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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1187 - (show annotations)
Mon Oct 29 11:30:45 2012 UTC (7 years, 1 month ago) by zherczeg
File size: 238694 byte(s)
JIT native interface.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.32
5 ------------
6
7 1. Improved JIT compiler optimizations for first character search and single
8 character iterators.
9
10 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11 Patch by Daniel Richard G.
12
13 3. Single character iterator optimizations in the JIT compiler.
14
15 4. Improved JIT compiler optimizations for character ranges.
16
17 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18 Reported by Giuseppe D'Angelo.
19
20 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21 of a line, was being set incorrectly in cases where .* appeared inside
22 atomic brackets at the start of a pattern, or where there was a subsequent
23 *PRUNE or *SKIP.
24
25 7. Improved instruction cache flush for POWER/PowerPC.
26 Patch by Daniel Richard G.
27
28 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
29 grep:
30
31 (a) There is now no limit to the number of patterns to be matched.
32
33 (b) An error is given if a pattern is too long.
34
35 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
36 are now supported.
37
38 (d) --exclude-from and --include-from (multiple use) have been added.
39
40 (e) Exclusions and inclusions now apply to all files and directories, not
41 just to those obtained from scanning a directory recursively.
42
43 (f) Multiple uses of -f and --file-list are now supported.
44
45 (g) In a Windows environment, the default for -d has been changed from
46 "read" (the GNU grep default) to "skip", because otherwise the presence
47 of a directory in the file list provokes an error.
48
49 (h) The documentation has been revised and clarified in places.
50
51 9. Improve the matching speed of capturing brackets.
52
53 10. Changed the meaning of \X so that it now matches a Unicode extended
54 grapheme cluster.
55
56 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
57 out POSIX threads when JIT support is configured.
58
59 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
60
61 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
62 structure could go wrong in environments where size_t is not the same size
63 as int.
64
65 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
66 set.
67
68 15. The EBCDIC support had decayed; later updates to the code had included
69 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
70 general tidy up of EBCDIC-related issues, and the documentation was also
71 not quite right. There is now a test that can be run on ASCII systems to
72 check some of the EBCDIC-related things (but is it not a full test).
73
74 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
75 in a small tidy to the code.
76
77 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
78
79 18. If the --only-matching (-o) option in pcregrep is specified multiple
80 times, each one causes appropriate output. For example, -o1 -o2 outputs the
81 substrings matched by the 1st and 2nd capturing parentheses. A separating
82 string can be specified by --om-separator (default empty).
83
84 19. Improving the first n character searches.
85
86 20. Turn case lists for horizontal and vertical white space into macros so that
87 they are defined only once.
88
89 21. This set of changes together give more compatible Unicode case-folding
90 behaviour for characters that have more than one other case when UCP
91 support is available.
92
93 (a) The Unicode property table now has offsets into a new table of sets of
94 three or more characters that are case-equivalent. The MultiStage2.py
95 script that generates these tables (the pcre_ucd.c file) now scans
96 CaseFolding.txt instead of UnicodeData.txt for character case
97 information.
98
99 (b) The code for adding characters or ranges of characters to a character
100 class has been abstracted into a generalized function that also handles
101 case-independence. In UTF-mode with UCP support, this uses the new data
102 to handle characters with more than one other case.
103
104 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
105 whose other case is greater than 256 are now correctly matched
106 caselessly. Previously, the high codepoint matched the low one, but not
107 vice versa.
108
109 (d) The processing of \h, \H, \v, and \ in character classes now makes use
110 of the new class addition function, using character lists defined as
111 macros alongside the case definitions of 20 above.
112
113 (e) Caseless back references now work with characters that have more than
114 one other case.
115
116 (f) General caseless matching of characters with more than one other case
117 is supported.
118
119 22. Unicode character properties were updated from Unicode 6.2.0
120
121 23. Improved CMake support under Windows. Patch by Daniel Richard G.
122
123 24. Add support for 32-bit character strings, and UTF-32
124
125 25. Major JIT compiler update (code refactoring and bugfixing).
126 Experimental Sparc 32 support is added.
127
128 26. Applied a modified version of Daniel Richard G's patch to create
129 pcre.h.generic and config.h.generic by "make" instead of in the
130 PrepareRelease script.
131
132 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
133 pcre_compile.c when checking for a zero character.
134
135 28. Introducing a native interface for JIT. Through this interface, the compiled
136 machine code can be directly executed. The purpose of this interface is to
137 provide fast pattern matching, so several sanity checks are not performed.
138 However, feature tests are still performed. The new interface provides
139 1.4x speedup compared to the old one.
140
141
142 Version 8.31 06-July-2012
143 -------------------------
144
145 1. Fixing a wrong JIT test case and some compiler warnings.
146
147 2. Removed a bashism from the RunTest script.
148
149 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
150 to unsigned type, result still unsigned" that was given by an MS compiler
151 on encountering the code "-sizeof(xxx)".
152
153 4. Partial matching support is added to the JIT compiler.
154
155 5. Fixed several bugs concerned with partial matching of items that consist
156 of more than one character:
157
158 (a) /^(..)\1/ did not partially match "aba" because checking references was
159 done on an "all or nothing" basis. This also applied to repeated
160 references.
161
162 (b) \R did not give a hard partial match if \r was found at the end of the
163 subject.
164
165 (c) \X did not give a hard partial match after matching one or more
166 characters at the end of the subject.
167
168 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
169 a partial match for the string "\r".
170
171 (e) When newline was set to CRLF, the metacharacter "." did not recognize
172 a partial match for a CR character at the end of the subject string.
173
174 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
175 running pcretest, the text "(JIT)" added to the output whenever JIT is
176 actually used to run the match.
177
178 7. Individual JIT compile options can be set in pcretest by following -s+[+]
179 or /S+[+] with a digit between 1 and 7.
180
181 8. OP_NOT now supports any UTF character not just single-byte ones.
182
183 9. (*MARK) control verb is now supported by the JIT compiler.
184
185 10. The command "./RunTest list" lists the available tests without actually
186 running any of them. (Because I keep forgetting what they all are.)
187
188 11. Add PCRE_INFO_MAXLOOKBEHIND.
189
190 12. Applied a (slightly modified) user-supplied patch that improves performance
191 when the heap is used for recursion (compiled with --disable-stack-for-
192 recursion). Instead of malloc and free for each heap frame each time a
193 logical recursion happens, frames are retained on a chain and re-used where
194 possible. This sometimes gives as much as 30% improvement.
195
196 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
197 call.
198
199 14. As documented, (*COMMIT) is now confined to within a positive assertion.
200
201 15. It is now possible to link pcretest with libedit as an alternative to
202 libreadline.
203
204 16. (*COMMIT) control verb is now supported by the JIT compiler.
205
206 17. The Unicode data tables have been updated to Unicode 6.1.0.
207
208 18. Added --file-list option to pcregrep.
