.
845
846 9. Added --match-limit and --recursion-limit to pcregrep.
847
848 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
849
850 11. When the -o option was used, pcregrep was setting a return code of 1, even
851 when matches were found, and --line-buffered was not being honoured.
852
853 12. Added an optional parentheses number to the -o and --only-matching options
854 of pcregrep.
855
856 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
857 can match an empty string. The code to do it in pcretest and pcredemo
858 needed fixing:
859
860 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
861 only one byte after an empty string match just before CRLF (this case
862 just got forgotten; "any" and "anycrlf" were OK).
863
864 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
865 mode when an empty string match preceded an ASCII character followed by
866 a non-ASCII character. (The code for advancing by one character rather
867 than one byte was nonsense.)
868
869 (c) The pcredemo.c sample program did not have any code at all to handle
870 the cases when CRLF is a valid newline sequence.
871
872 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
873 as a starting offset was within the subject string. There is now a new
874 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
875 negative or greater than the length of the string. In order to test this,
876 pcretest is extended to allow the setting of negative starting offsets.
877
878 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
879 starting offset points to the beginning of a UTF-8 character was
880 unnecessarily clumsy. I tidied it up.
881
882 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
883 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
884
885 17. Nobody had reported that the --include_dir option, which was added in
886 release 7.7 should have been called --include-dir (hyphen, not underscore)
887 for compatibility with GNU grep. I have changed it to --include-dir, but
888 left --include_dir as an undocumented synonym, and the same for
889 --exclude-dir, though that is not available in GNU grep, at least as of
890 release 2.5.4.
891
892 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
893 characters from a string of bytes) have been redefined so as not to use
894 loops, in order to improve performance in some environments. At the same
895 time, I abstracted some of the common code into auxiliary macros to save
896 repetition (this should not affect the compiled code).
897
898 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
899 compile-time error is now given if \c is not followed by an ASCII
900 character, that is, a byte less than 128. (In EBCDIC mode, the code is
901 different, and any byte value is allowed.)
902
903 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
904 START_OPTIMIZE option, which is now allowed at compile time - but just
905 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
906 to pcregrep and other applications that have no direct access to PCRE
907 options. The new /Y option in pcretest sets this option when calling
908 pcre_compile().
909
910 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
911 back references. Groups containing recursive back references were forced to
912 be atomic by that change, but in the case of named groups, the amount of
913 memory required was incorrectly computed, leading to "Failed: internal
914 error: code overflow". This has been fixed.
915
916 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
917 pcretest.c, to avoid build problems in some Borland environments.
918
919
920 Version 8.10 25-Jun-2010
921 ------------------------
922
923 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
924 THEN.
925
926 2. (*ACCEPT) was not working when inside an atomic group.
927
928 3. Inside a character class, \B is treated as a literal by default, but
929 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
930 causes the error). The code is unchanged, but I tidied the documentation.
931
932 4. Inside a character class, PCRE always treated \R and \X as literals,
933 whereas Perl faults them if its -w option is set. I have changed PCRE so
934 that it faults them when PCRE_EXTRA is set.
935
936 5. Added support for \N, which always matches any character other than
937 newline. (It is the same as "." when PCRE_DOTALL is not set.)
938
939 6. When compiling pcregrep with newer versions of gcc which may have
940 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
941 declared with attribute warn_unused_result" were given. Just casting the
942 result to (void) does not stop the warnings; a more elaborate fudge is
943 needed. I've used a macro to implement this.
944
945 7. Minor change to pcretest.c to avoid a compiler warning.
946
947 8. Added four artifical Unicode properties to help with an option to make
948 \s etc use properties (see next item). The new properties are: Xan
949 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
950
951 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
952 use Unicode properties. (*UCP) at the start of a pattern can be used to set
953 this option. Modified pcretest to add /W to test this facility. Added
954 REG_UCP to make it available via the POSIX interface.
955
956 10. Added --line-buffered to pcregrep.
957
958 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
959 studied, and the match started with a letter with a code point greater than
960 127 whose first byte was different to the first byte of the other case of
961 the letter, the other case of this starting letter was not recognized
962 (#976).
963
964 12. If a pattern that was studied started with a repeated Unicode property
965 test, for example, \p{Nd}+, there was the theoretical possibility of
966 setting up an incorrect bitmap of starting bytes, but fortunately it could
967 not have actually happened in practice until change 8 above was made (it
968 added property types that matched character-matching opcodes).
969
970 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
971 possible starting bytes for non-anchored patterns.
972
973 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
974 \R, and also a number of cases that involve Unicode properties, both
975 explicit and implicit when PCRE_UCP is set.
976
977 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
978 input, it could crash or give wrong results if characters with values
979 greater than 0xc0 were present in the subject string. (Detail: it assumed
980 UTF-8 input when processing these items.)
981
982 16. Added a lot of (int) casts to avoid compiler warnings in systems where
983 size_t is 64-bit (#991).
984
985 17. Added a check for running out of memory when PCRE is compiled with
986 --disable-stack-for-recursion (#990).
987
988 18. If the last data line in a file for pcretest does not have a newline on
989 the end, a newline was missing in the output.
990
991 19. The default pcre_chartables.c file recognizes only ASCII characters (values
992 less than 128) in its various bitmaps. However, there is a facility for
993 generating tables according to the current locale when PCRE is compiled. It
994 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
995 characters, are recognized by isspace() and therefore were getting set in
996 these tables, and indeed these tables seem to approximate to ISO 8859. This
997 caused a problem in UTF-8 mode when pcre_study() was used to create a list
998 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
999 which of course cannot start UTF-8 characters. I have changed the code so
1000 that only real ASCII characters (less than 128) and the correct starting
1001 bytes for UTF-8 encodings are set for characters greater than 127 when in
1002 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1003 altogether.)
1004
1005 20. Added the /T option to pcretest so as to be able to run tests with non-
1006 standard character tables, thus making it possible to include the tests
1007 used for 19 above in the standard set of tests.
1008
1009 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1010 reference to a subpattern the other side of a comment that contains an
1011 opening parenthesis caused either an internal compiling error, or a
1012 reference to the wrong subpattern.
1013
1014
1015 Version 8.02 19-Mar-2010
1016 ------------------------
1017
1018 1. The Unicode data tables have been updated to Unicode 5.2.0.
1019
1020 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1021 configured.
1022
1023 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1024 original author of that file, following a query about its status.
1025
1026 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1027 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1028
1029 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1030 quantifier applied to a forward-referencing subroutine call, could compile
1031 incorrect code or give the error "internal error: previously-checked
1032 referenced subpattern not found".
1033
1034 6. Both MS Visual Studio and Symbian OS have problems with initializing
1035 variables to point to external functions. For these systems, therefore,
1036 pcre_malloc etc. are now initialized to local functions that call the
1037 relevant global functions.
1038
1039 7. There were two entries missing in the vectors called coptable and poptable
1040 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1041 I've fixed the data, and added a kludgy way of testing at compile time that
1042 the lengths are correct (equal to the number of opcodes).
1043
1044 8. Following on from 7, I added a similar kludge to check the length of the
1045 eint vector in pcreposix.c.
1046
1047 9. Error texts for pcre_compile() are held as one long string to avoid too
1048 much relocation at load time. To find a text, the string is searched,
1049 counting zeros. There was no check for running off the end of the string,
1050 which could happen if a new error number was added without updating the
1051 string.
1052
1053 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1054
1055 11. \K was not working if it appeared in an atomic group or in a group that
1056 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1057 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1058 the assertion is positive, but not if it is negative.
1059
1060 12. Change 11 fortuitously reduced the size of the stack frame used in the
1061 "match()" function of pcre_exec.c by one pointer. Forthcoming
1062 implementation of support for (*MARK) will need an extra pointer on the
1063 stack; I have reserved it now, so that the stack frame size does not
1064 decrease.
1065
1066 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1067 item in branch that calls a recursion is a subroutine call - as in the
1068 second branch in the above example - was incorrectly given the compile-
1069 time error "recursive call could loop indefinitely" because pcre_compile()
1070 was not correctly checking the subroutine for matching a non-empty string.
1071
1072 14. The checks for overrunning compiling workspace could trigger after an
1073 overrun had occurred. This is a "should never occur" error, but it can be
1074 triggered by pathological patterns such as hundreds of nested parentheses.
1075 The checks now trigger 100 bytes before the end of the workspace.
1076
1077 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1078
1079
1080 Version 8.01 19-Jan-2010
1081 ------------------------
1082
1083 1. If a pattern contained a conditional subpattern with only one branch (in
1084 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1085 computed the wrong minimum data length (which is of course zero for such
1086 subpatterns). This could cause incorrect "no match" results.
1087
1088 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1089 the pattern is reset in the first branch, pcre_compile() failed with
1090 "internal error: code overflow at offset...". This happened only when
1091 the reset was to the original external option setting. (An optimization
1092 abstracts leading options settings into an external setting, which was the
1093 cause of this.)
1094
1095 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1096 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1097 assertion pattern did not match (meaning that the assertion was true), it
1098 was incorrectly treated as false if the SKIP had been reached during the
1099 matching. This also applied to assertions used as conditions.
1100
1101 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1102 assertion subpattern, including such a pattern used as a condition,
1103 unpredictable results occurred, instead of the error return
1104 PCRE_ERROR_DFA_UITEM.
1105
1106 5. The C++ GlobalReplace function was not working like Perl for the special
1107 situation when an empty string is matched. It now does the fancy magic
1108 stuff that is necessary.
1109
1110 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1111 removed. (These were left over from very, very early versions of PCRE.)
1112
1113 7. Some cosmetic changes to the code to make life easier when compiling it
1114 as part of something else:
1115
1116 (a) Change DEBUG to PCRE_DEBUG.
1117
1118 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1119 called "current" as "current_branch", to prevent a collision with the
1120 Linux macro when compiled as a kernel module.
1121
1122 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1123 prevent a collision with the Linux macro when compiled as a kernel
1124 module.
1125
1126 8. In pcre_compile() there are some checks for integer overflows that used to
1127 cast potentially large values to (double). This has been changed to that
1128 when building, a check for int64_t is made, and if it is found, it is used
1129 instead, thus avoiding the use of floating point arithmetic. (There is no
1130 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1131 double.
1132
1133 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1134 2005 (difference between two addresses compared to an unsigned value).
1135
1136 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1137 custom one, because of the following reported problem in Windows:
1138
1139 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1140 under Win32.
1141 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1142 therefore missing the function definition.
1143 - The compiler thus generates a "C" signature for the test function.
1144 - The linker fails to find the "C" function.
1145 - PCRE fails to configure if asked to do so against libbz2.
1146
1147 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1148 messages were output:
1149
1150 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1151 rerunning libtoolize, to keep the correct libtool macros in-tree.
1152 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1153
1154 I have done both of these things.
1155
1156 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1157 most of the time, it *can* run out if it is given a pattern that contains a
1158 runaway infinite recursion. I updated the discussion in the pcrestack man
1159 page.
1160
1161 13. Now that we have gone to the x.xx style of version numbers, the minor
1162 version may start with zero. Using 08 or 09 is a bad idea because users
1163 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1164 interpreted as invalid octal numbers. I've updated the previous comment in
1165 configure.ac, and also added a check that gives an error if 08 or 09 are
1166 used.
1167
1168 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1169 causing partial matching to fail when the end of the subject matched \W
1170 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1171
1172 15. There were some discrepancies between the declarations in pcre_internal.h
1173 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1174 their definitions. The declarations used "const uschar *" and the
1175 definitions used USPTR. Even though USPTR is normally defined as "const
1176 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1177 reported that: "This difference in casting confuses some C++ compilers, for
1178 example, SunCC recognizes above declarations as different functions and
1179 generates broken code for hbpcre." I have changed the declarations to use
1180 USPTR.
1181
1182 16. GNU libtool is named differently on some systems. The autogen.sh script now
1183 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1184 (FreeBSD).
1185
1186 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1187 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1188 comment: "Figure out how to create a longlong from a string: strtoll and
1189 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1190 instance, but it only takes 2 args instead of 3!"
1191
1192 18. A subtle bug concerned with back references has been fixed by a change of
1193 specification, with a corresponding code fix. A pattern such as
1194 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1195 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1196 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1197 same bug. Such groups have to be quantified to be useful, or contained
1198 inside another quantified group. (If there's no repetition, the reference
1199 can never match.) The problem arises because, having left the group and
1200 moved on to the rest of the pattern, a later failure that backtracks into
1201 the group uses the captured value from the final iteration of the group
1202 rather than the correct earlier one. I have fixed this in PCRE by forcing
1203 any group that contains a reference to itself to be an atomic group; that
1204 is, there cannot be any backtracking into it once it has completed. This is
1205 similar to recursive and subroutine calls.
1206
1207
1208 Version 8.00 19-Oct-09
1209 ----------------------
1210
1211 1. The table for translating pcre_compile() error codes into POSIX error codes
1212 was out-of-date, and there was no check on the pcre_compile() error code
1213 being within the table. This could lead to an OK return being given in
1214 error.
1215
1216 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1217 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1218 in a Windows environment.
1219
1220 3. The pcregrep --count option prints the count for each file even when it is
1221 zero, as does GNU grep. However, pcregrep was also printing all files when
1222 --files-with-matches was added. Now, when both options are given, it prints
1223 counts only for those files that have at least one match. (GNU grep just
1224 prints the file name in this circumstance, but including the count seems
1225 more useful - otherwise, why use --count?) Also ensured that the
1226 combination -clh just lists non-zero counts, with no names.
1227
1228 4. The long form of the pcregrep -F option was incorrectly implemented as
1229 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1230 but it seems right to fix it, and I didn't think it was worth preserving
1231 the old behaviour.
1232
1233 5. The command line items --regex=pattern and --regexp=pattern were not
1234 recognized by pcregrep, which required --regex pattern or --regexp pattern
1235 (with a space rather than an '='). The man page documented the '=' forms,
1236 which are compatible with GNU grep; these now work.
1237
1238 6. No libpcreposix.pc file was created for pkg-config; there was just
1239 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1240
1241 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1242 when UCP support is not needed, by modifying the Python script that
1243 generates it from Unicode data files. This should not matter if the module
1244 is correctly used as a library, but I received one complaint about 50K of
1245 unwanted data. My guess is that the person linked everything into his
1246 program rather than using a library. Anyway, it does no harm.
1247
1248 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1249 was a minimum greater than 1 for a wide character in a possessive
1250 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1251 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1252 character. Chaos in the form of incorrect output or a compiling loop could
1253 result.
1254
1255 9. The restrictions on what a pattern can contain when partial matching is
1256 requested for pcre_exec() have been removed. All patterns can now be
1257 partially matched by this function. In addition, if there are at least two
1258 slots in the offset vector, the offset of the earliest inspected character
1259 for the match and the offset of the end of the subject are set in them when
1260 PCRE_ERROR_PARTIAL is returned.
1261
1262 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1263 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1264 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1265 and may be more useful for multi-segment matching.
1266
1267 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1268 used to be given if ever the end of the subject was reached; now it is
1269 given only if matching could not proceed because another character was
1270 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1271 string "Z", which now yields "no match" instead of "partial match". In the
1272 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1273 final character ended with (*FAIL).
1274
1275 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1276 if the pattern had a "must contain" character that was already found in the
1277 earlier partial match, unless partial matching was again requested. For
1278 example, with the pattern /dog.(body)?/, the "must contain" character is
1279 "g". If the first part-match was for the string "dog", restarting with
1280 "sbody" failed. This bug has been fixed.
1281
1282 13. The string returned by pcre_dfa_exec() after a partial match has been
1283 changed so that it starts at the first inspected character rather than the
1284 first character of the match. This makes a difference only if the pattern
1285 starts with a lookbehind assertion or \b or \B (\K is not supported by
1286 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1287 matching functions compatible, and I think it's the right thing to do.
1288
1289 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1290 so that the demonstration program is easily available in environments where
1291 PCRE has not been installed from source.
1292
1293 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1294 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1295 library.
1296
1297 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1298 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1299 is not the first non-POSIX option to be added. Clearly some people find
1300 these options useful.
1301
1302 17. If a caller to the POSIX matching function regexec() passes a non-zero
1303 value for nmatch with a NULL value for pmatch, the value of
1304 nmatch is forced to zero.
1305
1306 18. RunGrepTest did not have a test for the availability of the -u option of
1307 the diff command, as RunTest does. It now checks in the same way as
1308 RunTest, and also checks for the -b option.
1309
1310 19. If an odd number of negated classes containing just a single character
1311 interposed, within parentheses, between a forward reference to a named
1312 subpattern and the definition of the subpattern, compilation crashed with
1313 an internal error, complaining that it could not find the referenced
1314 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1315 [The bug was that it was starting one character too far in when skipping
1316 over the character class, thus treating the ] as data rather than
1317 terminating the class. This meant it could skip too much.]
1318
1319 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1320 /g option in pcretest when the pattern contains \K, which makes it possible
1321 to have an empty string match not at the start, even when the pattern is
1322 anchored. Updated pcretest and pcredemo to use this option.
1323
1324 21. If the maximum number of capturing subpatterns in a recursion was greater
1325 than the maximum at the outer level, the higher number was returned, but
1326 with unset values at the outer level. The correct (outer level) value is
1327 now given.