209
210 19. Added binary file support to pcregrep, including the -a, --binary-files,
211 -I, and --text options.
212
213 20. The madvise function is renamed for posix_madvise for QNX compatibility
214 reasons. Fixed by Giuseppe D'Angelo.
215
216 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
217
218 22. Changed the option for creating symbolic links for 16-bit man pages from
219 -s to -sf so that re-installing does not cause issues.
220
221 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
222
223 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
224 matches in certain environments (the workspace was not being correctly
225 retained). Also added to pcre_dfa_exec() a simple plausibility check on
226 some of the workspace data at the beginning of a restart.
227
228 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
229 was not doing so when it should - probably a typo introduced by SVN 528
230 (change 8.10/14).
231
232 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
233 \w+ when the character tables indicated that \x{c4} was a word character.
234 There were several related cases, all because the tests for doing a table
235 lookup were testing for characters less than 127 instead of 255.
236
237 27. If a pattern contains capturing parentheses that are not used in a match,
238 their slots in the ovector are set to -1. For those that are higher than
239 any matched groups, this happens at the end of processing. In the case when
240 there were back references that the ovector was too small to contain
241 (causing temporary malloc'd memory to be used during matching), and the
242 highest capturing number was not used, memory off the end of the ovector
243 was incorrectly being set to -1. (It was using the size of the temporary
244 memory instead of the true size.)
245
246 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
247 ovector size, it uses memory at the end of the block that it has got.
248
249 29. Check for an overlong MARK name and give an error at compile time. The
250 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
251
252 30. JIT compiler update.
253
254 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
255 Rill for the patch.
256
257 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
258
259 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
260
261 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
262 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
263 was enabled.
264
265 35. Improve JIT code generation for greedy plus quantifier.
266
267 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
268 1 to "aa" instead of to an empty string. The bug affected repeated groups
269 that could potentially match an empty string.
270
271 37. Optimizing single character iterators in JIT.
272
273 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
274 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
275 codepoints that are too big for the mode are faulted, and in a UTF mode,
276 disallowed codepoints are also faulted.
277
278 39. If PCRE was compiled with UTF support, in three places in the DFA
279 matcher there was code that should only have been obeyed in UTF mode, but
280 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
281 processing when bytes with values greater than 127 were present. In 16-bit
282 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
283 both cases the values are those that cannot be the first data item in a UTF
284 character. The three items that might have provoked this were recursions,
285 possessively repeated groups, and atomic groups.
286
287 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
288 and pcregrep, because some OS require shared objects to be explicitly
289 passed to ld, causing the link step to fail if they are not.
290
291 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
292 mode, patterns that started with \h* or \R* might be incorrectly matched.
293
294
295 Version 8.30 04-February-2012
296 -----------------------------
297
298 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
299 name is defined in ctype.h.
300
301 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
302 only in quite long subpatterns.
303
304 3. Removed the function pcre_info(), which has been obsolete and deprecated
305 since it was replaced by pcre_fullinfo() in February 2000.
306
307 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
308 match a (*MARK), and the match failed at the start of the subject, a
309 reference to memory before the start of the subject could occur. This bug
310 was introduced by fix 17 of release 8.21.
311
312 5. A reference to an unset group with zero minimum repetition was giving
313 totally wrong answers (in non-JavaScript-compatibility mode). For example,
314 /(another)?(\1?)test/ matched against "hello world test". This bug was
315 introduced in release 8.13.
316
317 6. Add support for 16-bit character strings (a large amount of work involving
318 many changes and refactorings).
319
320 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
321 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
322 from a file.
323
324 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
325 rounding is not applied in this particular case).
326
327 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
328 if they appear, or are escaped, in patterns.
329
330 10. Get rid of a number of -Wunused-but-set-variable warnings.
331
332 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
333 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
334 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
335 also returns the mark "x". This bug applied to capturing parentheses,
336 non-capturing parentheses, and atomic parentheses. It also applied to some
337 assertions.
338
339 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
340 information out of configure.ac instead of relying on pcre.h.generic, which
341 is not stored in the repository.
342
343 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
344 -lreadline.
345
346 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
347
348 15. Applied Graycode's patch to put the top-level frame on the stack rather
349 than the heap when not using the stack for recursion. This gives a
350 performance improvement in many cases when recursion is not deep.
351
352 16. Experimental code added to "pcretest -C" to output the stack frame size.
353
354
355 Version 8.21 12-Dec-2011
356 ------------------------
357
358 1. Updating the JIT compiler.
359
360 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
361 are added as well.
362
363 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
364 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
365 calling _pcre_jit_exec. Some extra comments are added.
366
367 4. (*MARK) settings inside atomic groups that do not contain any capturing
368 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
369 was introduced by change 18 for 8.20.
370
371 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
372 ECMA-262 standard.
373
374 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
375 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
376 This bug was probably introduced by change 9 of 8.13.
377
378 7. While fixing 6 above, I noticed that a number of other items were being
379 incorrectly rejected as "not fixed length". This arose partly because newer
380 opcodes had not been added to the fixed-length checking code. I have (a)
381 corrected the bug and added tests for these items, and (b) arranged for an
382 error to occur if an unknown opcode is encountered while checking for fixed
383 length instead of just assuming "not fixed length". The items that were
384 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
385 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
386 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
387
388 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
389 being incorrectly compiled and would have given unpredicatble results.
390
391 9. A possessively repeated subpattern with minimum repeat count greater than
392 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
393 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
394 the first (A) could occur when it should not.
395
396 10. Add a cast and remove a redundant test from the code.
397
398 11. JIT should use pcre_malloc/pcre_free for allocation.
399
400 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
401 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
402 is anything other than /usr, -L is still shown).
403
404 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
405
406 14. Perl does not support \N without a following name in a [] class; PCRE now
407 also gives an error.
408
409 15. If a forward reference was repeated with an upper limit of around 2000,
410 it caused the error "internal error: overran compiling workspace". The
411 maximum number of forward references (including repeats) was limited by the
412 internal workspace, and dependent on the LINK_SIZE. The code has been
413 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
414 the default depends on LINK_SIZE. There is a new upper limit (for safety)
415 of around 200,000 forward references. While doing this, I also speeded up
416 the filling in of repeated forward references.
417
418 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
419 incorrectly expecting the subject to contain another "a" after the start.
420
421 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
422 in the match, the SKIP should be ignored. This was not happening; instead
423 the SKIP was being treated as NOMATCH. For patterns such as
424 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
425 tested.
426
427 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
428 now much more compatible with Perl, in particular in cases where the result
429 is a non-match for a non-anchored pattern. For example, if
430 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
431 "m", where previously it did not return a name. A side effect of this
432 change is that for partial matches, the last encountered mark name is
433 returned, as for non matches. A number of tests that were previously not
434 Perl-compatible have been moved into the Perl-compatible test files. The
435 refactoring has had the pleasing side effect of removing one argument from
436 the match() function, thus reducing its stack requirements.
437
438 19. If the /S+ option was used in pcretest to study a pattern using JIT,
439 subsequent uses of /S (without +) incorrectly behaved like /S+.
440
441 21. Retrieve executable code size support for the JIT compiler and fixing
442 some warnings.
443
444 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
445 not work when the shorter character appeared right at the end of the
446 subject string.
447
448 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
449 systems.
450
451 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
452 output it when the /M option is used in pcretest.
453
454 25. The CheckMan script was not being included in the distribution. Also, added
455 an explicit "perl" to run Perl scripts from the PrepareRelease script
456 because this is reportedly needed in Windows.
457
458 26. If study data was being save in a file and studying had not found a set of
459 "starts with" bytes for the pattern, the data written to the file (though
460 never used) was taken from uninitialized memory and so caused valgrind to
461 complain.
462
463 27. Updated RunTest.bat as provided by Sheri Pierce.
464
465 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
466
467 29. Computation of memory usage for the table of capturing group names was
468 giving an unnecessarily large value.
469
470
471 Version 8.20 21-Oct-2011
472 ------------------------
473
474 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
475 a POSIX class. After further experiments with Perl, which convinced me that
476 Perl has bugs and confusions, a closing square bracket is no longer allowed
477 in a POSIX name. This bug also affected patterns with classes that started
478 with full stops.
479
480 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
481 captured substring, but while checking the failing first alternative,
482 substring 1 is temporarily captured. If the output vector supplied to
483 pcre_exec() was not big enough for this capture, the yield of the function
484 was still zero ("insufficient space for captured substrings"). This cannot
485 be totally fixed without adding another stack variable, which seems a lot
486 of expense for a edge case. However, I have improved the situation in cases
487 such as /(a)(b)x|abc/ matched against "abc", where the return code
488 indicates that fewer than the maximum number of slots in the ovector have
489 been set.
490
491 3. Related to (2) above: when there are more back references in a pattern than
492 slots in the output vector, pcre_exec() uses temporary memory during
493 matching, and copies in the captures as far as possible afterwards. It was
494 using the entire output vector, but this conflicts with the specification
495 that only 2/3 is used for passing back captured substrings. Now it uses
496 only the first 2/3, for compatibility. This is, of course, another edge
497 case.
498
499 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
500 main code base, and can be used by building with --enable-jit. When this is
501 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
502 runtime --no-jit option is given.
503
504 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
505 ovector, the return from the function was zero, implying that there were
506 other matches that did not fit. The correct "exactly full" value is now
507 returned.
508
509 6. If a subpattern that was called recursively or as a subroutine contained
510 (*PRUNE) or any other control that caused it to give a non-standard return,
511 invalid errors such as "Error -26 (nested recursion at the same subject
512 position)" or even infinite loops could occur.
513
514 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
515 computing the minimum length on reaching *ACCEPT, and so ended up with the
516 wrong value of 1 rather than 0. Further investigation indicates that
517 computing a minimum subject length in the presence of *ACCEPT is difficult
518 (think back references, subroutine calls), and so I have changed the code
519 so that no minimum is registered for a pattern that contains *ACCEPT.
520
521 8. If (*THEN) was present in the first (true) branch of a conditional group,
522 it was not handled as intended. [But see 16 below.]
523
524 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
525 Sheri Pierce.
526
527 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
528 the first byte in a match must be "a".
529
530 11. Change 17 for 8.13 increased the recursion depth for patterns like
531 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
532 pattern contains any instances of (*THEN). If it does not, the old
533 optimizations are restored. It would be nice to do this on a per-group
534 basis, but at the moment that is not feasible.
535
536 12. In some environments, the output of pcretest -C is CRLF terminated. This
537 broke RunTest's code that checks for the link size. A single white space
538 character after the value is now allowed for.
539
540 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
541 For "fr", it uses the Windows-specific input and output files.
542
543 14. If (*THEN) appeared in a group that was called recursively or as a
544 subroutine, it did not work as intended. [But see next item.]
545
546 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
547 pattern fragments (but not containing any | characters). If A and B are
548 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
549 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
550 D. In other words, Perl considers parentheses that do not contain any |
551 characters to be part of a surrounding alternative, whereas PCRE was
552 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
553 differently. PCRE now behaves in the same way as Perl, except in the case
554 of subroutine/recursion calls such as (?1) which have in any case always
555 been different (but PCRE had them first :-).
556
557 16. Related to 15 above: Perl does not treat the | in a conditional group as
558 creating alternatives. Such a group is treated in the same way as an
559 ordinary group without any | characters when processing (*THEN). PCRE has
560 been changed to match Perl's behaviour.
561
562 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
563 RunGrepTest script failed.
564
565 18. Change 22 for version 13 caused atomic groups to use more stack. This is
566 inevitable for groups that contain captures, but it can lead to a lot of
567 stack use in large patterns. The old behaviour has been restored for atomic
568 groups that do not contain any capturing parentheses.
569
570 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
571 suppress the check for a minimum subject length at run time. (If it was
572 given to pcre_exec() or pcre_dfa_exec() it did work.)
573
574 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
575 fail to work when decoding hex characters in data strings in EBCDIC
576 environments.
577
578 21. It appears that in at least one Mac OS environment, the isxdigit() function
579 is implemented as a macro that evaluates to its argument more than once,
580 contravening the C 90 Standard (I haven't checked a later standard). There
581 was an instance in pcretest which caused it to go wrong when processing
582 \x{...} escapes in subject strings. The has been rewritten to avoid using
583 things like p++ in the argument of isxdigit().
584
585
586 Version 8.13 16-Aug-2011
587 ------------------------
588
589 1. The Unicode data tables have been updated to Unicode 6.0.0.
590
591 2. Two minor typos in pcre_internal.h have been fixed.
592
593 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
594 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
595 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
596
597 4. There were a number of related bugs in the code for matching backrefences
598 caselessly in UTF-8 mode when codes for the characters concerned were
599 different numbers of bytes. For example, U+023A and U+2C65 are an upper
600 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
601 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
602 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
603 2-byte code at the end of the subject (it thought there wasn't enough data
604 left).
605
606 5. Comprehensive information about what went wrong is now returned by
607 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
608 as the output vector has at least 2 elements. The offset of the start of
609 the failing character and a reason code are placed in the vector.
610
611 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
612 now returned is for the first byte of the failing character, instead of the
613 last byte inspected. This is an incompatible change, but I hope it is small
614 enough not to be a problem. It makes the returned offset consistent with
615 pcre_exec() and pcre_dfa_exec().
616
617 7. pcretest now gives a text phrase as well as the error number when
618 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
619 failure, the offset and reason code are output.
620
621 8. When \R was used with a maximizing quantifier it failed to skip backwards
622 over a \r\n pair if the subsequent match failed. Instead, it just skipped
623 back over a single character (\n). This seems wrong (because it treated the
624 two characters as a single entity when going forwards), conflicts with the
625 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
626 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
627 has been changed.