1328
1329 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1330 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1331 make it do so. The string so far is captured, making this feature
1332 compatible with Perl.
1333
1334 23. The tests have been re-organized, adding tests 11 and 12, to make it
1335 possible to check the Perl 5.10 features against Perl 5.10.
1336
1337 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1338 pattern matches a fixed length string. PCRE did not allow this; now it
1339 does. Neither allows recursion.
1340
1341 25. I finally figured out how to implement a request to provide the minimum
1342 length of subject string that was needed in order to match a given pattern.
1343 (It was back references and recursion that I had previously got hung up
1344 on.) This code has now been added to pcre_study(); it finds a lower bound
1345 to the length of subject needed. It is not necessarily the greatest lower
1346 bound, but using it to avoid searching strings that are too short does give
1347 some useful speed-ups. The value is available to calling programs via
1348 pcre_fullinfo().
1349
1350 26. While implementing 25, I discovered to my embarrassment that pcretest had
1351 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1352 study optimizations had never been tested with that matching function.
1353 Oops. What is worse, even when it was passed study data, there was a bug in
1354 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1355 were also very few tests of studied patterns with pcre_dfa_exec().
1356
1357 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1358 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1359 on the other side of the coin, they are no longer allowed to have different
1360 names, because these cannot be distinguished in PCRE, and this has caused
1361 confusion. (This is a difference from Perl.)
1362
1363 28. When duplicate subpattern names are present (necessarily with different
1364 numbers, as required by 27 above), and a test is made by name in a
1365 conditional pattern, either for a subpattern having been matched, or for
1366 recursion in such a pattern, all the associated numbered subpatterns are
1367 tested, and the overall condition is true if the condition is true for any
1368 one of them. This is the way Perl works, and is also more like the way
1369 testing by number works.
1370
1371
1372 Version 7.9 11-Apr-09
1373 ---------------------
1374
1375 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1376 (pcretest), all targets were linked against these libraries. This included
1377 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1378 libraries. This caused unwanted dependencies to be created. This problem
1379 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1380 pcretest is linked with readline.
1381
1382 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1383 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1384 moved outside it again, because FALSE and TRUE are already defined in AIX,
1385 but BOOL is not.
1386
1387 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1388 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1389
1390 4. The pcregrep documentation said spaces were inserted as well as colons (or
1391 hyphens) following file names and line numbers when outputting matching
1392 lines. This is not true; no spaces are inserted. I have also clarified the
1393 wording for the --colour (or --color) option.
1394
1395 5. In pcregrep, when --colour was used with -o, the list of matching strings
1396 was not coloured; this is different to GNU grep, so I have changed it to be
1397 the same.
1398
1399 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1400 each matching line was coloured. Now it goes on to look for further matches
1401 of any of the test patterns, which is the same behaviour as GNU grep.
1402
1403 7. A pattern that could match an empty string could cause pcregrep to loop; it
1404 doesn't make sense to accept an empty string match in pcregrep, so I have
1405 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1406 seems to be how GNU grep behaves.
1407
1408 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1409 start or after a newline", because the conditional assertion was not being
1410 correctly handled. The rule now is that both the assertion and what follows
1411 in the first alternative must satisfy the test.
1412
1413 9. If auto-callout was enabled in a pattern with a conditional group whose
1414 condition was an assertion, PCRE could crash during matching, both with
1415 pcre_exec() and pcre_dfa_exec().
1416
1417 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1418 used for matching.
1419
1420 11. Unicode property support in character classes was not working for
1421 characters (bytes) greater than 127 when not in UTF-8 mode.
1422
1423 12. Added the -M command line option to pcretest.
1424
1425 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1426
1427 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1428
1429 16. Added comments and documentation about mis-use of no_arg in the C++
1430 wrapper.
1431
1432 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1433 from Martin Jerabek that uses macro names for all relevant character and
1434 string constants.
1435
1436 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1437 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1438 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1439 these, but not everybody uses configure.
1440
1441 19. A conditional group that had only one branch was not being correctly
1442 recognized as an item that could match an empty string. This meant that an
1443 enclosing group might also not be so recognized, causing infinite looping
1444 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1445 with the subject "ab", where knowledge that the repeated group can match
1446 nothing is needed in order to break the loop.
1447
1448 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1449 exec(), but without supplying a callout function, matching went wrong.
1450
1451 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1452 leak if the size of the offset vector was greater than 30. When the vector
1453 is smaller, the saved offsets during recursion go onto a local stack
1454 vector, but for larger vectors malloc() is used. It was failing to free
1455 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1456 error, in fact).
1457
1458 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1459 heapframe that is used only when UTF-8 support is enabled. This caused no
1460 problem, but was untidy.
1461
1462 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1463 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1464 included within another project.
1465
1466 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1467 slightly modified by me:
1468
1469 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1470 not building pcregrep.
1471
1472 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1473 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1474
1475 25. Forward references, both numeric and by name, in patterns that made use of
1476 duplicate group numbers, could behave incorrectly or give incorrect errors,
1477 because when scanning forward to find the reference group, PCRE was not
1478 taking into account the duplicate group numbers. A pattern such as
1479 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1480
1481 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1482 the feature of a custom pointer more persuasive (as requested by a user).
1483
1484 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1485 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1486 pre-defined.
1487
1488 28. Added support for (*UTF8) at the start of a pattern.
1489
1490 29. Arrange for flags added by the "release type" setting in CMake to be shown
1491 in the configuration summary.
1492
1493
1494 Version 7.8 05-Sep-08
1495 ---------------------
1496
1497 1. Replaced UCP searching code with optimized version as implemented for Ad
1498 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1499 stage table and inline lookup instead of a function, giving speed ups of 2
1500 to 5 times on some simple patterns that I tested. Permission was given to
1501 distribute the MultiStage2.py script that generates the tables (it's not in
1502 the tarball, but is in the Subversion repository).
1503
1504 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1505 scripts.
1506
1507 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1508 a group with a zero qualifier. The result of the study could be incorrect,
1509 or the function might crash, depending on the pattern.
1510
1511 4. Caseless matching was not working for non-ASCII characters in back
1512 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1513 It now works when Unicode Property Support is available.
1514
1515 5. In pcretest, an escape such as \x{de} in the data was always generating
1516 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1517 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1518 truncation.
1519
1520 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1521
1522 7. Added two (int) casts to pcregrep when printing the difference of two
1523 pointers, in case they are 64-bit values.
1524
1525 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1526 test 2 if it fails.
1527
1528 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1529 and a #define of that name to empty if it is not externally set. This is to
1530 allow users of MSVC to set it if necessary.
1531
1532 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1533 the convenience functions in the pcre_get.c source file.
1534
1535 11. An option change at the start of a pattern that had top-level alternatives
1536 could cause overwriting and/or a crash. This command provoked a crash in
1537 some environments:
1538
1539 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1540
1541 This potential security problem was recorded as CVE-2008-2371.
1542
1543 12. For a pattern where the match had to start at the beginning or immediately
1544 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1545 pcre_dfa_exec() could read past the end of the passed subject if there was
1546 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1547 pcretest so that it places the subject at the end of its malloc-ed buffer.
1548
1549 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1550 exec() might read past the end of the data buffer in UTF-8 mode.
1551
1552 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1553 the data contained the byte 0x85 as part of a UTF-8 character within its
1554 first line. This applied both to normal and DFA matching.
1555
1556 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1557 /^[^d]*?$/8 failed to match "abc".
1558
1559 16. Added a missing copyright notice to pcrecpp_internal.h.
1560
1561 17. Make it more clear in the documentation that values returned from
1562 pcre_exec() in ovector are byte offsets, not character counts.
1563
1564 18. Tidied a few places to stop certain compilers from issuing warnings.
1565
1566 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1567 supplied by Stefan Weber. I made a further small update for 7.8 because
1568 there is a change of source arrangements: the pcre_searchfuncs.c module is
1569 replaced by pcre_ucd.c.
1570
1571
1572 Version 7.7 07-May-08
1573 ---------------------
1574
1575 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1576 a string to a long long, pretend we don't even have a long long." This is
1577 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1578
1579 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1580 pre-7.6 versions, which defined a global no_arg variable instead of putting
1581 it in the RE class. (See also #8 below.)
1582
1583 3. Remove a line of dead code, identified by coverity and reported by Nuno
1584 Lopes.
1585
1586 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1587
1588 (1) The include/exclude patterns were being applied to the whole pathnames
1589 of files, instead of just to the final components.
1590
1591 (2) If there was more than one level of directory, the subdirectories were
1592 skipped unless they satisfied the include/exclude conditions. This is
1593 inconsistent with GNU grep (and could even be seen as contrary to the
1594 pcregrep specification - which I improved to make it absolutely clear).
1595 The action now is always to scan all levels of directory, and just
1596 apply the include/exclude patterns to regular files.
1597
1598 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1599 --exclude_dir in the tests to avoid scanning .svn directories.
1600
1601 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1602 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1603 doesn't support NULs in patterns.
1604
1605 7. Added some missing "const"s to declarations of static tables in
1606 pcre_compile.c and pcre_dfa_exec.c.
1607
1608 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1609 caused by fix #2 above. (Subsequently also a second patch to fix the
1610 first patch. And a third patch - this was a messy problem.)
1611
1612 9. Applied Craig's patch to remove the use of push_back().
1613
1614 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1615 matching function regexec().
1616
1617 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1618 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1619 references. PCRE supports relative numbers with this syntax (I don't think
1620 Oniguruma does).
1621
1622 12. Previously, a group with a zero repeat such as (...){0} was completely
1623 omitted from the compiled regex. However, this means that if the group
1624 was called as a subroutine from elsewhere in the pattern, things went wrong
1625 (an internal error was given). Such groups are now left in the compiled
1626 pattern, with a new opcode that causes them to be skipped at execution
1627 time.
1628
1629 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1630 to the way PCRE behaves:
1631
1632 (a) A lone ] character is dis-allowed (Perl treats it as data).
1633
1634 (b) A back reference to an unmatched subpattern matches an empty string
1635 (Perl fails the current match path).
1636
1637 (c) A data ] in a character class must be notated as \] because if the
1638 first data character in a class is ], it defines an empty class. (In
1639 Perl it is not possible to have an empty class.) The empty class []
1640 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1641 The negative empty class [^] matches any one character, independently
1642 of the DOTALL setting.
1643
1644 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1645 non-existent subpattern following a character class starting with ']' and
1646 containing () gave an internal compiling error instead of "reference to
1647 non-existent subpattern". Fortunately, when the pattern did exist, the
1648 compiled code was correct. (When scanning forwards to check for the
1649 existencd of the subpattern, it was treating the data ']' as terminating
1650 the class, so got the count wrong. When actually compiling, the reference
1651 was subsequently set up correctly.)
1652
1653 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1654 it was being rejected as not supported by pcre_dfa_exec(), even though
1655 other assertions are supported. I have made pcre_dfa_exec() support
1656 (*FAIL).
1657
1658 16. The implementation of 13c above involved the invention of a new opcode,
1659 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1660 cannot be changed at match time, I realized I could make a small
1661 improvement to matching performance by compiling OP_ALLANY instead of
1662 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1663 on the OP_ANY path.
1664
1665 17. Compiling pcretest on Windows with readline support failed without the
1666 following two fixes: (1) Make the unistd.h include conditional on
1667 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1668
1669 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1670 ncurses library to be included for pcretest when ReadLine support is
1671 requested, but also to allow for it to be overridden. This patch came from
1672 Daniel Bergström.
1673
1674 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1675 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1676 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1677 spotting this.
1678
1679
1680 Version 7.6 28-Jan-08
1681 ---------------------
1682
1683 1. A character class containing a very large number of characters with
1684 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1685 overflow.
1686
1687 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1688 HAVE_LONG_LONG is not defined.
1689
1690 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1691 bring them up to date and include new features. This patch includes:
1692
1693 - Fixed PH's badly added libz and libbz2 support.
1694 - Fixed a problem with static linking.
1695 - Added pcredemo. [But later removed - see 7 below.]
1696 - Fixed dftables problem and added an option.
1697 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1698 HAVE_LONG_LONG.
1699 - Added readline support for pcretest.
1700 - Added an listing of the option settings after cmake has run.
1701
1702 4. A user submitted a patch to Makefile that makes it easy to create
1703 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1704 Makefile.am that cause it to include this special target, without
1705 affecting anything else. Note that the same mingw target plus all
1706 the other distribution libraries and programs are now supported
1707 when configuring with CMake (see 6 below) instead of with
1708 Configure/Make.
1709
1710 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1711 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1712 exported in the Windows port". It has not yet been confirmed that the patch
1713 solves the problem, but it does no harm.
1714
1715 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1716 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1717 with CMake, and also correct the comment about stack recursion.
1718
1719 7. Remove the automatic building of pcredemo from the ./configure system and
1720 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1721 of a program that users should build themselves after PCRE is installed, so
1722 building it automatically is not really right. What is more, it gave
1723 trouble in some build environments.
1724
1725 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1726
1727
1728 Version 7.5 10-Jan-08
1729 ---------------------
1730
1731 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1732 values in parens when parsing an RE using the C++ wrapper."
1733
1734 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1735 Characters greater than 255 were excluded from the class instead of being
1736 included.
1737
1738 3. The same bug as (2) above applied to negated POSIX classes such as
1739 [:^space:].
1740
1741 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1742 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1743 I have changed it.
1744
1745 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1746 first named subpattern) and a construct such as (?&a) would reference the
1747 first named subpattern whose name started with "a" (in other words, the
1748 length check was missing). Both these problems are fixed. "Subpattern name
1749 expected" is now given for (?&) (a zero-length name), and this patch also
1750 makes it give the same error for \k'' (previously it complained that that
1751 was a reference to a non-existent subpattern).
1752
1753 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1754 this is right because (?- can be followed by option settings as well as by
1755 digits. I have, however, made the messages clearer.
1756
1757 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1758 than the number used in the conditional) now cause a compile-time error.
1759 This is actually not compatible with Perl, which accepts such patterns, but
1760 treats the conditional as always being FALSE (as PCRE used to), but it
1761 seems to me that giving a diagnostic is better.
1762
1763 8. Change "alphameric" to the more common word "alphanumeric" in comments
1764 and messages.
1765
1766 9. Fix two occurrences of "backslash" in comments that should have been
1767 "backspace".
1768
1769 10. Remove two redundant lines of code that can never be obeyed (their function
1770 was moved elsewhere).
1771
1772 11. The program that makes PCRE's Unicode character property table had a bug
1773 which caused it to generate incorrect table entries for sequences of
1774 characters that have the same character type, but are in different scripts.
1775 It amalgamated them into a single range, with the script of the first of
1776 them. In other words, some characters were in the wrong script. There were
1777 thirteen such cases, affecting characters in the following ranges:
1778
1779 U+002b0 - U+002c1
1780 U+0060c - U+0060d
1781 U+0061e - U+00612
1782 U+0064b - U+0065e
1783 U+0074d - U+0076d
1784 U+01800 - U+01805
1785 U+01d00 - U+01d77
1786 U+01d9b - U+01dbf
1787 U+0200b - U+0200f
1788 U+030fc - U+030fe
1789 U+03260 - U+0327f
1790 U+0fb46 - U+0fbb1
1791 U+10450 - U+1049d
1792
1793 12. The -o option (show only the matching part of a line) for pcregrep was not
1794 compatible with GNU grep in that, if there was more than one match in a
1795 line, it showed only the first of them. It now behaves in the same way as
1796 GNU grep.
1797
1798 13. If the -o and -v options were combined for pcregrep, it printed a blank
1799 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1800 does the same. The return code can be used to tell if there were any
1801 non-matching lines.
1802
1803 14. Added --file-offsets and --line-offsets to pcregrep.
1804
1805 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1806 infinitely looping recursion. The bug was that positive lookaheads were not
1807 being skipped when checking for a possible empty match (negative lookaheads
1808 and both kinds of lookbehind were skipped).
1809
1810 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1811 inclusion of <windows.h> to before rather than after the definition of
1812 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1813
1814 17. Specifying a possessive quantifier with a specific limit for a Unicode
1815 character property caused pcre_compile() to compile bad code, which led at
1816 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1817 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1818 caused the error; without that there was no problem.
1819
1820 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1821
1822 19. Added --enable-pcretest-libreadline.
1823
1824 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1825 RE::GlobalReplace(). As a result, the number of replacements returned was
1826 double what it should be. I removed one of the increments, but Craig sent a
1827 later patch that removed the other one (the right fix) and added unit tests
1828 that check the return values (which was not done before).
1829
1830 21. Several CMake things:
1831
1832 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1833 the names libpcre and libpcreposix, not just pcre and pcreposix.
1834
1835 (2) The above change means that pcretest and pcregrep are now correctly
1836 linked with the newly-built libraries, not previously installed ones.
1837
1838 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1839
1840 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1841 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1842 UTF-8 newline character). The key issue is that the pattern starts .*;
1843 this means that the match must be either at the beginning, or after a
1844 newline. The bug was in the code for advancing after a failed match and
1845 checking that the new position followed a newline. It was not taking
1846 account of UTF-8 characters correctly.
1847
1848 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1849 character classes. PCRE was not treating the sequence [:...:] as a
1850 character class unless the ... were all letters. Perl, however, seems to
1851 allow any characters between [: and :], though of course it rejects as
1852 unknown any "names" that contain non-letters, because all the known class
1853 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1854 for example, whereas PCRE did not - it did not recognize a POSIX character
1855 class. This seemed a bit dangerous, so the code has been changed to be
1856 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1857 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1858 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1859 Perl does, and where it didn't before.
1860
1861 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1862 Windows environments %n is disabled by default.
1863
1864
1865 Version 7.4 21-Sep-07
1866 ---------------------
1867
1868 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1869 means that a class such as [\s] counted as "explicit reference to CR or
1870 LF". That isn't really right - the whole point of the change was to try to
1871 help when there was an actual mention of one of the two characters. So now
1872 the change happens only if \r or \n (or a literal CR or LF) character is
1873 encountered.
1874
1875 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1876 of both had grown to the point where there were only 3 bits left.
1877 Fortunately, there was spare space in the data structure, and so I have
1878 moved the internal flags into a new 16-bit field to free up more option
1879 bits.
1880
1881 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1882 but did not set the internal JCHANGED flag - either of these is enough to
1883 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1884 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1885 start sets both bits.
1886
1887 4. Added options (at build time, compile time, exec time) to change \R from
1888 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1889
1890 5. doc/pcresyntax.html was missing from the distribution.
1891
1892 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1893 compatibility, even though it is no longer used.
1894
1895 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1896 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1897 windows.h file is present (where different names are used). [This was
1898 reversed later after testing - see 16 below.]
1899
1900 8. Changed all #include <config.h> to #include "config.h". There were also
1901 some further <pcre.h> cases that I changed to "pcre.h".
1902
1903 9. When pcregrep was used with the --colour option, it missed the line ending
1904 sequence off the lines that it output.
1905
1906 10. It was pointed out to me that arrays of string pointers cause lots of
1907 relocations when a shared library is dynamically loaded. A technique of
1908 using a single long string with a table of offsets can drastically reduce
1909 these. I have refactored PCRE in four places to do this. The result is
1910 dramatic:
1911
1912 Originally: 290
1913 After changing UCP table: 187
1914 After changing error message table: 43
1915 After changing table of "verbs" 36
1916 After changing table of Posix names 22
1917
1918 Thanks to the folks working on Gregex for glib for this insight.
1919
1920 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1921 unicode-properties was also set.
1922
1923 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1924
1925 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1926 checked only for CRLF.
1927
1928 14. Added casts to pcretest.c to avoid compiler warnings.
1929
1930 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1931
1932 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1933 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1934 entirely. This removes changes made in 7 above.
1935
1936 17. The CMake files have been updated, and there is now more information about
1937 building with CMake in the NON-UNIX-USE document.
1938
1939
1940 Version 7.3 28-Aug-07
1941 ---------------------
1942
1943 1. In the rejigging of the build system that eventually resulted in 7.1, the
1944 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1945 brackets there is not right, since it causes compilers to look for an
1946 installed pcre.h, not the version that is in the source that is being
1947 compiled (which of course may be different). I have changed it back to:
1948
1949 #include "pcre.h"
1950
1951 I have a vague recollection that the change was concerned with compiling in
1952 different directories, but in the new build system, that is taken care of
1953 by the VPATH setting the Makefile.
1954
1955 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1956 when the subject happened to end in the byte 0x85 (e.g. if the last
1957 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1958 characters but of course it shouldn't be taken as a newline when it is part
1959 of another character. The bug was that, for an unlimited repeat of . in
1960 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1961 characters when looking for a newline.
1962
1963 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1964
1965 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1966 in debug output.
1967
1968 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1969 long printing in the pcrecpp unittest when running under MinGW.
1970
1971 6. ESC_K was left out of the EBCDIC table.
1972
1973 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1974 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1975 limit also applies to "virtual nesting" when a pattern is recursive, and in
1976 this case 1000 isn't so big. I have been able to remove this limit at the
1977 expense of backing off one optimization in certain circumstances. Normally,
1978 when pcre_exec() would call its internal match() function recursively and
1979 immediately return the result unconditionally, it uses a "tail recursion"
1980 feature to save stack. However, when a subpattern that can match an empty
1981 string has an unlimited repetition quantifier, it no longer makes this
1982 optimization. That gives it a stack frame in which to save the data for
1983 checking that an empty string has been matched. Previously this was taken
1984 from the 1000-entry workspace that had been reserved. So now there is no
1985 explicit limit, but more stack is used.
1986
1987 8. Applied Daniel's patches to solve problems with the import/export magic
1988 syntax that is required for Windows, and which was going wrong for the
1989 pcreposix and pcrecpp parts of the library. These were overlooked when this
1990 problem was solved for the main library.
1991
1992 9. There were some crude static tests to avoid integer overflow when computing
1993 the size of patterns that contain repeated groups with explicit upper
1994 limits. As the maximum quantifier is 65535, the maximum group length was
1995 set at 30,000 so that the product of these two numbers did not overflow a
1996 32-bit integer. However, it turns out that people want to use groups that
1997 are longer than 30,000 bytes (though not repeat them that many times).
1998 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1999 made it possible to implement the integer overflow checks in a much more
2000 dynamic way, which I have now done. The artificial limitation on group
2001 length has been removed - we now have only the limit on the total length of
2002 the compiled pattern, which depends on the LINK_SIZE setting.
2003
2004 10. Fixed a bug in the documentation for get/copy named substring when
2005 duplicate names are permitted. If none of the named substrings are set, the
2006 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2007 empty string.
2008
2009 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2010 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2011 because the ] is interpreted as the first data character and the
2012 terminating ] is not found. PCRE has been made compatible with Perl in this
2013 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2014 cause memory overwriting.
2015
2016 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2017 string has been matched (to stop an infinite loop). It was not recognizing
2018 a conditional subpattern that could match an empty string if that
2019 subpattern was within another subpattern. For example, it looped when
2020 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2021 condition was not nested. This bug has been fixed.
2022
2023 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2024 past the start of the subject in the presence of bytes with the top bit
2025 set, for example "\x8aBCD".
2026
2027 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2028 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2029
2030 14. Optimized (?!) to (*FAIL).
2031
2032 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2033 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2034 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2035 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2036 does: it's just the validity check that is more restrictive.
2037
2038 16. Inserted checks for integer overflows during escape sequence (backslash)
2039 processing, and also fixed erroneous offset values for syntax errors during
2040 backslash processing.
2041
2042 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2043 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2044
2045 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2046 caused an overrun.
2047
2048 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2049 something other than just ASCII characters) inside a group that had an
2050 unlimited repeat caused a loop at compile time (while checking to see
2051 whether the group could match an empty string).
2052
2053 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2054 [\P{Any}] did so. (Error in the code for printing property names.)
2055
2056 21. An orphan \E inside a character class could cause a crash.
2057
2058 22. A repeated capturing bracket such as (A)? could cause a wild memory
2059 reference during compilation.
2060
2061 23. There are several functions in pcre_compile() that scan along a compiled
2062 expression for various reasons (e.g. to see if it's fixed length for look
2063 behind). There were bugs in these functions when a repeated \p or \P was
2064 present in the pattern. These operators have additional parameters compared
2065 with \d, etc, and these were not being taken into account when moving along
2066 the compiled data. Specifically:
2067
2068 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2069 length.
2070
2071 (b) An item such as \pL+ within a repeated group could cause crashes or
2072 loops.
2073
2074 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2075 "reference to non-existent subpattern" error.
2076
2077 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2078
2079 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2080 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2081
2082 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2083
2084 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2085 character were causing crashes (broken optimization).
2086
2087 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2088 \p or \P) caused a compile-time loop.
2089
2090 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2091 break. For example, the unstudied pattern [\r\n]A does not match the string
2092 "\r\nA" because change 7.0/46 below moves the current point on by two
2093 characters after failing to match at the start. However, the pattern \nA
2094 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2095 the same is true. There doesn't seem any very clean way out of this, but
2096 what I have chosen to do makes the common cases work: PCRE now takes note
2097 of whether there can be an explicit match for \r or \n anywhere in the
2098 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2099 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2100 pattern has explicit CR or LF references.
2101
2102 29. Added (*CR) etc for changing newline setting at start of pattern.
2103
2104
2105 Version 7.2 19-Jun-07
2106 ---------------------
2107
2108 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2109 which is apparently normally available under Windows.
2110
2111 2. Re-jig the pcregrep tests with different newline settings in an attempt
2112 to make them independent of the local environment's newline setting.
2113
2114 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2115
2116 4. Some of the "internals" tests were previously cut out when the link size
2117 was not 2, because the output contained actual offsets. The recent new
2118 "Z" feature of pcretest means that these can be cut out, making the tests
2119 usable with all link sizes.
2120
2121 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2122 stack recursion. This gives a massive performance boost under BSD, but just
2123 a small improvement under Linux. However, it saves one field in the frame
2124 in all cases.
2125
2126 6. Added more features from the forthcoming Perl 5.10:
2127
2128 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2129 recursion call. It refers to the nth most recently opened parentheses.
2130
2131 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2132 to be opened parentheses.
2133
2134 (c) Conditions that refer to capturing parentheses can be specified
2135 relatively, for example, (?(-2)... or (?(+3)...
2136
2137 (d) \K resets the start of the current match so that everything before
2138 is not part of it.
2139
2140 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2141
2142 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2143 reference syntax.
2144
2145 (g) (?| introduces a group in which the numbering of parentheses in each
2146 alternative starts with the same number.
2147
2148 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2149
2150 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2151 PCRE_INFO_JCHANGED.
2152
2153 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2154 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2155 for detecting groups that can match an empty string.
2156
2157 9. A pattern with a very large number of alternatives (more than several
2158 hundred) was running out of internal workspace during the pre-compile
2159 phase, where pcre_compile() figures out how much memory will be needed. A
2160 bit of new cunning has reduced the workspace needed for groups with
2161 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2162 workspace instead of running out of the 4096 that are available.
2163
2164 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2165
2166 11. Applied patch from Google to remove an optimization that didn't quite work.
2167 The report of the bug said:
2168
2169 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2170 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2171 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2172
2173 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2174 it matched the wrong number of bytes.
2175
2176
2177 Version 7.1 24-Apr-07
2178 ---------------------
2179
2180 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2181 that is more "standard", making use of automake and other Autotools. There
2182 is some re-arrangement of the files and adjustment of comments consequent
2183 on this.
2184
2185 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2186 for recursive directory scanning broke on some systems because the files
2187 are not scanned in any specific order and on different systems the order
2188 was different. A call to "sort" has been inserted into RunGrepTest for the
2189 approprate test as a short-term fix. In the longer term there may be an
2190 alternative.
2191
2192 3. I had an email from Eric Raymond about problems translating some of PCRE's
2193 man pages to HTML (despite the fact that I distribute HTML pages, some
2194 people do their own conversions for various reasons). The problems
2195 concerned the use of low-level troff macros .br and .in. I have therefore
2196 removed all such uses from the man pages (some were redundant, some could
2197 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2198 HTML has been updated to handle .nf/.fi and to complain if it encounters
2199 .br or .in.
2200
2201 4. Updated comments in configure.ac that get placed in config.h.in and also
2202 arranged for config.h to be included in the distribution, with the name
2203 config.h.generic, for the benefit of those who have to compile without
2204 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2205
2206 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2207 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2208 makevp.bat for the current PCRE, using the additional files
2209 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2210
2211 6. A Windows user reported a minor discrepancy with test 2, which turned out
2212 to be caused by a trailing space on an input line that had got lost in his
2213 copy. The trailing space was an accident, so I've just removed it.
2214
2215 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2216 that is needed.
2217
2218 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2219 as "const" (a) because they are and (b) because it helps the PHP
2220 maintainers who have recently made a script to detect big data structures
2221 in the php code that should be moved to the .rodata section. I remembered
2222 to update Builducptable as well, so it won't revert if ucptable.h is ever
2223 re-created.
2224
2225 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2226 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2227 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2228 support is not required. This saves 1.5-2K of code, which is important in
2229 some applications.
2230
2231 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2232 so as not to refer to the tables, even though these functions will never be
2233 called when UTF-8 support is disabled. Otherwise there are problems with a
2234 shared library.
2235
2236 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2237
2238 (a) It was defining its arguments as char * instead of void *.
2239
2240 (b) It was assuming that all moves were upwards in memory; this was true
2241 a long time ago when I wrote it, but is no longer the case.
2242
2243 The emulated memove() is provided for those environments that have neither
2244 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2245 is clearly not the case, as these two bugs were recently reported.
2246
2247 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2248 and Detrail to create the HTML documentation, the .txt form of the man
2249 pages, and it removes trailing spaces from listed files. It also creates
2250 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2251 case, it wraps all the #defines with #ifndefs. This script should be run
2252 before "make dist".
2253
2254 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2255 with Unicode property support.
2256
2257 (a) For a maximizing quantifier, if the two different cases of the
2258 character were of different lengths in their UTF-8 codings (there are
2259 some cases like this - I found 11), and the matching function had to
2260 back up over a mixture of the two cases, it incorrectly assumed they
2261 were both the same length.
2262
2263 (b) When PCRE was configured to use the heap rather than the stack for
2264 recursion during matching, it was not correctly preserving the data for
2265 the other case of a UTF-8 character when checking ahead for a match
2266 while processing a minimizing repeat. If the check also involved
2267 matching a wide character, but failed, corruption could cause an
2268 erroneous result when trying to check for a repeat of the original
2269 character.
2270
2271 13. Some tidying changes to the testing mechanism:
2272
2273 (a) The RunTest script now detects the internal link size and whether there
2274 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2275 values substituted by "configure". (The RunGrepTest script already did
2276 this for UTF-8.) The configure.ac script no longer substitutes the
2277 relevant variables.
2278
2279 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2280 with length and offset values. This means that the output is different
2281 for different internal link sizes. Test 2 is skipped for link sizes
2282 other than 2 because of this, bypassing the problem. Unfortunately,
2283 there was also a test in test 3 (the locale tests) that used /B and
2284 failed for link sizes other than 2. Rather than cut the whole test out,
2285 I have added a new /Z option to pcretest that replaces the length and
2286 offset values with spaces. This is now used to make test 3 independent
2287 of link size. (Test 2 will be tidied up later.)
2288
2289 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2290 segmentation fault instead of returning the appropriate error message.
2291
2292 15. In multiline mode when the newline sequence was set to "any", the pattern
2293 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2294 This doesn't seem right; it now treats the CRLF combination as the line
2295 ending, and so does not match in that case. It's only a pattern such as ^$
2296 that would hit this one: something like ^ABC$ would have failed after \r
2297 and then tried again after \r\n.
2298
2299 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2300 in an attempt to make files that differ only in their line terminators
2301 compare equal. This works on Linux.
2302
2303 17. Under certain error circumstances pcregrep might try to free random memory
2304 as it exited. This is now fixed, thanks to valgrind.
2305
2306 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2307 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2308 was because its rules for how to advance for /g after matching an empty
2309 string at the end of a line did not allow for this case. They now check for
2310 it specially.
2311
2312 20. pcretest is supposed to handle patterns and data of any length, by
2313 extending its buffers when necessary. It was getting this wrong when the
2314 buffer for a data line had to be extended.
2315
2316 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2317 CRLF as a newline sequence.
2318
2319 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2320 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2321 I have nevertheless tidied it up.
2322
2323 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2324
2325 24. Added a man page for pcre-config.
2326
2327
2328 Version 7.0 19-Dec-06
2329 ---------------------
2330
2331 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2332 moving to gcc 4.1.1.
2333
2334 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2335 sys/time.h, which is documented as needed for this function. It doesn't
2336 seem to matter on Linux, but it showed up on some releases of OS X.
2337
2338 3. It seems that there are systems where bytes whose values are greater than
2339 127 match isprint() in the "C" locale. The "C" locale should be the
2340 default when a C program starts up. In most systems, only ASCII printing
2341 characters match isprint(). This difference caused the output from pcretest
2342 to vary, making some of the tests fail. I have changed pcretest so that:
2343
2344 (a) When it is outputting text in the compiled version of a pattern, bytes
2345 other than 32-126 are always shown as hex escapes.
2346
2347 (b) When it is outputting text that is a matched part of a subject string,
2348 it does the same, unless a different locale has been set for the match
2349 (using the /L modifier). In this case, it uses isprint() to decide.
2350
2351 4. Fixed a major bug that caused incorrect computation of the amount of memory
2352 required for a compiled pattern when options that changed within the
2353 pattern affected the logic of the preliminary scan that determines the
2354 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2355 that the computed length was too small. The symptoms of this bug were
2356 either the PCRE error "internal error: code overflow" from pcre_compile(),
2357 or a glibc crash with a message such as "pcretest: free(): invalid next
2358 size (fast)". Examples of patterns that provoked this bug (shown in
2359 pcretest format) are:
2360
2361 /(?-x: )/x
2362 /(?x)(?-x: \s*#\s*)/
2363 /((?i)[\x{c0}])/8
2364 /(?i:[\x{c0}])/8
2365
2366 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2367 is now done differently.