628
629 9. Some internal refactoring has changed the processing so that the handling
630 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
631 time (the PCRE_DOTALL option was changed this way some time ago: version
632 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
633 which was always a bit of a fudge. It also means that there is one less
634 argument for the match() function, which reduces its stack requirements
635 slightly. This change also fixes an incompatibility with Perl: the pattern
636 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
637
638 10. More internal refactoring has drastically reduced the number of recursive
639 calls to match() for possessively repeated groups such as (abc)++ when
640 using pcre_exec().
641
642 11. While implementing 10, a number of bugs in the handling of groups were
643 discovered and fixed:
644
645 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
646 (a|)*(?1) gave a compile-time internal error.
647 ((a|)+)+ did not notice that the outer group could match an empty string.
648 (^a|^)+ was not marked as anchored.
649 (.*a|.*)+ was not marked as matching at start or after a newline.
650
651 12. Yet more internal refactoring has removed another argument from the match()
652 function. Special calls to this function are now indicated by setting a
653 value in a variable in the "match data" data block.
654
655 13. Be more explicit in pcre_study() instead of relying on "default" for
656 opcodes that mean there is no starting character; this means that when new
657 ones are added and accidentally left out of pcre_study(), testing should
658 pick them up.
659
660 14. The -s option of pcretest has been documented for ages as being an old
661 synonym of -m (show memory usage). I have changed it to mean "force study
662 for every regex", that is, assume /S for every regex. This is similar to -i
663 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
664 using it. It makes it easier to run collections of tests with and without
665 study enabled, and thereby test pcre_study() more easily. All the standard
666 tests are now run with and without -s (but some patterns can be marked as
667 "never study" - see 20 below).
668
669 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
670 restoration of the capturing data to the outer values was not happening
671 correctly.
672
673 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
674 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
675 pattern had matched an empty string, and so incorrectly returned a no
676 match.
677
678 17. There was optimizing code for the last branch of non-capturing parentheses,
679 and also for the obeyed branch of a conditional subexpression, which used
680 tail recursion to cut down on stack usage. Unfortunately, now that there is
681 the possibility of (*THEN) occurring in these branches, tail recursion is
682 no longer possible because the return has to be checked for (*THEN). These
683 two optimizations have therefore been removed. [But see 8.20/11 above.]
684
685 18. If a pattern containing \R was studied, it was assumed that \R always
686 matched two bytes, thus causing the minimum subject length to be
687 incorrectly computed because \R can also match just one byte.
688
689 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
690 was incorrectly computed.
691
692 20. If /S is present twice on a test pattern in pcretest input, it now
693 *disables* studying, thereby overriding the use of -s on the command line
694 (see 14 above). This is necessary for one or two tests to keep the output
695 identical in both cases.
696
697 21. When (*ACCEPT) was used in an assertion that matched an empty string and
698 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
699
700 22. When an atomic group that contained a capturing parenthesis was
701 successfully matched, but the branch in which it appeared failed, the
702 capturing was not being forgotten if a higher numbered group was later
703 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
704 group 1 to "a", when in fact it should be unset. This applied to multi-
705 branched capturing and non-capturing groups, repeated or not, and also to
706 positive assertions (capturing in negative assertions does not happen
707 in PCRE) and also to nested atomic groups.
708
709 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
710 subject after a captured substring, to make it easier to tell which of a
711 number of identical substrings has been captured.
712
713 24. The way atomic groups are processed by pcre_exec() has been changed so that
714 if they are repeated, backtracking one repetition now resets captured
715 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
716 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
717 "aaa". Previously, it would have been "a". As part of this code
718 refactoring, the way recursive calls are handled has also been changed.
719
720 25. If an assertion condition captured any substrings, they were not passed
721 back unless some other capturing happened later. For example, if
722 (?(?=(a))a) was matched against "a", no capturing was returned.
723
724 26. When studying a pattern that contained subroutine calls or assertions,
725 the code for finding the minimum length of a possible match was handling
726 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
727 group 1 called group 2 while simultaneously a separate group 2 called group
728 1). A stack overflow occurred in this case. I have fixed this by limiting
729 the recursion depth to 10.
730
731 27. Updated RunTest.bat in the distribution to the version supplied by Tom
732 Fortmann. This supports explicit test numbers on the command line, and has
733 argument validation and error reporting.
734
735 28. An instance of \X with an unlimited repeat could fail if at any point the
736 first character it looked at was a mark character.
737
738 29. Some minor code refactoring concerning Unicode properties and scripts
739 should reduce the stack requirement of match() slightly.
740
741 30. Added the '=' option to pcretest to check the setting of unused capturing
742 slots at the end of the pattern, which are documented as being -1, but are
743 not included in the return count.
744
745 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
746 compiled something random. Now it gives a compile-time error (as does
747 Perl).
748
749 32. A *MARK encountered during the processing of a positive assertion is now
750 recorded and passed back (compatible with Perl).
751
752 33. If --only-matching or --colour was set on a pcregrep call whose pattern
753 had alternative anchored branches, the search for a second match in a line
754 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
755 matched the line "0102" twice. The same bug affected patterns that started
756 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
757 twice.
758
759 34. Previously, PCRE did not allow quantification of assertions. However, Perl
760 does, and because of capturing effects, quantifying parenthesized
761 assertions may at times be useful. Quantifiers are now allowed for
762 parenthesized assertions.
763
764 35. A minor code tidy in pcre_compile() when checking options for \R usage.
765
766 36. \g was being checked for fancy things in a character class, when it should
767 just be a literal "g".
768
769 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
770 appearance of a nested POSIX class supersedes an apparent external class.
771 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
772 unescaped square brackets may also appear as part of class names. For
773 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
774 more like Perl. (But see 8.20/1 above.)
775
776 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
777 was because it thought it was \N{name}, which is not supported).
778
779 39. Add minix to OS list not supporting the -S option in pcretest.
780
781 40. PCRE tries to detect cases of infinite recursion at compile time, but it
782 cannot analyze patterns in sufficient detail to catch mutual recursions
783 such as ((?1))((?2)). There is now a runtime test that gives an error if a
784 subgroup is called recursively as a subpattern for a second time at the
785 same position in the subject string. In previous releases this might have
786 been caught by the recursion limit, or it might have run out of stack.
787
788 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
789 happen only once. PCRE was, however incorrectly giving a compile time error
790 "recursive call could loop indefinitely" because it cannot analyze the
791 pattern in sufficient detail. The compile time test no longer happens when
792 PCRE is compiling a conditional subpattern, but actual runaway loops are
793 now caught at runtime (see 40 above).
794
795 42. It seems that Perl allows any characters other than a closing parenthesis
796 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
797 has been changed to be the same.