2368
2369 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2370 wrapper classes; (b) implement a new function in the C++ scanner that is
2371 more efficient than the old way of doing things because it avoids levels of
2372 recursion in the regex matching; (c) add a paragraph to the documentation
2373 for the FullMatch() function.
2374
2375 6. The escape sequence \n was being treated as whatever was defined as
2376 "newline". Not only was this contrary to the documentation, which states
2377 that \n is character 10 (hex 0A), but it also went horribly wrong when
2378 "newline" was defined as CRLF. This has been fixed.
2379
2380 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2381 was being set to -1 for the "end of line" case (supposedly a value that no
2382 character can have). Though this value is never used (the check for end of
2383 line is "zero bytes in current character"), it caused compiler complaints.
2384 I've changed it to 0xffffffff.
2385
2386 8. In pcre_version.c, the version string was being built by a sequence of
2387 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2388 string (as it is for production releases) called a macro with an empty
2389 argument. The C standard says the result of this is undefined. The gcc
2390 compiler treats it as an empty string (which was what was wanted) but it is
2391 reported that Visual C gives an error. The source has been hacked around to
2392 avoid this problem.
2393
2394 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2395 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2396 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2397 of them did).
2398
2399 10. Originally, pcretest opened its input and output without "b"; then I was
2400 told that "b" was needed in some environments, so it was added for release
2401 5.0 to both the input and output. (It makes no difference on Unix-like
2402 systems.) Later I was told that it is wrong for the input on Windows. I've
2403 now abstracted the modes into two macros, to make it easier to fiddle with
2404 them, and removed "b" from the input mode under Windows.
2405
2406 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2407
2408 12. Added -help and --help to pcretest as an official way of being reminded
2409 of the options.
2410
2411 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2412 and pcrecpp.cc because they annoy compilers at high warning levels.
2413
2414 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2415
2416 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2417 scripts are not C programs :-) and which was not noticed because it works
2418 on Linux.
2419
2420 16. pcretest is supposed to handle any length of pattern and data line (as one
2421 line or as a continued sequence of lines) by extending its input buffer if
2422 necessary. This feature was broken for very long pattern lines, leading to
2423 a string of junk being passed to pcre_compile() if the pattern was longer
2424 than about 50K.
2425
2426 17. I have done a major re-factoring of the way pcre_compile() computes the
2427 amount of memory needed for a compiled pattern. Previously, there was code
2428 that made a preliminary scan of the pattern in order to do this. That was
2429 OK when PCRE was new, but as the facilities have expanded, it has become
2430 harder and harder to keep it in step with the real compile phase, and there
2431 have been a number of bugs (see for example, 4 above). I have now found a
2432 cunning way of running the real compile function in a "fake" mode that
2433 enables it to compute how much memory it would need, while actually only
2434 ever using a few hundred bytes of working memory and without too many
2435 tests of the mode. This should make future maintenance and development
2436 easier. A side effect of this work is that the limit of 200 on the nesting
2437 depth of parentheses has been removed (though this was never a serious
2438 limitation, I suspect). However, there is a downside: pcre_compile() now
2439 runs more slowly than before (30% or more, depending on the pattern). I
2440 hope this isn't a big issue. There is no effect on runtime performance.
2441
2442 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2443 newline (only possible for the last line of a file) and it was a
2444 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2445
2446 19. Added additional timing features to pcretest. (1) The -tm option now times
2447 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2448 separate command line item, by a number that specifies the number of
2449 repeats to use when timing. The default is 50000; this gives better
2450 precision, but takes uncomfortably long for very large patterns.
2451
2452 20. Extended pcre_study() to be more clever in cases where a branch of a
2453 subpattern has no definite first character. For example, (a*|b*)[cd] would
2454 previously give no result from pcre_study(). Now it recognizes that the
2455 first character must be a, b, c, or d.
2456
2457 21. There was an incorrect error "recursive call could loop indefinitely" if
2458 a subpattern (or the entire pattern) that was being tested for matching an
2459 empty string contained only one non-empty item after a nested subpattern.
2460 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2461 incorrectly, because the \d was being skipped in the check.
2462
2463 22. The pcretest program now has a new pattern option /B and a command line
2464 option -b, which is equivalent to adding /B to every pattern. This causes
2465 it to show the compiled bytecode, without the additional information that
2466 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2467 is the same as /B/I).
2468
2469 23. A new optimization is now able automatically to treat some sequences such
2470 as a*b as a*+b. More specifically, if something simple (such as a character
2471 or a simple class like \d) has an unlimited quantifier, and is followed by
2472 something that cannot possibly match the quantified thing, the quantifier
2473 is automatically "possessified".
2474
2475 24. A recursive reference to a subpattern whose number was greater than 39
2476 went wrong under certain circumstances in UTF-8 mode. This bug could also
2477 have affected the operation of pcre_study().
2478
2479 25. Realized that a little bit of performance could be had by replacing
2480 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2481
2482 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2483
2484 27. Possessive quantifiers such as a++ were previously implemented by turning
2485 them into atomic groups such as ($>a+). Now they have their own opcodes,
2486 which improves performance. This includes the automatically created ones
2487 from 23 above.
2488
2489 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2490 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2491 the first matched character to be a colon. This applied both to named and
2492 numbered groups.
2493
2494 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2495
2496 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2497 building PCRE on Windows easier, so I have included it in the distribution.
2498
2499 31. There is now a check in pcretest against a ridiculously large number being
2500 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2501 loop, the loop is abandoned.
2502
2503 32. Forward references to subpatterns in conditions such as (?(2)...) where
2504 subpattern 2 is defined later cause pcre_compile() to search forwards in
2505 the pattern for the relevant set of parentheses. This search went wrong
2506 when there were unescaped parentheses in a character class, parentheses
2507 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2508
2509 33. "Subroutine" calls and backreferences were previously restricted to
2510 referencing subpatterns earlier in the regex. This restriction has now
2511 been removed.
2512
2513 34. Added a number of extra features that are going to be in Perl 5.10. On the
2514 whole, these are just syntactic alternatives for features that PCRE had
2515 previously implemented using the Python syntax or my own invention. The
2516 other formats are all retained for compatibility.
2517
2518 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2519 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2520 also .NET compatible.
2521
2522 (b) A recursion or subroutine call to a named group can now be defined as
2523 (?&name) as well as (?P>name).
2524
2525 (c) A backreference to a named group can now be defined as \k<name> or
2526 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2527 5.10, are also .NET compatible.
2528
2529 (d) A conditional reference to a named group can now use the syntax
2530 (?(<name>) or (?('name') as well as (?(name).
2531
2532 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2533 groups (named and numbered) that are never evaluated inline, but can be
2534 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2535 is always false. There may be only one alternative in such a group.
2536
2537 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2538 as the simple (?(R). The condition is true only if the most recent
2539 recursion is that of the given number or name. It does not search out
2540 through the entire recursion stack.
2541
2542 (g) The escape \gN or \g{N} has been added, where N is a positive or
2543 negative number, specifying an absolute or relative reference.
2544
2545 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2546 some "unreachable code" warnings.
2547
2548 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2549 things, this adds five new scripts.
2550
2551 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2552 There were also incompatibilities regarding the handling of \Q..\E inside
2553 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2554 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2555
2556 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2557 matches an empty string, and forcibly breaks the loop. There were bugs in
2558 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2559 against aaaa the result was just "a" rather than "aaaa", for example. Two
2560 separate and independent bugs (that affected different cases) have been
2561 fixed.
2562
2563 39. Refactored the code to abolish the use of different opcodes for small
2564 capturing bracket numbers. This is a tidy that I avoided doing when I
2565 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2566 The new approach is not only tidier, it makes it possible to reduce the
2567 memory needed to fix the previous bug (38).
2568
2569 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2570 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2571 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2572 mode.
2573
2574 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2575 report.
2576
2577 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2578 copy construction and assignment in the C++ wrapper.
2579
2580 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2581 couple of bugs that could have given wrong results in the "--newline=crlf"
2582 case.
2583
2584 44. Added a number of casts and did some reorganization of signed/unsigned int
2585 variables following suggestions from Dair Grant. Also renamed the variable
2586 "this" as "item" because it is a C++ keyword.
2587
2588 45. Arranged for dftables to add
2589
2590 #include "pcre_internal.h"
2591
2592 to pcre_chartables.c because without it, gcc 4.x may remove the array
2593 definition from the final binary if PCRE is built into a static library and
2594 dead code stripping is activated.
2595
2596 46. For an unanchored pattern, if a match attempt fails at the start of a
2597 newline sequence, and the newline setting is CRLF or ANY, and the next two
2598 characters are CRLF, advance by two characters instead of one.
2599
2600
2601 Version 6.7 04-Jul-06
2602 ---------------------
2603
2604 1. In order to handle tests when input lines are enormously long, pcretest has
2605 been re-factored so that it automatically extends its buffers when
2606 necessary. The code is crude, but this _is_ just a test program. The
2607 default size has been increased from 32K to 50K.
2608
2609 2. The code in pcre_study() was using the value of the re argument before
2610 testing it for NULL. (Of course, in any sensible call of the function, it
2611 won't be NULL.)
2612
2613 3. The memmove() emulation function in pcre_internal.h, which is used on
2614 systems that lack both memmove() and bcopy() - that is, hardly ever -
2615 was missing a "static" storage class specifier.
2616
2617 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2618 containing an extended class (one that cannot be represented by a bitmap
2619 because it contains high-valued characters or Unicode property items, e.g.
2620 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2621 pattern, but PCRE should not loop if you do not (it no longer does).
2622 [Detail: two cases were found: (a) a repeated subpattern containing an
2623 extended class; (b) a recursive reference to a subpattern that followed a
2624 previous extended class. It wasn't skipping over the extended class
2625 correctly when UTF-8 mode was not set.]
2626
2627 5. A negated single-character class was not being recognized as fixed-length
2628 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2629 compile error "lookbehind assertion is not fixed length".
2630
2631 6. The RunPerlTest auxiliary script was showing an unexpected difference
2632 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2633 write a Perl script that can interpret lines of an input file either as
2634 byte characters or as UTF-8, which is what "perltest" was being required to
2635 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2636 can't do is switch easily at run time between having the "use utf8;" pragma
2637 or not. In the end, I fudged it by using the RunPerlTest script to insert
2638 "use utf8;" explicitly for the UTF-8 tests.
2639
2640 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2641 the end of the subject string, contrary to the documentation and to what
2642 Perl does. This was true of both matching functions. Now it matches only at
2643 the start of the subject and immediately after *internal* newlines.
2644
2645 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2646 a pointer to an int instead of a pointer to an unsigned long int. This
2647 caused problems on 64-bit systems.
2648
2649 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2650 instance of the 'standard' template library not being so standard".
2651
2652 10. There was no check on the number of named subpatterns nor the maximum
2653 length of a subpattern name. The product of these values is used to compute
2654 the size of the memory block for a compiled pattern. By supplying a very
2655 long subpattern name and a large number of named subpatterns, the size
2656 computation could be caused to overflow. This is now prevented by limiting
2657 the length of names to 32 characters, and the number of named subpatterns
2658 to 10,000.
2659
2660 11. Subpatterns that are repeated with specific counts have to be replicated in
2661 the compiled pattern. The size of memory for this was computed from the
2662 length of the subpattern and the repeat count. The latter is limited to
2663 65535, but there was no limit on the former, meaning that integer overflow
2664 could in principle occur. The compiled length of a repeated subpattern is
2665 now limited to 30,000 bytes in order to prevent this.
2666
2667 12. Added the optional facility to have named substrings with the same name.
2668
2669 13. Added the ability to use a named substring as a condition, using the
2670 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2671 are numbers (not recommended). Forward references are permitted.
2672
2673 14. Added forward references in named backreferences (if you see what I mean).
2674
2675 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2676 pattern could run off the end of the subject. For example, the pattern
2677 "(?s)(.{1,5})"8 did this with the subject "ab".
2678
2679 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2680 PCRE_CASELESS was set when matching characters that were quantified with ?
2681 or *.
2682
2683 17. A character class other than a single negated character that had a minimum
2684 but no maximum quantifier - for example [ab]{6,} - was not handled
2685 correctly by pce_dfa_exec(). It would match only one character.
2686
2687 18. A valid (though odd) pattern that looked like a POSIX character
2688 class but used an invalid character after [ (for example [[,abc,]]) caused
2689 pcre_compile() to give the error "Failed: internal error: code overflow" or
2690 in some cases to crash with a glibc free() error. This could even happen if
2691 the pattern terminated after [[ but there just happened to be a sequence of
2692 letters, a binary zero, and a closing ] in the memory that followed.
2693
2694 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2695 over the years. Originally (before any Unicode support), just the bottom 8
2696 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2697 output from "man perlunicode" includes this:
2698
2699 The regular expression compiler produces polymorphic opcodes. That
2700 is, the pattern adapts to the data and automatically switches to
2701 the Unicode character scheme when presented with Unicode data--or
2702 instead uses a traditional byte scheme when presented with byte
2703 data.
2704
2705 Sadly, a wide octal escape does not cause a switch, and in a string with
2706 no other multibyte characters, these octal escapes are treated as before.
2707 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2708 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2709 Unicode string.
2710
2711 I have not perpetrated such confusion in PCRE. Up till now, it took just
2712 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2713 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2714 translate to the appropriate multibyte character.
2715
2716 29. Applied some refactoring to reduce the number of warnings from Microsoft
2717 and Borland compilers. This has included removing the fudge introduced
2718 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2719 a warning about an unused variable.
2720
2721 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2722 characters since release 4.0, because Perl (from release 5.004) does not.
2723 [Or at least, is documented not to: some releases seem to be in conflict
2724 with the documentation.] However, when a pattern was studied with
2725 pcre_study() and all its branches started with \s, PCRE still included VT
2726 as a possible starting character. Of course, this did no harm; it just
2727 caused an unnecessary match attempt.
2728
2729 22. Removed a now-redundant internal flag bit that recorded the fact that case
2730 dependency changed within the pattern. This was once needed for "required
2731 byte" processing, but is no longer used. This recovers a now-scarce options
2732 bit. Also moved the least significant internal flag bit to the most-
2733 significant bit of the word, which was not previously used (hangover from
2734 the days when it was an int rather than a uint) to free up another bit for
2735 the future.
2736
2737 23. Added support for CRLF line endings as well as CR and LF. As well as the
2738 default being selectable at build time, it can now be changed at runtime
2739 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2740 specify that it is scanning data with non-default line endings.
2741
2742 24. Changed the definition of CXXLINK to make it agree with the definition of
2743 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2744
2745 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2746 recursions. This makes a big different to stack usage for some patterns.
2747
2748 26. If a subpattern containing a named recursion or subroutine reference such
2749 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2750 the space required for the compiled pattern went wrong and gave too small a
2751 value. Depending on the environment, this could lead to "Failed: internal
2752 error: code overflow at offset 49" or "glibc detected double free or
2753 corruption" errors.
2754
2755 27. Applied patches from Google (a) to support the new newline modes and (b) to
2756 advance over multibyte UTF-8 characters in GlobalReplace.
2757
2758 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2759 difference for some implementation of PCRE in some Windows version.
2760
2761 29. Added some extra testing facilities to pcretest:
2762
2763 \q<number> in a data line sets the "match limit" value
2764 \Q<number> in a data line sets the "match recursion limt" value
2765 -S <number> sets the stack size, where <number> is in megabytes
2766
2767 The -S option isn't available for Windows.
2768
2769
2770 Version 6.6 06-Feb-06
2771 ---------------------
2772
2773 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2774 in pcreposix.h. I have copied the definition from pcre.h.
2775
2776 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2777 because pcre.h is no longer a built file.
2778
2779 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2780 not normally included in the compiled code.
2781
2782
2783 Version 6.5 01-Feb-06
2784 ---------------------
2785
2786 1. When using the partial match feature with pcre_dfa_exec(), it was not
2787 anchoring the second and subsequent partial matches at the new starting
2788 point. This could lead to incorrect results. For example, with the pattern
2789 /1234/, partially matching against "123" and then "a4" gave a match.
2790
2791 2. Changes to pcregrep:
2792
2793 (a) All non-match returns from pcre_exec() were being treated as failures
2794 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2795 error message is output. Some extra information is given for the
2796 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2797 probably the only errors that are likely to be caused by users (by
2798 specifying a regex that has nested indefinite repeats, for instance).
2799 If there are more than 20 of these errors, pcregrep is abandoned.
2800
2801 (b) A binary zero was treated as data while matching, but terminated the
2802 output line if it was written out. This has been fixed: binary zeroes
2803 are now no different to any other data bytes.
2804
2805 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2806 used to set a locale for matching. The --locale=xxxx long option has
2807 been added (no short equivalent) to specify a locale explicitly on the
2808 pcregrep command, overriding the environment variables.
2809
2810 (d) When -B was used with -n, some line numbers in the output were one less
2811 than they should have been.
2812
2813 (e) Added the -o (--only-matching) option.
2814
2815 (f) If -A or -C was used with -c (count only), some lines of context were
2816 accidentally printed for the final match.
2817
2818 (g) Added the -H (--with-filename) option.
2819
2820 (h) The combination of options -rh failed to suppress file names for files
2821 that were found from directory arguments.
2822
2823 (i) Added the -D (--devices) and -d (--directories) options.
2824
2825 (j) Added the -F (--fixed-strings) option.
2826
2827 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2828
2829 (l) Added the --colo(u)r option.
2830
2831 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2832 is not present by default.
2833
2834 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2835 items such as (?R) or (?1), when the recursion could match a number of
2836 alternatives. If it matched one of the alternatives, but subsequently,
2837 outside the recursion, there was a failure, the code tried to back up into
2838 the recursion. However, because of the way PCRE is implemented, this is not
2839 possible, and the result was an incorrect result from the match.
2840
2841 In order to prevent this happening, the specification of recursion has
2842 been changed so that all such subpatterns are automatically treated as
2843 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2844
2845 4. I had overlooked the fact that, in some locales, there are characters for
2846 which isalpha() is true but neither isupper() nor islower() are true. In
2847 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2848 and ordfeminine) are like this. This affected the treatment of \w and \W
2849 when they appeared in character classes, but not when they appeared outside
2850 a character class. The bit map for "word" characters is now created
2851 separately from the results of isalnum() instead of just taking it from the
2852 upper, lower, and digit maps. (Plus the underscore character, of course.)
2853
2854 5. The above bug also affected the handling of POSIX character classes such as
2855 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2856 permanent tables. Instead, the bit maps for such a class were previously
2857 created as the appropriate unions of the upper, lower, and digit bitmaps.
2858 Now they are created by subtraction from the [[:word:]] class, which has
2859 its own bitmap.
2860
2861 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2862 It is created by subtracting the vertical space characters (\x09, \x0a,
2863 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2864 subtraction was done in the overall bitmap for a character class, meaning
2865 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2866 be recognized. This bug has been fixed.
2867
2868 7. Patches from the folks at Google:
2869
2870 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2871 real life, but is still worth protecting against".
2872
2873 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2874 regular expressions".
2875
2876 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2877 have it.
2878
2879 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2880 "configure" and the latter not, in order to fix a problem somebody had
2881 with compiling the Arg class on HP-UX.
2882
2883 (e) Improve the error-handling of the C++ wrapper a little bit.
2884
2885 (f) New tests for checking recursion limiting.
2886
2887 8. The pcre_memmove() function, which is used only if the environment does not
2888 have a standard memmove() function (and is therefore rarely compiled),
2889 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2890 returning a result (though PCRE never actually uses the result).
2891
2892 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2893 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2894 returned instead of calling malloc() with an overflowing number that would
2895 most likely cause subsequent chaos.
2896
2897 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2898
2899 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2900 with this option is matched, the nmatch and pmatch options of regexec() are
2901 ignored.
2902
2903 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2904 provided in case anyone wants to the the POSIX interface with UTF-8
2905 strings.
2906
2907 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2908 C++ linking (needed for some HP-UX environments).
2909
2910 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2911 (unused parameter) and in the pcre_printint() function (omitted "default"
2912 switch label when the default is to do nothing).
2913
2914 15. Added some code to make it possible, when PCRE is compiled as a C++
2915 library, to replace subject pointers for pcre_exec() with a smart pointer
2916 class, thus making it possible to process discontinuous strings.
2917
2918 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2919 much the same function. They were added by different people who were trying
2920 to make PCRE easy to compile on non-Unix systems. It has been suggested
2921 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2922 for compiling on Windows systems. I have therefore replaced it with
2923 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2924 defaults to "extern" for C or "extern C" for C++, which works fine on
2925 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2926 SCOPE with something explicit in config.h. In addition:
2927
2928 (a) pcreposix.h still had just "extern" instead of either of these macros;
2929 I have replaced it with PCRE_DATA_SCOPE.
2930
2931 (b) Functions such as _pcre_xclass(), which are internal to the library,
2932 but external in the C sense, all had PCRE_EXPORT in their definitions.
2933 This is apparently wrong for the Windows case, so I have removed it.
2934 (It makes no difference on Unix-like systems.)
2935
2936 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2937 of recursive calls to match(). This is different to MATCH_LIMIT because
2938 that limits the total number of calls to match(), not all of which increase
2939 the depth of recursion. Limiting the recursion depth limits the amount of
2940 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2941 when PCRE is compiled, and changed at run time. A patch from Google adds
2942 this functionality to the C++ interface.
2943
2944 18. Changes to the handling of Unicode character properties:
2945
2946 (a) Updated the table to Unicode 4.1.0.
2947
2948 (b) Recognize characters that are not in the table as "Cn" (undefined).
2949
2950 (c) I revised the way the table is implemented to a much improved format
2951 which includes recognition of ranges. It now supports the ranges that
2952 are defined in UnicodeData.txt, and it also amalgamates other
2953 characters into ranges. This has reduced the number of entries in the
2954 table from around 16,000 to around 3,000, thus reducing its size
2955 considerably. I realized I did not need to use a tree structure after
2956 all - a binary chop search is just as efficient. Having reduced the
2957 number of entries, I extended their size from 6 bytes to 8 bytes to
2958 allow for more data.
2959
2960 (d) Added support for Unicode script names via properties such as \p{Han}.
2961
2962 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2963 matching that character.
2964
2965 20. When matching a repeated Unicode property with a minimum greater than zero,
2966 (for example \pL{2,}), PCRE could look past the end of the subject if it
2967 reached it while seeking the minimum number of characters. This could
2968 happen only if some of the characters were more than one byte long, because
2969 there is a check for at least the minimum number of bytes.
2970
2971 21. Refactored the implementation of \p and \P so as to be more general, to
2972 allow for more different types of property in future. This has changed the
2973 compiled form incompatibly. Anybody with saved compiled patterns that use
2974 \p or \P will have to recompile them.
2975
2976 22. Added "Any" and "L&" to the supported property types.
2977
2978 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2979 but give a compile time error if the value is greater than 0xff.
2980
2981 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2982 accidentally not being installed or uninstalled.
2983
2984 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2985 made were to insert the current release number. This seemed silly, because
2986 it made things harder for people building PCRE on systems that don't run
2987 "configure". I have turned pcre.h into a distributed file, no longer built
2988 by "configure", with the version identification directly included. There is
2989 no longer a pcre.h.in file.
2990
2991 However, this change necessitated a change to the pcre-config script as
2992 well. It is built from pcre-config.in, and one of the substitutions was the
2993 release number. I have updated configure.ac so that ./configure now finds
2994 the release number by grepping pcre.h.
2995
2996 26. Added the ability to run the tests under valgrind.
2997
2998
2999 Version 6.4 05-Sep-05
3000 ---------------------
3001
3002 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3003 "--" to be printed when multiple files were scanned, even when none of the
3004 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3005 consider it to be a bug, and have restored the previous behaviour.
3006
3007 2. A couple of code tidies to get rid of compiler warnings.
3008
3009 3. The pcretest program used to cheat by referring to symbols in the library
3010 whose names begin with _pcre_. These are internal symbols that are not
3011 really supposed to be visible externally, and in some environments it is
3012 possible to suppress them. The cheating is now confined to including
3013 certain files from the library's source, which is a bit cleaner.
3014
3015 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3016 file's purpose clearer.
3017
3018 5. Reorganized pcre_ucp_findchar().
3019
3020
3021 Version 6.3 15-Aug-05
3022 ---------------------
3023
3024 1. The file libpcre.pc.in did not have general read permission in the tarball.
3025
3026 2. There were some problems when building without C++ support:
3027
3028 (a) If C++ support was not built, "make install" and "make test" still
3029 tried to test it.
3030
3031 (b) There were problems when the value of CXX was explicitly set. Some
3032 changes have been made to try to fix these, and ...
3033
3034 (c) --disable-cpp can now be used to explicitly disable C++ support.
3035
3036 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3037 backslash in a target when C++ was disabled. This confuses some
3038 versions of "make", apparently. Using an intermediate variable solves
3039 this. (Same for CPP_LOBJ.)
3040
3041 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3042 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3043 necessary on certain architectures.
3044
3045 4. Added a setting of -export-symbols-regex to the link command to remove
3046 those symbols that are exported in the C sense, but actually are local
3047 within the library, and not documented. Their names all begin with
3048 "_pcre_". This is not a perfect job, because (a) we have to except some
3049 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3050 available (and never for static libraries). I have made a note to try to
3051 find a way round (a) in the future.
3052
3053
3054 Version 6.2 01-Aug-05
3055 ---------------------
3056
3057 1. There was no test for integer overflow of quantifier values. A construction
3058 such as {1111111111111111} would give undefined results. What is worse, if
3059 a minimum quantifier for a parenthesized subpattern overflowed and became
3060 negative, the calculation of the memory size went wrong. This could have
3061 led to memory overwriting.
3062
3063 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3064
3065 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3066 operating environments where this matters.
3067
3068 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3069 PCRE options from within the C++ wrapper.
3070
3071 5. Named capturing subpatterns were not being correctly counted when a pattern
3072 was compiled. This caused two problems: (a) If there were more than 100
3073 such subpatterns, the calculation of the memory needed for the whole
3074 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3075 back references of the form \12, where the number was greater than 9, were
3076 not recognized as back references, even though there were sufficient
3077 previous subpatterns.
3078
3079 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3080 versions of gcc, e.g. 2.95.4.
3081
3082
3083 Version 6.1 21-Jun-05
3084 ---------------------
3085
3086 1. There was one reference to the variable "posix" in pcretest.c that was not
3087 surrounded by "#if !defined NOPOSIX".
3088
3089 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3090 the cross-check on the old pcre_info() function, for the benefit of the
3091 cut-down version of PCRE that is currently imported into Exim.
3092
3093 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3094 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3095 patterns that start (?i) but I don't think that matters. The use of (?i) is
3096 just an example; this all applies to the other options as well.
3097
3098 4. Since libtool seems to echo the compile commands it is issuing, the output
3099 from "make" can be reduced a bit by putting "@" in front of each libtool
3100 compile command.
3101
3102 5. Patch from the folks at Google for configure.in to be a bit more thorough
3103 in checking for a suitable C++ installation before trying to compile the
3104 C++ stuff. This should fix a reported problem when a compiler was present,
3105 but no suitable headers.
3106
3107 6. The man pages all had just "PCRE" as their title. I have changed them to
3108 be the relevant file name. I have also arranged that these names are
3109 retained in the file doc/pcre.txt, which is a concatenation in text format
3110 of all the man pages except the little individual ones for each function.
3111
3112 7. The NON-UNIX-USE file had not been updated for the different set of source
3113 files that come with release 6. I also added a few comments about the C++
3114 wrapper.
3115
3116
3117 Version 6.0 07-Jun-05
3118 ---------------------
3119
3120 1. Some minor internal re-organization to help with my DFA experiments.
3121
3122 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3123 didn't matter for the library itself when fully configured, but did matter
3124 when compiling without UCP support, or within Exim, where the ucp files are
3125 not imported.
3126
3127 3. Refactoring of the library code to split up the various functions into
3128 different source modules. The addition of the new DFA matching code (see
3129 below) to a single monolithic source would have made it really too
3130 unwieldy, quite apart from causing all the code to be include in a
3131 statically linked application, when only some functions are used. This is
3132 relevant even without the DFA addition now that patterns can be compiled in
3133 one application and matched in another.
3134
3135 The downside of splitting up is that there have to be some external
3136 functions and data tables that are used internally in different modules of
3137 the library but which are not part of the API. These have all had their
3138 names changed to start with "_pcre_" so that they are unlikely to clash
3139 with other external names.
3140
3141 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3142 a different (DFA) algorithm. Although it is slower than the original
3143 function, it does have some advantages for certain types of matching
3144 problem.
3145
3146 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3147 including restarting after a partial match.
3148
3149 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3150 defined when compiling for Windows was sent to me. I have put it into the
3151 code, though I have no means of testing or verifying it.
3152
3153 7. Added the pcre_refcount() auxiliary function.
3154
3155 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3156 match before or at the first newline in the subject string. In pcretest,
3157 the /f option on a pattern can be used to set this.
3158
3159 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3160 would behave wrongly. This has been present in PCRE since release 4.0.
3161
3162 10. A number of changes to the pcregrep command:
3163
3164 (a) Refactored how -x works; insert ^(...)$ instead of setting
3165 PCRE_ANCHORED and checking the length, in preparation for adding
3166 something similar for -w.
3167
3168 (b) Added the -w (match as a word) option.
3169
3170 (c) Refactored the way lines are read and buffered so as to have more
3171 than one at a time available.
3172
3173 (d) Implemented a pcregrep test script.
3174
3175 (e) Added the -M (multiline match) option. This allows patterns to match
3176 over several lines of the subject. The buffering ensures that at least
3177 8K, or the rest of the document (whichever is the shorter) is available
3178 for matching (and similarly the previous 8K for lookbehind assertions).
3179
3180 (f) Changed the --help output so that it now says
3181
3182 -w, --word-regex(p)
3183
3184 instead of two lines, one with "regex" and the other with "regexp"
3185 because that confused at least one person since the short forms are the
3186 same. (This required a bit of code, as the output is generated
3187 automatically from a table. It wasn't just a text change.)
3188
3189 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3190 option but starts with a hyphen. Could be a pattern or a path name
3191 starting with a hyphen, for instance.
3192
3193 (h) "-" can be given as a file name to represent stdin.
3194
3195 (i) When file names are being printed, "(standard input)" is used for
3196 the standard input, for compatibility with GNU grep. Previously
3197 "<stdin>" was used.
3198
3199 (j) The option --label=xxx can be used to supply a name to be used for
3200 stdin when file names are being printed. There is no short form.
3201
3202 (k) Re-factored the options decoding logic because we are going to add
3203 two more options that take data. Such options can now be given in four
3204 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3205
3206 (l) Added the -A, -B, and -C options for requesting that lines of context
3207 around matches be printed.
3208
3209 (m) Added the -L option to print the names of files that do not contain
3210 any matching lines, that is, the complement of -l.
3211
3212 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3213 continue to scan other files.
3214
3215 (o) The -s option was incorrectly implemented. For compatibility with other
3216 greps, it now suppresses the error message for a non-existent or non-
3217 accessible file (but not the return code). There is a new option called
3218 -q that suppresses the output of matching lines, which was what -s was
3219 previously doing.
3220
3221 (p) Added --include and --exclude options to specify files for inclusion
3222 and exclusion when recursing.
3223
3224 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3225 Hopefully, it now does.
3226
3227 12. Missing cast in pcre_study().
3228
3229 13. Added an "uninstall" target to the makefile.
3230
3231 14. Replaced "extern" in the function prototypes in Makefile.in with
3232 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3233 world, but is set differently for Windows.
3234
3235 15. Added a second compiling function called pcre_compile2(). The only
3236 difference is that it has an extra argument, which is a pointer to an
3237 integer error code. When there is a compile-time failure, this is set
3238 non-zero, in addition to the error test pointer being set to point to an
3239 error message. The new argument may be NULL if no error number is required
3240 (but then you may as well call pcre_compile(), which is now just a
3241 wrapper). This facility is provided because some applications need a
3242 numeric error indication, but it has also enabled me to tidy up the way
3243 compile-time errors are handled in the POSIX wrapper.
3244
3245 16. Added VPATH=.libs to the makefile; this should help when building with one
3246 prefix path and installing with another. (Or so I'm told by someone who
3247 knows more about this stuff than I do.)
3248
3249 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3250 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3251 match everything, including newlines. This is not POSIX-compatible, but
3252 somebody wanted the feature. From pcretest it can be activated by using
3253 both the P and the s flags.
3254
3255 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3256
3257 19. libpcre.pc was being incorrectly installed as executable.
3258
3259 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3260 it now also looks for '\r' so that it will work unmodified on Windows.
3261
3262 21. Added Google's contributed C++ wrapper to the distribution.
3263
3264 22. Added some untidy missing memory free() calls in pcretest, to keep
3265 Electric Fence happy when testing.
3266
3267
3268
3269 Version 5.0 13-Sep-04
3270 ---------------------
3271
3272 1. Internal change: literal characters are no longer packed up into items
3273 containing multiple characters in a single byte-string. Each character
3274 is now matched using a separate opcode. However, there may be more than one
3275 byte in the character in UTF-8 mode.
3276
3277 2. The pcre_callout_block structure has two new fields: pattern_position and
3278 next_item_length. These contain the offset in the pattern to the next match
3279 item, and its length, respectively.
3280
3281 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3282 insertion of callouts before each pattern item. Added the /C option to
3283 pcretest to make use of this.
3284
3285 4. On the advice of a Windows user, the lines
3286
3287 #if defined(_WIN32) || defined(WIN32)
3288 _setmode( _fileno( stdout ), 0x8000 );
3289 #endif /* defined(_WIN32) || defined(WIN32) */
3290
3291 have been added to the source of pcretest. This apparently does useful
3292 magic in relation to line terminators.
3293
3294 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3295 for the benefit of those environments where the "b" makes a difference.