798
799 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
800 as not to get warnings when autogen.sh is called. Also changed
801 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
802
803 44. To help people who use pcregrep to scan files containing exceedingly long
804 lines, the following changes have been made:
805
806 (a) The default value of the buffer size parameter has been increased from
807 8K to 20K. (The actual buffer used is three times this size.)
808
809 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
810 PCRE is built.
811
812 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
813 to be set at run time.
814
815 (d) Numerical values in pcregrep options can be followed by K or M, for
816 example --buffer-size=50K.
817
818 (e) If a line being scanned overflows pcregrep's buffer, an error is now
819 given and the return code is set to 2.
820
821 45. Add a pointer to the latest mark to the callout data block.
822
823 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
824 partial match of an empty string instead of no match. This was specific to
825 the use of ".".
826
827 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
828 complete match instead of a partial match. This bug was dependent on both
829 the PCRE_UTF8 and PCRE_DOTALL options being set.
830
831 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
832 starting byte set, because \b was not being ignored.
833
834
835 Version 8.12 15-Jan-2011
836 ------------------------
837
838 1. Fixed some typos in the markup of the man pages, and wrote a script that
839 checks for such things as part of the documentation building process.
840
841 2. On a big-endian 64-bit system, pcregrep did not correctly process the
842 --match-limit and --recursion-limit options (added for 8.11). In
843 particular, this made one of the standard tests fail. (The integer value
844 went into the wrong half of a long int.)
845
846 3. If the --colour option was given to pcregrep with -v (invert match), it
847 did strange things, either producing crazy output, or crashing. It should,
848 of course, ignore a request for colour when reporting lines that do not
849 match.
850
851 4. Another pcregrep bug caused similar problems if --colour was specified with
852 -M (multiline) and the pattern match finished with a line ending.
853
854 5. In pcregrep, when a pattern that ended with a literal newline sequence was
855 matched in multiline mode, the following line was shown as part of the
856 match. This seems wrong, so I have changed it.
857
858 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
859 the check for further matches in the same line (so they could be coloured)
860 to overrun the end of the current line. If another match was found, it was
861 incorrectly shown (and then shown again when found in the next line).
862
863 7. If pcregrep was compiled under Windows, there was a reference to the
864 function pcregrep_exit() before it was defined. I am assuming this was
865 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
866 reported by a user. I've moved the definition above the reference.
867
868
869 Version 8.11 10-Dec-2010
870 ------------------------
871
872 1. (*THEN) was not working properly if there were untried alternatives prior
873 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
874 backtracked to try for "b" instead of moving to the next alternative branch
875 at the same level (in this case, to look for "c"). The Perl documentation
876 is clear that when (*THEN) is backtracked onto, it goes to the "next
877 alternative in the innermost enclosing group".
878
879 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
880 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
881 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
882 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
883 (*THEN).
884
885 3. If \s appeared in a character class, it removed the VT character from
886 the class, even if it had been included by some previous item, for example
887 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
888 of \s, but is part of the POSIX "space" class.)
889
890 4. A partial match never returns an empty string (because you can always
891 match an empty string at the end of the subject); however the checking for
892 an empty string was starting at the "start of match" point. This has been
893 changed to the "earliest inspected character" point, because the returned
894 data for a partial match starts at this character. This means that, for
895 example, /(?<=abc)def/ gives a partial match for the subject "abc"
896 (previously it gave "no match").
897
898 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
899 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
900 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
901 has an implication that the given string is incomplete (because a partial
902 match is preferred over a full match). For this reason, these items now
903 give a partial match in this situation. [Aside: previously, the one case
904 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
905 match rather than a full match, which was wrong by the old rules, but is
906 now correct.]
907
908 6. There was a bug in the handling of #-introduced comments, recognized when
909 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
910 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
911 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
912 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
913 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
914 places in pcre_compile().
915
916 7. Related to (6) above, when pcre_compile() was skipping #-introduced
917 comments when looking ahead for named forward references to subpatterns,
918 the only newline sequence it recognized was NL. It now handles newlines
919 according to the set newline convention.
920
921 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
922 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
923 cater for a lack of strerror(). These oversights have been fixed.
924
925 9. Added --match-limit and --recursion-limit to pcregrep.
926
927 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
928
929 11. When the -o option was used, pcregrep was setting a return code of 1, even
930 when matches were found, and --line-buffered was not being honoured.
931
932 12. Added an optional parentheses number to the -o and --only-matching options
933 of pcregrep.
934
935 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
936 can match an empty string. The code to do it in pcretest and pcredemo
937 needed fixing:
938
939 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
940 only one byte after an empty string match just before CRLF (this case
941 just got forgotten; "any" and "anycrlf" were OK).
942
943 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
944 mode when an empty string match preceded an ASCII character followed by
945 a non-ASCII character. (The code for advancing by one character rather
946 than one byte was nonsense.)
947
948 (c) The pcredemo.c sample program did not have any code at all to handle
949 the cases when CRLF is a valid newline sequence.
950
951 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
952 as a starting offset was within the subject string. There is now a new
953 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
954 negative or greater than the length of the string. In order to test this,
955 pcretest is extended to allow the setting of negative starting offsets.
956
957 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
958 starting offset points to the beginning of a UTF-8 character was
959 unnecessarily clumsy. I tidied it up.
960
961 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
962 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
963
964 17. Nobody had reported that the --include_dir option, which was added in
965 release 7.7 should have been called --include-dir (hyphen, not underscore)
966 for compatibility with GNU grep. I have changed it to --include-dir, but
967 left --include_dir as an undocumented synonym, and the same for
968 --exclude-dir, though that is not available in GNU grep, at least as of
969 release 2.5.4.
970
971 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
972 characters from a string of bytes) have been redefined so as not to use
973 loops, in order to improve performance in some environments. At the same
974 time, I abstracted some of the common code into auxiliary macros to save
975 repetition (this should not affect the compiled code).
976
977 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
978 compile-time error is now given if \c is not followed by an ASCII
979 character, that is, a byte less than 128. (In EBCDIC mode, the code is
980 different, and any byte value is allowed.)
981
982 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
983 START_OPTIMIZE option, which is now allowed at compile time - but just
984 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
985 to pcregrep and other applications that have no direct access to PCRE
986 options. The new /Y option in pcretest sets this option when calling
987 pcre_compile().
988
989 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
990 back references. Groups containing recursive back references were forced to
991 be atomic by that change, but in the case of named groups, the amount of
992 memory required was incorrectly computed, leading to "Failed: internal
993 error: code overflow". This has been fixed.
994
995 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
996 pcretest.c, to avoid build problems in some Borland environments.
997
998
999 Version 8.10 25-Jun-2010
1000 ------------------------
1001
1002 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1003 THEN.
1004
1005 2. (*ACCEPT) was not working when inside an atomic group.
1006
1007 3. Inside a character class, \B is treated as a literal by default, but
1008 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1009 causes the error). The code is unchanged, but I tidied the documentation.