3296
3297 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3298 to know about it. I have put a hack into configure.in that adds in code
3299 to set GCC=yes if CC=icc. This seems to end up at a point in the
3300 generated configure script that is early enough to affect the setting of
3301 compiler options, which is what is needed, but I have no means of testing
3302 whether it really works. (The user who reported this had patched the
3303 generated configure script, which of course I cannot do.)
3304
3305 LATER: After change 22 below (new libtool files), the configure script
3306 seems to know about icc (and also ecc). Therefore, I have commented out
3307 this hack in configure.in.
3308
3309 7. Added support for pkg-config (2 patches were sent in).
3310
3311 8. Negated POSIX character classes that used a combination of internal tables
3312 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3313 [[:^ascii]]. Typically, they would match almost any characters. The other
3314 POSIX classes were not broken in this way.
3315
3316 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3317 to find the match, as PCRE was deluded into thinking that the match had to
3318 start at the start point or following a newline. The same bug applied to
3319 patterns with negative forward assertions or any backward assertions
3320 preceding ".*" at the start, unless the pattern required a fixed first
3321 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3322
3323 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3324 starting at the last subject character, bytes beyond the end of the subject
3325 string were read.
3326
3327 11. Renamed the variable "class" as "classbits" to make life easier for C++
3328 users. (Previously there was a macro definition, but it apparently wasn't
3329 enough.)
3330
3331 12. Added the new field "tables" to the extra data so that tables can be passed
3332 in at exec time, or the internal tables can be re-selected. This allows
3333 a compiled regex to be saved and re-used at a later time by a different
3334 program that might have everything at different addresses.
3335
3336 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3337 -R library as well as a -L library.
3338
3339 14. The debugging options of pcretest (-d on the command line or D on a
3340 pattern) showed incorrect output for anything following an extended class
3341 that contained multibyte characters and which was followed by a quantifier.
3342
3343 15. Added optional support for general category Unicode character properties
3344 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3345 support. It adds about 90K to the size of the library. The meanings of the
3346 inbuilt class escapes such as \d and \s have NOT been changed.
3347
3348 16. Updated pcredemo.c to include calls to free() to release the memory for the
3349 compiled pattern.
3350
3351 17. The generated file chartables.c was being created in the source directory
3352 instead of in the building directory. This caused the build to fail if the
3353 source directory was different from the building directory, and was
3354 read-only.
3355
3356 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3357 file. No doubt somebody will tell me if they don't make sense... Also added
3358 Dan Mooney's comments about building on OpenVMS.
3359
3360 19. Added support for partial matching via the PCRE_PARTIAL option for
3361 pcre_exec() and the \P data escape in pcretest.
3362
3363 20. Extended pcretest with 3 new pattern features:
3364
3365 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3366 write the compiled pattern to the file whose name is "rest-of-line".
3367 This is a straight binary dump of the data, with the saved pointer to
3368 the character tables forced to be NULL. The study data, if any, is
3369 written too. After writing, pcretest reads a new pattern.
3370
3371 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3372 compiled pattern from the given file. There must not be any
3373 occurrences of "<" in the file name (pretty unlikely); if there are,
3374 pcretest will instead treat the initial "<" as a pattern delimiter.
3375 After reading in the pattern, pcretest goes on to read data lines as
3376 usual.
3377
3378 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3379 and 16-bit fields in a compiled pattern, to simulate a pattern that
3380 was compiled on a host of opposite endianness.
3381
3382 21. The pcre-exec() function can now cope with patterns that were compiled on
3383 hosts of opposite endianness, with this restriction:
3384
3385 As for any compiled expression that is saved and used later, the tables
3386 pointer field cannot be preserved; the extra_data field in the arguments
3387 to pcre_exec() should be used to pass in a tables address if a value
3388 other than the default internal tables were used at compile time.
3389
3390 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3391 now diagnosed as an error. Previously, most of the time, a negative number
3392 would have been treated as zero, but if in addition "ovector" was passed as
3393 NULL, a crash could occur.
3394
3395 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3396 new versions from the libtool 1.5 distribution (the last one is a copy of
3397 a file called libtool.m4). This seems to have fixed the need to patch
3398 "configure" to support Darwin 1.3 (which I used to do). However, I still
3399 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3400 workstation).
3401
3402 24. Changed the PCRE licence to be the more standard "BSD" licence.
3403
3404
3405 Version 4.5 01-Dec-03
3406 ---------------------
3407
3408 1. There has been some re-arrangement of the code for the match() function so
3409 that it can be compiled in a version that does not call itself recursively.
3410 Instead, it keeps those local variables that need separate instances for
3411 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3412 needs to "recurse". Keeping track of where control must go is done by means
3413 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3414 hide most of the details from the main code, and operates only if
3415 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3416 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3417 operating.
3418
3419 To make it easier for callers to provide specially tailored get/free
3420 functions for this usage, two new functions, pcre_stack_malloc, and
3421 pcre_stack_free, are used. They are always called in strict stacking order,
3422 and the size of block requested is always the same.
3423
3424 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3425 PCRE has been compiled to use the stack or the heap for recursion. The
3426 -C option of pcretest uses this to show which version is compiled.
3427
3428 A new data escape \S, is added to pcretest; it causes the amounts of store
3429 obtained and freed by both kinds of malloc/free at match time to be added
3430 to the output.
3431
3432 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3433 what's available on my current Linux desktop machine.
3434
3435 3. When matching a UTF-8 string, the test for a valid string at the start has
3436 been extended. If start_offset is not zero, PCRE now checks that it points
3437 to a byte that is the start of a UTF-8 character. If not, it returns
3438 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3439 this is necessary because there may be backward assertions in the pattern.
3440 When matching the same subject several times, it may save resources to use
3441 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3442
3443 4. The code for checking the validity of UTF-8 strings has been tightened so
3444 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3445 containing "overlong sequences".
3446
3447 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3448 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3449 should have been "&", but it just so happened that all the cases this let
3450 through by mistake were picked up later in the function.
3451
3452 6. I had used a variable called "isblank" - this is a C99 function, causing
3453 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3454
3455 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3456 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3457 so the version is shown; (c) stop "make test" echoing "./RunTest".
3458
3459 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3460
3461 9. The prototype for memmove() for systems that don't have it was using
3462 size_t, but the inclusion of the header that defines size_t was later. I've
3463 moved the #includes for the C headers earlier to avoid this.
3464
3465 10. Added some adjustments to the code to make it easier to compiler on certain
3466 special systems:
3467
3468 (a) Some "const" qualifiers were missing.
3469 (b) Added the macro EXPORT before all exported functions; by default this
3470 is defined to be empty.
3471 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3472 that it reads its output file name as an argument instead of writing
3473 to the standard output and assuming this can be redirected.
3474
3475 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3476 class containing characters with values greater than 255, PCRE compilation
3477 went into a loop.
3478
3479 12. A recursive reference to a subpattern that was within another subpattern
3480 that had a minimum quantifier of zero caused PCRE to crash. For example,
3481 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3482 recursion. If the recursively-called subpattern itself had a zero repeat,
3483 that was OK.
3484
3485 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3486 buffer into which it was copied (for escape processing) was still set at
3487 1024, so long lines caused crashes.
3488
3489 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3490 "internal error: code overflow...". This applied to any character class
3491 that was followed by a possessive quantifier.
3492
3493 15. Modified the Makefile to add libpcre.la as a prerequisite for
3494 libpcreposix.la because I was told this is needed for a parallel build to
3495 work.
3496
3497 16. If a pattern that contained .* following optional items at the start was
3498 studied, the wrong optimizing data was generated, leading to matching
3499 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3500 matching string must start with a or b or c. The correct conclusion for
3501 this pattern is that a match can start with any character.
3502
3503
3504 Version 4.4 13-Aug-03
3505 ---------------------
3506
3507 1. In UTF-8 mode, a character class containing characters with values between
3508 127 and 255 was not handled correctly if the compiled pattern was studied.
3509 In fixing this, I have also improved the studying algorithm for such
3510 classes (slightly).
3511
3512 2. Three internal functions had redundant arguments passed to them. Removal
3513 might give a very teeny performance improvement.
3514
3515 3. Documentation bug: the value of the capture_top field in a callout is *one
3516 more than* the number of the hightest numbered captured substring.
3517
3518 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3519 in incorrectly linking with a previously installed version. They now link
3520 explicitly with libpcre.la.
3521
3522 5. configure.in no longer needs to recognize Cygwin specially.
3523
3524 6. A problem in pcre.in for Windows platforms is fixed.
3525
3526 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3527 pcretest, it used to include the size of the study block as part of its
3528 output. Unfortunately, the structure contains a field that has a different
3529 size on different hardware architectures. This meant that the tests that
3530 showed this size failed. As the block is currently always of a fixed size,
3531 this information isn't actually particularly useful in pcretest output, so
3532 I have just removed it.
3533
3534 8. Three pre-processor statements accidentally did not start in column 1.
3535 Sadly, there are *still* compilers around that complain, even though
3536 standard C has not required this for well over a decade. Sigh.
3537
3538 9. In pcretest, the code for checking callouts passed small integers in the
3539 callout_data field, which is a void * field. However, some picky compilers
3540 complained about the casts involved for this on 64-bit systems. Now
3541 pcretest passes the address of the small integer instead, which should get
3542 rid of the warnings.
3543
3544 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3545 both compile and run time, and gives an error if an invalid UTF-8 sequence
3546 is found. There is a option for disabling this check in cases where the
3547 string is known to be correct and/or the maximum performance is wanted.
3548
3549 11. In response to a bug report, I changed one line in Makefile.in from
3550
3551 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3552 to
3553 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3554
3555 to look similar to other lines, but I have no way of telling whether this
3556 is the right thing to do, as I do not use Windows. No doubt I'll get told
3557 if it's wrong...
3558
3559
3560 Version 4.3 21-May-03
3561 ---------------------
3562
3563 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3564 Makefile.
3565
3566 2. Some refactoring to improve the quality of the code:
3567
3568 (i) The utf8_table... variables are now declared "const".
3569
3570 (ii) The code for \cx, which used the "case flipping" table to upper case
3571 lower case letters, now just substracts 32. This is ASCII-specific,
3572 but the whole concept of \cx is ASCII-specific, so it seems
3573 reasonable.
3574
3575 (iii) PCRE was using its character types table to recognize decimal and
3576 hexadecimal digits in the pattern. This is silly, because it handles
3577 only 0-9, a-f, and A-F, but the character types table is locale-
3578 specific, which means strange things might happen. A private
3579 table is now used for this - though it costs 256 bytes, a table is
3580 much faster than multiple explicit tests. Of course, the standard
3581 character types table is still used for matching digits in subject
3582 strings against \d.
3583
3584 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3585 ending in _t are). So I've renamed it as ESC_tee.
3586
3587 3. The first argument for regexec() in the POSIX wrapper should have been
3588 defined as "const".
3589
3590 4. Changed pcretest to use malloc() for its buffers so that they can be
3591 Electric Fenced for debugging.
3592
3593 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3594 to read one or more bytes before the start of the subject string. Often this
3595 had no effect on PCRE's behaviour, but in some circumstances it could
3596 provoke a segmentation fault.
3597
3598 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3599 to try to read one or more bytes before the start of the subject string.
3600
3601 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3602 UTF-8 support could misbehave in various ways if the subject string
3603 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3604 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3605 back over UTF-8 characters.)
3606
3607
3608 Version 4.2 14-Apr-03
3609 ---------------------
3610
3611 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3612
3613 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3614 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3615 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3616 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3617 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3618 and BUILD_EXEEXT
3619 Note: automatic setting of the BUILD variables is not yet working
3620 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3621 compile-time but not at link-time
3622 [LINK]: use for linking executables only
3623 make different versions for Windows and non-Windows
3624 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3625 libraries
3626 [LINK_FOR_BUILD]: new variable
3627 [OBJEXT]: use throughout
3628 [EXEEXT]: use throughout
3629 <winshared>: new target
3630 <wininstall>: new target
3631 <dftables.o>: use native compiler
3632 <dftables>: use native linker
3633 <install>: handle Windows platform correctly
3634 <clean>: ditto
3635 <check>: ditto
3636 copy DLL to top builddir before testing
3637
3638 As part of these changes, -no-undefined was removed again. This was reported
3639 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3640 in any case.
3641
3642 3. Some tidies to get rid of compiler warnings:
3643
3644 . In the match_data structure, match_limit was an unsigned long int, whereas
3645 match_call_count was an int. I've made them both unsigned long ints.
3646
3647 . In pcretest the fact that a const uschar * doesn't automatically cast to
3648 a void * provoked a warning.
3649
3650 . Turning on some more compiler warnings threw up some "shadow" variables
3651 and a few more missing casts.
3652
3653 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3654 option, a class that contained a single character with a value between 128
3655 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3656
3657 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3658 option, a class that contained several characters, but with at least one
3659 whose value was between 128 and 255 caused PCRE to crash.
3660
3661
3662 Version 4.1 12-Mar-03
3663 ---------------------
3664
3665 1. Compiling with gcc -pedantic found a couple of places where casts were
3666 needed, and a string in dftables.c that was longer than standard compilers are
3667 required to support.
3668
3669 2. Compiling with Sun's compiler found a few more places where the code could
3670 be tidied up in order to avoid warnings.
3671
3672 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3673 first of these names is deprecated in the latest Autoconf in favour of the name
3674 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3675 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3676 analogy I have changed it to CFLAGS_FOR_BUILD.
3677
3678 4. Added -no-undefined to the linking command in the Makefile, because this is
3679 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3680 linking step for the pcreposix library.
3681
3682 5. PCRE was failing to diagnose the case of two named groups with the same
3683 name.
3684
3685 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3686 literal character that is needed in the subject for a match, and scans along to
3687 ensure that it is present before embarking on the full matching process. This
3688 saves time in cases of nested unlimited repeats that are never going to match.
3689 Problem: the scan can take a lot of time if the subject is very long (e.g.
3690 megabytes), thus penalizing straightforward matches. It is now done only if the
3691 amount of subject to be scanned is less than 1000 bytes.
3692
3693 7. A lesser problem with the same optimization is that it was recording the
3694 first character of an anchored pattern as "needed", thus provoking a search
3695 right along the subject, even when the first match of the pattern was going to
3696 fail. The "needed" character is now not set for anchored patterns, unless it
3697 follows something in the pattern that is of non-fixed length. Thus, it still
3698 fulfils its original purpose of finding quick non-matches in cases of nested
3699 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3700
3701
3702 Version 4.0 17-Feb-03
3703 ---------------------
3704
3705 1. If a comment in an extended regex that started immediately after a meta-item
3706 extended to the end of string, PCRE compiled incorrect data. This could lead to
3707 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3708
3709 2. Moved to autoconf 2.53 and libtool 1.4.2.
3710
3711 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3712 the special perltest8 script is no longer needed - all the tests can be run
3713 from a single perltest script.
3714
3715 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3716 by \s. It has now been removed in PCRE. This means it isn't recognized as
3717 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3718 class [:space:] *does* include VT, thereby creating a mess.
3719
3720 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3721 space and tab.
3722
3723 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3724 its new features into the main test script, reducing the number of scripts.
3725
3726 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3727 were backward compatible, and made the (?i) apply to the whole pattern, as if
3728 /i were given. Now it behaves more logically, and applies the option setting
3729 only to what follows. PCRE has been changed to follow suit. However, if it
3730 finds options settings right at the start of the pattern, it extracts them into
3731 the global options, as before. Thus, they show up in the info data.
3732
3733 8. Added support for the \Q...\E escape sequence. Characters in between are
3734 treated as literals. This is slightly different from Perl in that $ and @ are
3735 also handled as literals inside the quotes. In Perl, they will cause variable
3736 interpolation. Note the following examples:
3737
3738 Pattern PCRE matches Perl matches
3739
3740 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3741 \Qabc\$xyz\E abc\$xyz abc\$xyz
3742 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3743
3744 For compatibility with Perl, \Q...\E sequences are recognized inside character
3745 classes as well as outside them.
3746
3747 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3748 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3749 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3750 signed/unsigned warnings.
3751
3752 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3753 option for pcretest, so I've replaced it by a simple function that does just
3754 that job.
3755
3756 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3757 "pcregrep -".
3758
3759 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3760 Java package. This provides some syntactic sugar for simple cases of what my
3761 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3762 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3763 item, you can use this simplified notation. Note that only makes sense with
3764 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3765 greediness, whatever the setting of the PCRE_UNGREEDY option.
3766
3767 13. A change of greediness default within a pattern was not taking effect at
3768 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3769 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3770 was abstracted outside.
3771
3772 14. PCRE now supports the \G assertion. It is true when the current matching
3773 position is at the start point of the match. This differs from \A when the
3774 starting offset is non-zero. Used with the /g option of pcretest (or similar
3775 code), it works in the same way as it does for Perl's /g option. If all
3776 alternatives of a regex begin with \G, the expression is anchored to the start
3777 match position, and the "anchored" flag is set in the compiled expression.
3778
3779 15. Some bugs concerning the handling of certain option changes within patterns
3780 have been fixed. These applied to options other than (?ims). For example,
3781 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3782 been the other way round. Some of this was related to change 7 above.