1010
1011 4. Inside a character class, PCRE always treated \R and \X as literals,
1012 whereas Perl faults them if its -w option is set. I have changed PCRE so
1013 that it faults them when PCRE_EXTRA is set.
1014
1015 5. Added support for \N, which always matches any character other than
1016 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1017
1018 6. When compiling pcregrep with newer versions of gcc which may have
1019 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1020 declared with attribute warn_unused_result" were given. Just casting the
1021 result to (void) does not stop the warnings; a more elaborate fudge is
1022 needed. I've used a macro to implement this.
1023
1024 7. Minor change to pcretest.c to avoid a compiler warning.
1025
1026 8. Added four artifical Unicode properties to help with an option to make
1027 \s etc use properties (see next item). The new properties are: Xan
1028 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1029
1030 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1031 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1032 this option. Modified pcretest to add /W to test this facility. Added
1033 REG_UCP to make it available via the POSIX interface.
1034
1035 10. Added --line-buffered to pcregrep.
1036
1037 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1038 studied, and the match started with a letter with a code point greater than
1039 127 whose first byte was different to the first byte of the other case of
1040 the letter, the other case of this starting letter was not recognized
1041 (#976).
1042
1043 12. If a pattern that was studied started with a repeated Unicode property
1044 test, for example, \p{Nd}+, there was the theoretical possibility of
1045 setting up an incorrect bitmap of starting bytes, but fortunately it could
1046 not have actually happened in practice until change 8 above was made (it
1047 added property types that matched character-matching opcodes).
1048
1049 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1050 possible starting bytes for non-anchored patterns.
1051
1052 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1053 \R, and also a number of cases that involve Unicode properties, both
1054 explicit and implicit when PCRE_UCP is set.
1055
1056 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1057 input, it could crash or give wrong results if characters with values
1058 greater than 0xc0 were present in the subject string. (Detail: it assumed
1059 UTF-8 input when processing these items.)
1060
1061 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1062 size_t is 64-bit (#991).
1063
1064 17. Added a check for running out of memory when PCRE is compiled with
1065 --disable-stack-for-recursion (#990).
1066
1067 18. If the last data line in a file for pcretest does not have a newline on
1068 the end, a newline was missing in the output.
1069
1070 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1071 less than 128) in its various bitmaps. However, there is a facility for
1072 generating tables according to the current locale when PCRE is compiled. It
1073 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1074 characters, are recognized by isspace() and therefore were getting set in
1075 these tables, and indeed these tables seem to approximate to ISO 8859. This
1076 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1077 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1078 which of course cannot start UTF-8 characters. I have changed the code so
1079 that only real ASCII characters (less than 128) and the correct starting
1080 bytes for UTF-8 encodings are set for characters greater than 127 when in
1081 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1082 altogether.)
1083
1084 20. Added the /T option to pcretest so as to be able to run tests with non-
1085 standard character tables, thus making it possible to include the tests
1086 used for 19 above in the standard set of tests.
1087
1088 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1089 reference to a subpattern the other side of a comment that contains an
1090 opening parenthesis caused either an internal compiling error, or a
1091 reference to the wrong subpattern.
1092
1093
1094 Version 8.02 19-Mar-2010
1095 ------------------------
1096
1097 1. The Unicode data tables have been updated to Unicode 5.2.0.
1098
1099 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1100 configured.
1101
1102 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1103 original author of that file, following a query about its status.
1104
1105 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1106 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1107
1108 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1109 quantifier applied to a forward-referencing subroutine call, could compile
1110 incorrect code or give the error "internal error: previously-checked
1111 referenced subpattern not found".
1112
1113 6. Both MS Visual Studio and Symbian OS have problems with initializing
1114 variables to point to external functions. For these systems, therefore,
1115 pcre_malloc etc. are now initialized to local functions that call the
1116 relevant global functions.
1117
1118 7. There were two entries missing in the vectors called coptable and poptable
1119 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1120 I've fixed the data, and added a kludgy way of testing at compile time that
1121 the lengths are correct (equal to the number of opcodes).
1122
1123 8. Following on from 7, I added a similar kludge to check the length of the
1124 eint vector in pcreposix.c.
1125
1126 9. Error texts for pcre_compile() are held as one long string to avoid too
1127 much relocation at load time. To find a text, the string is searched,
1128 counting zeros. There was no check for running off the end of the string,
1129 which could happen if a new error number was added without updating the
1130 string.
1131
1132 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1133
1134 11. \K was not working if it appeared in an atomic group or in a group that
1135 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1136 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1137 the assertion is positive, but not if it is negative.
1138
1139 12. Change 11 fortuitously reduced the size of the stack frame used in the
1140 "match()" function of pcre_exec.c by one pointer. Forthcoming
1141 implementation of support for (*MARK) will need an extra pointer on the
1142 stack; I have reserved it now, so that the stack frame size does not
1143 decrease.
1144
1145 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1146 item in branch that calls a recursion is a subroutine call - as in the
1147 second branch in the above example - was incorrectly given the compile-
1148 time error "recursive call could loop indefinitely" because pcre_compile()
1149 was not correctly checking the subroutine for matching a non-empty string.
1150
1151 14. The checks for overrunning compiling workspace could trigger after an
1152 overrun had occurred. This is a "should never occur" error, but it can be
1153 triggered by pathological patterns such as hundreds of nested parentheses.
1154 The checks now trigger 100 bytes before the end of the workspace.
1155
1156 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1157
1158
1159 Version 8.01 19-Jan-2010
1160 ------------------------
1161
1162 1. If a pattern contained a conditional subpattern with only one branch (in
1163 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1164 computed the wrong minimum data length (which is of course zero for such
1165 subpatterns). This could cause incorrect "no match" results.
1166
1167 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1168 the pattern is reset in the first branch, pcre_compile() failed with
1169 "internal error: code overflow at offset...". This happened only when
1170 the reset was to the original external option setting. (An optimization
1171 abstracts leading options settings into an external setting, which was the
1172 cause of this.)
1173
1174 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1175 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1176 assertion pattern did not match (meaning that the assertion was true), it
1177 was incorrectly treated as false if the SKIP had been reached during the
1178 matching. This also applied to assertions used as conditions.
1179
1180 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1181 assertion subpattern, including such a pattern used as a condition,
1182 unpredictable results occurred, instead of the error return
1183 PCRE_ERROR_DFA_UITEM.
1184
1185 5. The C++ GlobalReplace function was not working like Perl for the special
1186 situation when an empty string is matched. It now does the fancy magic
1187 stuff that is necessary.
1188
1189 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1190 removed. (These were left over from very, very early versions of PCRE.)
1191
1192 7. Some cosmetic changes to the code to make life easier when compiling it
1193 as part of something else:
1194
1195 (a) Change DEBUG to PCRE_DEBUG.