3783
3784 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3785 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3786 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3787 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3788
3789 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3790 mode. Unlike ".", it always matches newline, whatever the setting of
3791 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3792 assertions. Perl allows it, but it doesn't (in general) work because it can't
3793 calculate the length of the lookbehind. At least, that's the case for Perl
3794 5.8.0 - I've been told they are going to document that it doesn't work in
3795 future.
3796
3797 18. Added an error diagnosis for escapes that PCRE does not support: these are
3798 \L, \l, \N, \P, \p, \U, \u, and \X.
3799
3800 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3801 reading past the end of the pattern in cases such as /[abcd/.
3802
3803 20. PCRE was getting more memory than necessary for patterns with classes that
3804 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3805
3806 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3807 compiling PCRE for use with Virtual Pascal.
3808
3809 22. Small fix to the Makefile to make it work properly if the build is done
3810 outside the source tree.
3811
3812 23. Added a new extension: a condition to go with recursion. If a conditional
3813 subpattern starts with (?(R) the "true" branch is used if recursion has
3814 happened, whereas the "false" branch is used only at the top level.
3815
3816 24. When there was a very long string of literal characters (over 255 bytes
3817 without UTF support, over 250 bytes with UTF support), the computation of how
3818 much memory was required could be incorrect, leading to segfaults or other
3819 strange effects.
3820
3821 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3822 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3823 there was a subsequent back reference to those brackets. This meant that, for
3824 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3825 possible to check for precisely this case. All we can do is abandon the
3826 optimization if .* occurs inside capturing brackets when there are any back
3827 references whatsoever. (See below for a better fix that came later.)
3828
3829 26. The handling of the optimization for finding the first character of a
3830 non-anchored pattern, and for finding a character that is required later in the
3831 match were failing in some cases. This didn't break the matching; it just
3832 failed to optimize when it could. The way this is done has been re-implemented.
3833
3834 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3835
3836 28. Added a new feature that provides some of the functionality that Perl
3837 provides with (?{...}). The facility is termed a "callout". The way it is done
3838 in PCRE is for the caller to provide an optional function, by setting
3839 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3840 global variable. By default it is unset, which disables all calling out. To get
3841 the function called, the regex must include (?C) at appropriate points. This
3842 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3843 This provides a means of identifying different callout points. When PCRE
3844 reaches such a point in the regex, if pcre_callout has been set, the external
3845 function is called. It is provided with data in a structure called
3846 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3847 matching continues; if it returns a non-zero value, the match at the current
3848 point fails. However, backtracking will occur if possible. [This was changed
3849 later and other features added - see item 49 below.]
3850
3851 29. pcretest is upgraded to test the callout functionality. It provides a
3852 callout function that displays information. By default, it shows the start of
3853 the match and the current position in the text. There are some new data escapes
3854 to vary what happens:
3855
3856 \C+ in addition, show current contents of captured substrings
3857 \C- do not supply a callout function
3858 \C!n return 1 when callout number n is reached
3859 \C!n!m return 1 when callout number n is reached for the mth time
3860
3861 30. If pcregrep was called with the -l option and just a single file name, it
3862 output "<stdin>" if a match was found, instead of the file name.
3863
3864 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3865 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3866 pcre_exec(). This saves a malloc/free per call. The default value of
3867 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3868 when configuring.
3869
3870 32. The default maximum size of a compiled pattern is 64K. There have been a
3871 few cases of people hitting this limit. The code now uses macros to handle the
3872 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3873 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3874 configuring. Tests 2 and 5 work only with 2-byte links because they output
3875 debugging information about compiled patterns.
3876
3877 33. Internal code re-arrangements:
3878
3879 (a) Moved the debugging function for printing out a compiled regex into
3880 its own source file (printint.c) and used #include to pull it into
3881 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3882 separate copies.
3883
3884 (b) Defined the list of op-code names for debugging as a macro in
3885 internal.h so that it is next to the definition of the opcodes.
3886
3887 (c) Defined a table of op-code lengths for simpler skipping along compiled
3888 code. This is again a macro in internal.h so that it is next to the
3889 definition of the opcodes.
3890
3891 34. Added support for recursive calls to individual subpatterns, along the
3892 lines of Robin Houston's patch (but implemented somewhat differently).
3893
3894 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3895 allow it to read and process whole directories in Win32. This code was
3896 contributed by Lionel Fourquaux; it has not been tested by me.
3897
3898 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3899 used to name a group. Names consist of alphanumerics and underscores, and must
3900 be unique. Back references use the syntax (?P=name) and recursive calls use
3901 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3902 numbers. The function pcre_fullinfo() can be used after compilation to extract
3903 a name/number map. There are three relevant calls:
3904
3905 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3906 PCRE_INFO_NAMECOUNT yields the number of entries
3907 PCRE_INFO_NAMETABLE yields a pointer to the map.
3908
3909 The map is a vector of fixed-size entries. The size of each entry depends on
3910 the length of the longest name used. The first two bytes of each entry are the
3911 group number, most significant byte first. There follows the corresponding
3912 name, zero terminated. The names are in alphabetical order.
3913
3914 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3915 case instead of 255. Making it the same both with and without UTF-8 support
3916 means that the same test output works with both.
3917
3918 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3919 calling malloc() with a zero argument.
3920
3921 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3922 optimization. I've improved things by keeping a bitmap of backreferences with
3923 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3924 fact referenced, the optimization can be applied. It is unlikely that a
3925 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3926 the match to follow \n) will appear inside brackets with a number greater than
3927 31, but if it does, any back reference > 31 suppresses the optimization.
3928
3929 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3930 of disabling numbered capturing parentheses. Any opening parenthesis that is
3931 not followed by ? behaves as if it were followed by ?: but named parentheses
3932 can still be used for capturing (and they will acquire numbers in the usual
3933 way).
3934
3935 41. Redesigned the return codes from the match() function into yes/no/error so
3936 that errors can be passed back from deep inside the nested calls. A malloc
3937 failure while inside a recursive subpattern call now causes the
3938 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3939
3940 42. It is now possible to set a limit on the number of times the match()
3941 function is called in a call to pcre_exec(). This facility makes it possible to
3942 limit the amount of recursion and backtracking, though not in a directly
3943 obvious way, because the match() function is used in a number of different
3944 circumstances. The count starts from zero for each position in the subject
3945 string (for non-anchored patterns). The default limit is, for compatibility, a
3946 large number, namely 10 000 000. You can change this in two ways:
3947
3948 (a) When configuring PCRE before making, you can use --with-match-limit=n
3949 to set a default value for the compiled library.
3950
3951 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3952 a different value is set. See 45 below.
3953
3954 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3955
3956 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3957 of things that can be changed at compile time. The first argument specifies
3958 what is wanted and the second points to where the information is to be placed.
3959 The current list of available information is:
3960
3961 PCRE_CONFIG_UTF8
3962
3963 The output is an integer that is set to one if UTF-8 support is available;
3964 otherwise it is set to zero.
3965
3966 PCRE_CONFIG_NEWLINE
3967
3968 The output is an integer that it set to the value of the code that is used for
3969 newline. It is either LF (10) or CR (13).
3970
3971 PCRE_CONFIG_LINK_SIZE
3972
3973 The output is an integer that contains the number of bytes used for internal
3974 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3975
3976 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3977
3978 The output is an integer that contains the threshold above which the POSIX
3979 interface uses malloc() for output vectors. See item 31 above.
3980
3981 PCRE_CONFIG_MATCH_LIMIT
3982
3983 The output is an unsigned integer that contains the default limit of the number
3984 of match() calls in a pcre_exec() execution. See 42 above.
3985
3986 44. pcretest has been upgraded by the addition of the -C option. This causes it
3987 to extract all the available output from the new pcre_config() function, and to
3988 output it. The program then exits immediately.
3989
3990 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3991 order to support additional features. One way would have been to define
3992 pcre_exec2() (for example) with extra arguments, but this would not have been
3993 extensible, and would also have required all calls to the original function to
3994 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3995 is used for passing in "extra" data from pcre_study().
3996
3997 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3998 contains the following fields:
3999
4000 flags a bitmap indicating which of the following fields are set
4001 study_data opaque data from pcre_study()
4002 match_limit a way of specifying a limit on match() calls for a specific
4003 call to pcre_exec()
4004 callout_data data for callouts (see 49 below)
4005
4006 The flag bits are also defined in pcre.h, and are
4007
4008 PCRE_EXTRA_STUDY_DATA
4009 PCRE_EXTRA_MATCH_LIMIT
4010 PCRE_EXTRA_CALLOUT_DATA
4011
4012 The pcre_study() function now returns one of these new pcre_extra blocks, with
4013 the actual study data pointed to by the study_data field, and the
4014 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4015 before. That is, this change is entirely upwards-compatible and requires no
4016 change to existing code.
4017
4018 If you want to pass in additional data to pcre_exec(), you can either place it
4019 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4020 block.
4021
4022 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4023 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4024 times with different match limits, until it finds the minimum value needed for
4025 pcre_exec() to complete. The value is then output. This can be instructive; for
4026 most simple matches the number is quite small, but for pathological cases it
4027 gets very large very quickly.
4028
4029 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4030 returns the size of the data block pointed to by the study_data field in a
4031 pcre_extra block, that is, the value that was passed as the argument to
4032 pcre_malloc() when PCRE was getting memory in which to place the information
4033 created by pcre_study(). The fourth argument should point to a size_t variable.
4034 pcretest has been extended so that this information is shown after a successful
4035 pcre_study() call when information about the compiled regex is being displayed.
4036
4037 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4038 because what follows is always an absolute path. (Later: it turns out that this
4039 is more than cosmetic for MinGW, because it doesn't like empty path
4040 components.)
4041
4042 49. Some changes have been made to the callout feature (see 28 above):
4043
4044 (i) A callout function now has three choices for what it returns:
4045
4046 0 => success, carry on matching
4047 > 0 => failure at this point, but backtrack if possible
4048 < 0 => serious error, return this value from pcre_exec()
4049
4050 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4051 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4052 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4053 use by callout functions. It will never be used by PCRE itself.
4054
4055 (ii) The pcre_extra structure (see 45 above) has a void * field called
4056 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4057 pcre_callout_block structure has a field of the same name. The contents of
4058 the field passed in the pcre_extra structure are passed to the callout
4059 function in the corresponding field in the callout block. This makes it
4060 easier to use the same callout-containing regex from multiple threads. For
4061 testing, the pcretest program has a new data escape
4062
4063 \C*n pass the number n (may be negative) as callout_data
4064
4065 If the callout function in pcretest receives a non-zero value as
4066 callout_data, it returns that value.
4067
4068 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4069 there were some redundant $(CFLAGS) in commands that are now specified as
4070 $(LINK), which already includes $(CFLAGS).
4071
4072 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4073 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4074 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4075 one-byte characters throughout. Note that case-insensitive matching applies
4076 only to characters whose values are less than 256. PCRE doesn't support the
4077 notion of cases for higher-valued characters.
4078
4079 (i) A character class whose characters are all within 0-255 is handled as
4080 a bit map, and the map is inverted for negative classes. Previously, a
4081 character > 255 always failed to match such a class; however it should
4082 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4083
4084 (ii) A negated character class with a single character < 255 is coded as
4085 "not this character" (OP_NOT). This wasn't working properly when the test
4086 character was multibyte, either singly or repeated.
4087
4088 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4089 mode, for example: \x{100}{2,3}.
4090
4091 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4092 singly or repeated) now correctly test multibyte characters. However,
4093 PCRE doesn't recognize any characters with values greater than 255 as
4094 digits, spaces, or word characters. Such characters always match \D, \S,
4095 and \W, and never match \d, \s, or \w.
4096
4097 (v) Classes may now contain characters and character ranges with values
4098 greater than 255. For example: [ab\x{100}-\x{400}].
4099
4100 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4101 PCRE in UTF-8 mode.
4102
4103 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4104 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4105 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4106 value.)
4107
4108 53. The single man page has become too large. I have therefore split it up into
4109 a number of separate man pages. These also give rise to individual HTML pages;
4110 these are now put in a separate directory, and there is an index.html page that
4111 lists them all. Some hyperlinking between the pages has been installed.
4112
4113 54. Added convenience functions for handling named capturing parentheses.
4114
4115 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4116 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4117 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4118 are faulted.
4119
4120 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4121 calling configure. These values are used when compiling the dftables.c program
4122 which is run to generate the source of the default character tables. They
4123 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4124 you will need to set these values.
4125
4126 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4127
4128
4129 Version 3.9 02-Jan-02
4130 ---------------------
4131
4132 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4133
4134 2. If --disable-static was given, the building process failed when trying to
4135 build pcretest and pcregrep. (For some reason it was using libtool to compile
4136 them, which is not right, as they aren't part of the library.)
4137
4138
4139 Version 3.8 18-Dec-01
4140 ---------------------
4141
4142 1. The experimental UTF-8 code was completely screwed up. It was packing the
4143 bytes in the wrong order. How dumb can you get?
4144
4145
4146 Version 3.7 29-Oct-01
4147 ---------------------
4148
4149 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4150 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4151 this didn't happen under Solaris 8, where I normally test things.
4152
4153 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4154 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4155 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4156 this fix an hour or so after the initial 3.7 release.)
4157
4158
4159 Version 3.6 23-Oct-01
4160 ---------------------
4161
4162 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4163 offsets passed as NULL with zero offset count.
4164
4165 2. The config.guess and config.sub files had not been updated when I moved to
4166 the latest autoconf.
4167
4168
4169 Version 3.5 15-Aug-01
4170 ---------------------
4171
4172 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4173 had been forgotten.
4174
4175 2. By using declared but undefined structures, we can avoid using "void"
4176 definitions in pcre.h while keeping the internal definitions of the structures
4177 private.
4178
4179 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4180 user point of view, this means that both static and shared libraries are built
4181 by default, but this can be individually controlled. More of the work of
4182 handling this static/shared cases is now inside libtool instead of PCRE's make
4183 file.
4184
4185 4. The pcretest utility is now installed along with pcregrep because it is
4186 useful for users (to test regexs) and by doing this, it automatically gets
4187 relinked by libtool. The documentation has been turned into a man page, so
4188 there are now .1, .txt, and .html versions in /doc.
4189
4190 5. Upgrades to pcregrep:
4191 (i) Added long-form option names like gnu grep.
4192 (ii) Added --help to list all options with an explanatory phrase.
4193 (iii) Added -r, --recursive to recurse into sub-directories.
4194 (iv) Added -f, --file to read patterns from a file.
4195
4196 6. pcre_exec() was referring to its "code" argument before testing that
4197 argument for NULL (and giving an error if it was NULL).
4198
4199 7. Upgraded Makefile.in to allow for compiling in a different directory from
4200 the source directory.
4201
4202 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4203 options bits, the pointer it was passed was to an int instead of to an unsigned
4204 long int. This mattered only on 64-bit systems.
4205
4206 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4207 generated) instead of pcre.in, which it its source. Also made the same change
4208 in several of the .c files.
4209
4210 10. A new release of gcc defines printf() as a macro, which broke pcretest
4211 because it had an ifdef in the middle of a string argument for printf(). Fixed
4212 by using separate calls to printf().
4213
4214 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4215 script, to force use of CR or LF instead of \n in the source. On non-Unix
4216 systems, the value can be set in config.h.
4217
4218 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4219 absolute limit. Changed the text of the error message to make this clear, and
4220 likewise updated the man page.
4221
4222 13. The limit of 99 on the number of capturing subpatterns has been removed.
4223 The new limit is 65535, which I hope will not be a "real" limit.
4224
4225
4226 Version 3.4 22-Aug-00
4227 ---------------------
4228
4229 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4230
4231 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4232
4233
4234 Version 3.3 01-Aug-00
4235 ---------------------
4236
4237 1. If an octal character was given, but the value was greater than \377, it
4238 was not getting masked to the least significant bits, as documented. This could
4239 lead to crashes in some systems.
4240
4241 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4242 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4243
4244 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4245 These just pass their arguments on to (pcre_free)(), but they are provided
4246 because some uses of PCRE bind it to non-C systems that can call its functions,
4247 but cannot call free() or pcre_free() directly.
4248
4249 4. Add "make test" as a synonym for "make check". Corrected some comments in
4250 the Makefile.
4251
4252 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4253 Makefile.
4254
4255 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4256 command called pgrep for grepping around the active processes.
4257
4258 7. Added the beginnings of support for UTF-8 character strings.
4259
4260 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4261 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4262 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4263 out for the ar command.)
4264
4265
4266 Version 3.2 12-May-00
4267 ---------------------
4268
4269 This is purely a bug fixing release.
4270
4271 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4272 of ZA. This was just one example of several cases that could provoke this bug,
4273 which was introduced by change 9 of version 2.00. The code for breaking
4274 infinite loops after an iteration that matches an empty string was't working
4275 correctly.
4276
4277 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4278 when matched against abbab (for example). After matching an empty string, it
4279 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4280 caused it to match further down the string than it should.
4281
4282 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4283 was there because it doesn't seem to be needed, and it causes trouble on some
4284 systems, as it is not a Standard C header. It has been removed.