1196
1197 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1198 called "current" as "current_branch", to prevent a collision with the
1199 Linux macro when compiled as a kernel module.
1200
1201 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1202 prevent a collision with the Linux macro when compiled as a kernel
1203 module.
1204
1205 8. In pcre_compile() there are some checks for integer overflows that used to
1206 cast potentially large values to (double). This has been changed to that
1207 when building, a check for int64_t is made, and if it is found, it is used
1208 instead, thus avoiding the use of floating point arithmetic. (There is no
1209 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1210 double.
1211
1212 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1213 2005 (difference between two addresses compared to an unsigned value).
1214
1215 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1216 custom one, because of the following reported problem in Windows:
1217
1218 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1219 under Win32.
1220 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1221 therefore missing the function definition.
1222 - The compiler thus generates a "C" signature for the test function.
1223 - The linker fails to find the "C" function.
1224 - PCRE fails to configure if asked to do so against libbz2.
1225
1226 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1227 messages were output:
1228
1229 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1230 rerunning libtoolize, to keep the correct libtool macros in-tree.
1231 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1232
1233 I have done both of these things.
1234
1235 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1236 most of the time, it *can* run out if it is given a pattern that contains a
1237 runaway infinite recursion. I updated the discussion in the pcrestack man
1238 page.
1239
1240 13. Now that we have gone to the x.xx style of version numbers, the minor
1241 version may start with zero. Using 08 or 09 is a bad idea because users
1242 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1243 interpreted as invalid octal numbers. I've updated the previous comment in
1244 configure.ac, and also added a check that gives an error if 08 or 09 are
1245 used.
1246
1247 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1248 causing partial matching to fail when the end of the subject matched \W
1249 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1250
1251 15. There were some discrepancies between the declarations in pcre_internal.h
1252 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1253 their definitions. The declarations used "const uschar *" and the
1254 definitions used USPTR. Even though USPTR is normally defined as "const
1255 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1256 reported that: "This difference in casting confuses some C++ compilers, for
1257 example, SunCC recognizes above declarations as different functions and
1258 generates broken code for hbpcre." I have changed the declarations to use
1259 USPTR.
1260
1261 16. GNU libtool is named differently on some systems. The autogen.sh script now
1262 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1263 (FreeBSD).
1264
1265 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1266 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1267 comment: "Figure out how to create a longlong from a string: strtoll and
1268 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1269 instance, but it only takes 2 args instead of 3!"
1270
1271 18. A subtle bug concerned with back references has been fixed by a change of
1272 specification, with a corresponding code fix. A pattern such as
1273 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1274 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1275 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1276 same bug. Such groups have to be quantified to be useful, or contained
1277 inside another quantified group. (If there's no repetition, the reference
1278 can never match.) The problem arises because, having left the group and
1279 moved on to the rest of the pattern, a later failure that backtracks into
1280 the group uses the captured value from the final iteration of the group
1281 rather than the correct earlier one. I have fixed this in PCRE by forcing
1282 any group that contains a reference to itself to be an atomic group; that
1283 is, there cannot be any backtracking into it once it has completed. This is
1284 similar to recursive and subroutine calls.
1285
1286
1287 Version 8.00 19-Oct-09
1288 ----------------------
1289
1290 1. The table for translating pcre_compile() error codes into POSIX error codes
1291 was out-of-date, and there was no check on the pcre_compile() error code
1292 being within the table. This could lead to an OK return being given in
1293 error.
1294
1295 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1296 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1297 in a Windows environment.
1298
1299 3. The pcregrep --count option prints the count for each file even when it is
1300 zero, as does GNU grep. However, pcregrep was also printing all files when
1301 --files-with-matches was added. Now, when both options are given, it prints
1302 counts only for those files that have at least one match. (GNU grep just
1303 prints the file name in this circumstance, but including the count seems
1304 more useful - otherwise, why use --count?) Also ensured that the
1305 combination -clh just lists non-zero counts, with no names.
1306
1307 4. The long form of the pcregrep -F option was incorrectly implemented as
1308 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1309 but it seems right to fix it, and I didn't think it was worth preserving
1310 the old behaviour.
1311
1312 5. The command line items --regex=pattern and --regexp=pattern were not
1313 recognized by pcregrep, which required --regex pattern or --regexp pattern
1314 (with a space rather than an '='). The man page documented the '=' forms,
1315 which are compatible with GNU grep; these now work.
1316
1317 6. No libpcreposix.pc file was created for pkg-config; there was just
1318 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1319
1320 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1321 when UCP support is not needed, by modifying the Python script that
1322 generates it from Unicode data files. This should not matter if the module
1323 is correctly used as a library, but I received one complaint about 50K of
1324 unwanted data. My guess is that the person linked everything into his
1325 program rather than using a library. Anyway, it does no harm.
1326
1327 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1328 was a minimum greater than 1 for a wide character in a possessive
1329 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1330 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1331 character. Chaos in the form of incorrect output or a compiling loop could
1332 result.
1333
1334 9. The restrictions on what a pattern can contain when partial matching is
1335 requested for pcre_exec() have been removed. All patterns can now be
1336 partially matched by this function. In addition, if there are at least two
1337 slots in the offset vector, the offset of the earliest inspected character
1338 for the match and the offset of the end of the subject are set in them when
1339 PCRE_ERROR_PARTIAL is returned.
1340
1341 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1342 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1343 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1344 and may be more useful for multi-segment matching.
1345
1346 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1347 used to be given if ever the end of the subject was reached; now it is
1348 given only if matching could not proceed because another character was
1349 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1350 string "Z", which now yields "no match" instead of "partial match". In the
1351 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1352 final character ended with (*FAIL).
1353
1354 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1355 if the pattern had a "must contain" character that was already found in the
1356 earlier partial match, unless partial matching was again requested. For
1357 example, with the pattern /dog.(body)?/, the "must contain" character is
1358 "g". If the first part-match was for the string "dog", restarting with
1359 "sbody" failed. This bug has been fixed.
1360
1361 13. The string returned by pcre_dfa_exec() after a partial match has been
1362 changed so that it starts at the first inspected character rather than the
1363 first character of the match. This makes a difference only if the pattern
1364 starts with a lookbehind assertion or \b or \B (\K is not supported by
1365 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1366 matching functions compatible, and I think it's the right thing to do.
1367
1368 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1369 so that the demonstration program is easily available in environments where
1370 PCRE has not been installed from source.
1371
1372 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1373 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1374 library.
1375
1376 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1377 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1378 is not the first non-POSIX option to be added. Clearly some people find
1379 these options useful.
1380
1381 17. If a caller to the POSIX matching function regexec() passes a non-zero
1382 value for nmatch with a NULL value for pmatch, the value of
1383 nmatch is forced to zero.