4285
4286 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4287 were reported on the Macintosh. The changes were from
4288
4289 while ((c = *(++ptr)) != 0 && c != '\n');
4290 to
4291 while ((c = *(++ptr)) != 0 && c != '\n') ;
4292
4293 Totally extraordinary, but if that's what it takes...
4294
4295 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4296 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4297 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4298 assumes the way PCRE uses memmove() (always moving upwards).
4299
4300 6. PCRE is being used in one environment where strchr() is not available. There
4301 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4302 faster code anyway.
4303
4304
4305 Version 3.1 09-Feb-00
4306 ---------------------
4307
4308 The only change in this release is the fixing of some bugs in Makefile.in for
4309 the "install" target:
4310
4311 (1) It was failing to install pcreposix.h.
4312
4313 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4314
4315
4316 Version 3.0 01-Feb-00
4317 ---------------------
4318
4319 1. Add support for the /+ modifier to perltest (to output $` like it does in
4320 pcretest).
4321
4322 2. Add support for the /g modifier to perltest.
4323
4324 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4325 matches null strings.
4326
4327 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4328 pattern. Perl treats empty patterns specially - it reuses the most recent
4329 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4330 effect.
4331
4332 5. The POSIX interface was broken in that it was just handing over the POSIX
4333 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4334 required a bigger vector, with some working space on the end. This means that
4335 the POSIX wrapper now has to get and free some memory, and copy the results.
4336
4337 6. Added some simple autoconf support, placing the test data and the
4338 documentation in separate directories, re-organizing some of the
4339 information files, and making it build pcre-config (a GNU standard). Also added
4340 libtool support for building PCRE as a shared library, which is now the
4341 default.
4342
4343 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4344 09 are not valid octal constants. Single digits will be used for minor values
4345 less than 10.
4346
4347 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4348 existing programs that set these in the POSIX interface can use PCRE without
4349 modification.
4350
4351 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4352 return all that pcre_info() returns, plus additional data. The pcre_info()
4353 function is retained for compatibility, but is considered to be obsolete.
4354
4355 10. Added experimental recursion feature (?R) to handle one common case that
4356 Perl 5.6 will be able to do with (?p{...}).
4357
4358 11. Added support for POSIX character classes like [:alpha:], which Perl is
4359 adopting.
4360
4361
4362 Version 2.08 31-Aug-99
4363 ----------------------
4364
4365 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4366 trying to match at the startoffset position, but instead was moving forward to
4367 the next newline as if a previous match had failed.
4368
4369 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4370 and could get into a loop if a null string was matched other than at the start
4371 of the subject.
4372
4373 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4374 be distinguished at compile time, and for completeness also added PCRE_DATE.
4375
4376 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4377 in GnuWin32 environments.
4378
4379
4380 Version 2.07 29-Jul-99
4381 ----------------------
4382
4383 1. The documentation is now supplied in plain text form and HTML as well as in
4384 the form of man page sources.
4385
4386 2. C++ compilers don't like assigning (void *) values to other pointer types.
4387 In particular this affects malloc(). Although there is no problem in Standard
4388 C, I've put in casts to keep C++ compilers happy.
4389
4390 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4391 should be (const char *).
4392
4393 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4394 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4395 However, I haven't made this a standard facility. The documentation doesn't
4396 mention it, and the Makefile doesn't support it.
4397
4398 5. The Makefile now contains an "install" target, with editable destinations at
4399 the top of the file. The pcretest program is not installed.
4400
4401 6. pgrep -V now gives the PCRE version number and date.
4402
4403 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4404 causing the entire string to be ignored, instead of just the last character.
4405
4406 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4407 non-matching string, it can take a very, very long time, even for strings of
4408 quite modest length, because of the nested recursion. PCRE now does better in
4409 some of these cases. It does this by remembering the last required literal
4410 character in the pattern, and pre-searching the subject to ensure it is present
4411 before running the real match. In other words, it applies a heuristic to detect
4412 some types of certain failure quickly, and in the above example, if presented
4413 with a string that has no trailing " it gives "no match" very quickly.
4414
4415 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4416 other alternatives are tried instead.
4417
4418
4419 Version 2.06 09-Jun-99
4420 ----------------------
4421
4422 1. Change pcretest's output for amount of store used to show just the code
4423 space, because the remainder (the data block) varies in size between 32-bit and
4424 64-bit systems.
4425
4426 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4427 start matching at. This allows lookbehinds to work when searching for multiple
4428 occurrences in a string.
4429
4430 3. Added additional options to pcretest for testing multiple occurrences:
4431
4432 /+ outputs the rest of the string that follows a match
4433 /g loops for multiple occurrences, using the new startoffset argument
4434 /G loops for multiple occurrences by passing an incremented pointer
4435
4436 4. PCRE wasn't doing the "first character" optimization for patterns starting
4437 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4438 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4439 the letter 'x'. On long subject strings, this gives a significant speed-up.
4440
4441
4442 Version 2.05 21-Apr-99
4443 ----------------------
4444
4445 1. Changed the type of magic_number from int to long int so that it works
4446 properly on 16-bit systems.
4447
4448 2. Fixed a bug which caused patterns starting with .* not to work correctly
4449 when the subject string contained newline characters. PCRE was assuming
4450 anchoring for such patterns in all cases, which is not correct because .* will
4451 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4452 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4453 must be retried after every newline in the subject.
4454
4455
4456 Version 2.04 18-Feb-99
4457 ----------------------
4458
4459 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4460 computation of the store needed to hold the pattern was incorrect (too large).
4461 If such patterns were nested a few deep, this could multiply and become a real
4462 problem.
4463
4464 2. Added /M option to pcretest to show the memory requirement of a specific
4465 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4466
4467 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4468 compiled in such a way that the backtracking after subsequent failure was
4469 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4470 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4471
4472
4473 Version 2.03 02-Feb-99
4474 ----------------------
4475
4476 1. Fixed typo and small mistake in man page.
4477
4478 2. Added 4th condition (GPL supersedes if conflict) and created separate
4479 LICENCE file containing the conditions.
4480
4481 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4482 Perl, that is the internal \ allows the delimiter to be included in the
4483 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4484 the final delimiter, add \ to the end of the pattern (to test the error).
4485
4486 4. Added the convenience functions for extracting substrings after a successful
4487 match. Updated pcretest to make it able to test these functions.
4488
4489
4490 Version 2.02 14-Jan-99
4491 ----------------------
4492
4493 1. Initialized the working variables associated with each extraction so that
4494 their saving and restoring doesn't refer to uninitialized store.
4495
4496 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4497 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4498 fix the problem.
4499
4500 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4501 calls, and wasn't printing the correct value for compiling calls. Increased the
4502 default value of LOOPREPEAT, and the number of significant figures in the
4503 times.
4504
4505 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4506
4507 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4508 a building problem on Windows NT with a FAT file system.
4509
4510
4511 Version 2.01 21-Oct-98
4512 ----------------------
4513
4514 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4515 to character tables built by pcre_maketables() in the current locale. If NULL
4516 is passed, the default tables are used.
4517
4518
4519 Version 2.00 24-Sep-98
4520 ----------------------
4521
4522 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4523 it any more.
4524
4525 2. Allow quantification of (?>) groups, and make it work correctly.
4526
4527 3. The first character computation wasn't working for (?>) groups.
4528
4529 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4530 end of the subject) and add 5.005's \z, which really does match only at the
4531 very end of the subject.
4532
4533 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4534
4535 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4536 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4537 localized options. All options to pcre_study() were also removed.
4538
4539 7. Add other new features from 5.005:
4540
4541 $(?<= positive lookbehind
4542 $(?<! negative lookbehind
4543 (?imsx-imsx) added the unsetting capability
4544 such a setting is global if at outer level; local otherwise
4545 (?imsx-imsx:) non-capturing groups with option setting
4546 (?(cond)re|re) conditional pattern matching
4547
4548 A backreference to itself in a repeated group matches the previous
4549 captured string.
4550
4551 8. General tidying up of studying (both automatic and via "study")
4552 consequential on the addition of new assertions.
4553
4554 9. As in 5.005, unlimited repeated groups that could match an empty substring
4555 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4556 runtime if any iteration does actually match an empty substring.
4557
4558 10. Include the RunTest script in the distribution.
4559
4560 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4561 discrepancies, some of which were old and were also with respect to 5.004. They
4562 have now been fixed.
4563
4564
4565 Version 1.09 28-Apr-98
4566 ----------------------
4567
4568 1. A negated single character class followed by a quantifier with a minimum
4569 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4570 program crashes, or just wrong answers. This did not apply to negated classes
4571 containing more than one character, or to minima other than one.
4572
4573
4574 Version 1.08 27-Mar-98
4575 ----------------------
4576
4577 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4578
4579 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4580 latter must appear before anything that relies on it in the pattern.
4581
4582
4583 Version 1.07 16-Feb-98
4584 ----------------------
4585
4586 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4587 repeat of a potentially empty string).
4588
4589
4590 Version 1.06 23-Jan-98
4591 ----------------------
4592
4593 1. Added Markus Oberhumer's little patches for C++.
4594
4595 2. Literal strings longer than 255 characters were broken.
4596
4597
4598 Version 1.05 23-Dec-97
4599 ----------------------
4600
4601 1. Negated character classes containing more than one character were failing if
4602 PCRE_CASELESS was set at run time.
4603
4604
4605 Version 1.04 19-Dec-97
4606 ----------------------
4607
4608 1. Corrected the man page, where some "const" qualifiers had been omitted.
4609
4610 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4611 input syntax.
4612
4613 3. Fixed memory leak which occurred when a regex with back references was
4614 matched with an offsets vector that wasn't big enough. The temporary memory
4615 that is used in this case wasn't being freed if the match failed.
4616
4617 4. Tidied pcretest to ensure it frees memory that it gets.
4618
4619 5. Temporary memory was being obtained in the case where the passed offsets
4620 vector was exactly big enough.
4621
4622 6. Corrected definition of offsetof() from change 5 below.
4623
4624 7. I had screwed up change 6 below and broken the rules for the use of
4625 setjmp(). Now fixed.
4626
4627
4628 Version 1.03 18-Dec-97
4629 ----------------------
4630
4631 1. A erroneous regex with a missing opening parenthesis was correctly
4632 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4633 on some systems.
4634
4635 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4636 it was reported that one broken compiler failed on the former because "code" is
4637 also an independent variable.
4638
4639 3. The erroneous regex a[]b caused an array overrun reference.
4640
4641 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4642 fail on data ending with that character. (It was going on too far, and checking
4643 the next character, typically a binary zero.) This was specific to the
4644 optimized code for single-character negative classes.
4645
4646 5. Added a contributed patch from the TIN world which does the following:
4647
4648 + Add an undef for memmove, in case the the system defines a macro for it.
4649
4650 + Add a definition of offsetof(), in case there isn't one. (I don't know
4651 the reason behind this - offsetof() is part of the ANSI standard - but
4652 it does no harm).
4653
4654 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4655 most of the places where whitespace preceded '#'. I have given up and
4656 allowed the remaining 2 cases to be at the margin.
4657
4658 + Rename some variables in pcre to eliminate shadowing. This seems very
4659 pedantic, but does no harm, of course.
4660
4661 6. Moved the call to setjmp() into its own function, to get rid of warnings
4662 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4663
4664 7. Constructs such as \d{8,} were compiling into the equivalent of
4665 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4666 outcome, but in this particular case used more store than had been allocated,
4667 which caused the bug to be discovered because it threw up an internal error.
4668
4669 8. The debugging code in both pcre and pcretest for outputting the compiled
4670 form of a regex was going wrong in the case of back references followed by
4671 curly-bracketed repeats.
4672
4673
4674 Version 1.02 12-Dec-97
4675 ----------------------
4676
4677 1. Typos in pcre.3 and comments in the source fixed.
4678
4679 2. Applied a contributed patch to get rid of places where it used to remove
4680 'const' from variables, and fixed some signed/unsigned and uninitialized
4681 variable warnings.
4682
4683 3. Added the "runtest" target to Makefile.
4684
4685 4. Set default compiler flag to -O2 rather than just -O.
4686
4687
4688 Version 1.01 19-Nov-97
4689 ----------------------
4690
4691 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4692 like /([ab]*)*/, that is, for classes with more than one character in them.
4693
4694 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4695 as /((?>a*))*/ (a PCRE_EXTRA facility).
4696
4697
4698 Version 1.00 18-Nov-97
4699 ----------------------
4700
4701 1. Added compile-time macros to support systems such as SunOS4 which don't have
4702 memmove() or strerror() but have other things that can be used instead.
4703
4704 2. Arranged that "make clean" removes the executables.
4705
4706
4707 Version 0.99 27-Oct-97
4708 ----------------------
4709
4710 1. Fixed bug in code for optimizing classes with only one character. It was
4711 initializing a 32-byte map regardless, which could cause it to run off the end
4712 of the memory it had got.
4713
4714 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4715
4716
4717 Version 0.98 22-Oct-97
4718 ----------------------
4719
4720 1. Fixed bug in code for handling temporary memory usage when there are more
4721 back references than supplied space in the ovector. This could cause segfaults.
4722
4723
4724 Version 0.97 21-Oct-97
4725 ----------------------
4726
4727 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4728
4729 2. Optimized negated single characters not to use a bit map.
4730
4731 3. Brought error texts together as macro definitions; clarified some of them;
4732 fixed one that was wrong - it said "range out of order" when it meant "invalid
4733 escape sequence".
4734
4735 4. Changed some char * arguments to const char *.
4736
4737 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4738
4739 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4740 pcretest.
4741
4742
4743 Version 0.96 16-Oct-97
4744 ----------------------
4745
4746 1. Added a simple "pgrep" utility to the distribution.
4747
4748 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4749 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4750 where "ddd" means "one or more decimal digits".
4751
4752 3. Fixed serious bug. If a pattern had a back reference, but the call to
4753 pcre_exec() didn't supply a large enough ovector to record the related
4754 identifying subpattern, the match always failed. PCRE now remembers the number
4755 of the largest back reference, and gets some temporary memory in which to save
4756 the offsets during matching if necessary, in order to ensure that
4757 backreferences always work.
4758
4759 4. Increased the compatibility with Perl in a number of ways:
4760
4761 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4762 to request this handling. The option can be set at compile or exec time.
4763
4764 (b) $ matches before a terminating newline by default; an option
4765 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4766 mode). The option can be set at compile or exec time.
4767
4768 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4769 the same as Perl's. If the decimal number it represents is less than 10
4770 or there aren't that many previous left capturing parentheses, an octal
4771 escape is read. Inside a character class, it's always an octal escape,
4772 even if it is a single digit.
4773
4774 (d) An escaped but undefined alphabetic character is taken as a literal,
4775 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4776 escapes.
4777
4778 (e) {0} is now permitted. (The previous item is removed from the compiled
4779 pattern).
4780
4781 5. Changed all the names of code files so that the basic parts are no longer
4782 than 10 characters, and abolished the teeny "globals.c" file.
4783
4784 6. Changed the handling of character classes; they are now done with a 32-byte
4785 bit map always.
4786
4787 7. Added the -d and /D options to pcretest to make it possible to look at the
4788 internals of compilation without having to recompile pcre.
4789
4790
4791 Version 0.95 23-Sep-97
4792 ----------------------
4793
4794 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4795 \x20 at the start of a run of normal characters. These were being treated as
4796 real characters, instead of the source characters being re-checked.
4797
4798
4799 Version 0.94 18-Sep-97
4800 ----------------------
4801
4802 1. The functions are now thread-safe, with the caveat that the global variables
4803 containing pointers to malloc() and free() or alternative functions are the
4804 same for all threads.
4805
4806 2. Get pcre_study() to generate a bitmap of initial characters for non-
4807 anchored patterns when this is possible, and use it if passed to pcre_exec().
4808
4809
4810 Version 0.93 15-Sep-97
4811 ----------------------
4812
4813 1. /(b)|(:+)/ was computing an incorrect first character.
4814
4815 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4816 but not actually doing anything yet.
4817
4818 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4819 as Perl does (e.g. [-az] or [az-]).
4820
4821 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4822 all possible positions.
4823
4824 5. Split up into different modules to avoid including unneeded functions in a
4825 compiled binary. However, compile and exec are still in one module. The "study"
4826 function is split off.
4827
4828 6. The character tables are now in a separate module whose source is generated
4829 by an auxiliary program - but can then be edited by hand if required. There are
4830 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4831 toupper() in the code.
4832
4833 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4834 make them global. Abolish the function for setting them, as the caller can now
4835 set them directly.
4836
4837
4838 Version 0.92 11-Sep-97
4839 ----------------------
4840
4841 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4842 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4843
4844 2. Caseless matching was not working in character classes if the characters in
4845 the pattern were in upper case.
4846
4847 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4848
4849 4. Make PCRE_ANCHORED public and accept as a compile option.
4850
4851 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4852 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4853 pass them.
4854
4855 6. Give an error if bad option bits passed at compile or run time.
4856
4857 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4858 pcretest to cause it to pass that flag.
4859
4860 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4861 options, and the first character, if set.
4862
4863 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4864
4865
4866 Version 0.91 10-Sep-97
4867 ----------------------
4868
4869 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4870 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4871
4872 2. PCRE was looping on encountering an indefinitely repeated back reference to
4873 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4874 Perl does - treats the match as successful.
4875
4876 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5