1384
1385 18. RunGrepTest did not have a test for the availability of the -u option of
1386 the diff command, as RunTest does. It now checks in the same way as
1387 RunTest, and also checks for the -b option.
1388
1389 19. If an odd number of negated classes containing just a single character
1390 interposed, within parentheses, between a forward reference to a named
1391 subpattern and the definition of the subpattern, compilation crashed with
1392 an internal error, complaining that it could not find the referenced
1393 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1394 [The bug was that it was starting one character too far in when skipping
1395 over the character class, thus treating the ] as data rather than
1396 terminating the class. This meant it could skip too much.]
1397
1398 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1399 /g option in pcretest when the pattern contains \K, which makes it possible
1400 to have an empty string match not at the start, even when the pattern is
1401 anchored. Updated pcretest and pcredemo to use this option.
1402
1403 21. If the maximum number of capturing subpatterns in a recursion was greater
1404 than the maximum at the outer level, the higher number was returned, but
1405 with unset values at the outer level. The correct (outer level) value is
1406 now given.
1407
1408 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1409 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1410 make it do so. The string so far is captured, making this feature
1411 compatible with Perl.
1412
1413 23. The tests have been re-organized, adding tests 11 and 12, to make it
1414 possible to check the Perl 5.10 features against Perl 5.10.
1415
1416 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1417 pattern matches a fixed length string. PCRE did not allow this; now it
1418 does. Neither allows recursion.
1419
1420 25. I finally figured out how to implement a request to provide the minimum
1421 length of subject string that was needed in order to match a given pattern.
1422 (It was back references and recursion that I had previously got hung up
1423 on.) This code has now been added to pcre_study(); it finds a lower bound
1424 to the length of subject needed. It is not necessarily the greatest lower
1425 bound, but using it to avoid searching strings that are too short does give
1426 some useful speed-ups. The value is available to calling programs via
1427 pcre_fullinfo().
1428
1429 26. While implementing 25, I discovered to my embarrassment that pcretest had
1430 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1431 study optimizations had never been tested with that matching function.
1432 Oops. What is worse, even when it was passed study data, there was a bug in
1433 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1434 were also very few tests of studied patterns with pcre_dfa_exec().
1435
1436 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1437 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1438 on the other side of the coin, they are no longer allowed to have different
1439 names, because these cannot be distinguished in PCRE, and this has caused
1440 confusion. (This is a difference from Perl.)
1441
1442 28. When duplicate subpattern names are present (necessarily with different
1443 numbers, as required by 27 above), and a test is made by name in a
1444 conditional pattern, either for a subpattern having been matched, or for
1445 recursion in such a pattern, all the associated numbered subpatterns are
1446 tested, and the overall condition is true if the condition is true for any
1447 one of them. This is the way Perl works, and is also more like the way
1448 testing by number works.
1449
1450
1451 Version 7.9 11-Apr-09
1452 ---------------------
1453
1454 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1455 (pcretest), all targets were linked against these libraries. This included
1456 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1457 libraries. This caused unwanted dependencies to be created. This problem
1458 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1459 pcretest is linked with readline.
1460
1461 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1462 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1463 moved outside it again, because FALSE and TRUE are already defined in AIX,
1464 but BOOL is not.
1465
1466 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1467 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1468
1469 4. The pcregrep documentation said spaces were inserted as well as colons (or
1470 hyphens) following file names and line numbers when outputting matching
1471 lines. This is not true; no spaces are inserted. I have also clarified the
1472 wording for the --colour (or --color) option.
1473
1474 5. In pcregrep, when --colour was used with -o, the list of matching strings
1475 was not coloured; this is different to GNU grep, so I have changed it to be
1476 the same.
1477
1478 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1479 each matching line was coloured. Now it goes on to look for further matches
1480 of any of the test patterns, which is the same behaviour as GNU grep.
1481
1482 7. A pattern that could match an empty string could cause pcregrep to loop; it
1483 doesn't make sense to accept an empty string match in pcregrep, so I have
1484 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1485 seems to be how GNU grep behaves.
1486
1487 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1488 start or after a newline", because the conditional assertion was not being
1489 correctly handled. The rule now is that both the assertion and what follows
1490 in the first alternative must satisfy the test.
1491
1492 9. If auto-callout was enabled in a pattern with a conditional group whose
1493 condition was an assertion, PCRE could crash during matching, both with
1494 pcre_exec() and pcre_dfa_exec().
1495
1496 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1497 used for matching.
1498
1499 11. Unicode property support in character classes was not working for
1500 characters (bytes) greater than 127 when not in UTF-8 mode.
1501
1502 12. Added the -M command line option to pcretest.
1503
1504 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1505
1506 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1507
1508 16. Added comments and documentation about mis-use of no_arg in the C++
1509 wrapper.
1510
1511 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1512 from Martin Jerabek that uses macro names for all relevant character and
1513 string constants.
1514
1515 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1516 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1517 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1518 these, but not everybody uses configure.
1519
1520 19. A conditional group that had only one branch was not being correctly
1521 recognized as an item that could match an empty string. This meant that an
1522 enclosing group might also not be so recognized, causing infinite looping
1523 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1524 with the subject "ab", where knowledge that the repeated group can match
1525 nothing is needed in order to break the loop.
1526
1527 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1528 exec(), but without supplying a callout function, matching went wrong.
1529
1530 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1531 leak if the size of the offset vector was greater than 30. When the vector
1532 is smaller, the saved offsets during recursion go onto a local stack
1533 vector, but for larger vectors malloc() is used. It was failing to free
1534 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1535 error, in fact).
1536
1537 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1538 heapframe that is used only when UTF-8 support is enabled. This caused no
1539 problem, but was untidy.
1540
1541 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1542 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1543 included within another project.
1544
1545 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1546 slightly modified by me:
1547
1548 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1549 not building pcregrep.
1550
1551 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1552 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1553
1554 25. Forward references, both numeric and by name, in patterns that made use of
1555 duplicate group numbers, could behave incorrectly or give incorrect errors,
1556 because when scanning forward to find the reference group, PCRE was not
1557 taking into account the duplicate group numbers. A pattern such as
1558 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1559
1560 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1561 the feature of a custom pointer more persuasive (as requested by a user).
1562
1563 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1564 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1565 pre-defined.
1566
1567 28. Added support for (*UTF8) at the start of a pattern.
1568
1569 29. Arrange for flags added by the "release type" setting in CMake to be shown
1570 in the configuration summary.
1571
1572
1573 Version 7.8 05-Sep-08
1574 ---------------------
1575
1576 1. Replaced UCP searching code with optimized version as implemented for Ad
1577 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1578 stage table and inline lookup instead of a function, giving speed ups of 2
1579 to 5 times on some simple patterns that I tested. Permission was given to
1580 distribute the MultiStage2.py script that generates the tables (it's not in
1581 the tarball, but is in the Subversion repository).
1582
1583 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1584 scripts.
1585
1586 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1587 a group with a zero qualifier. The result of the study could be incorrect,
1588 or the function might crash, depending on the pattern.
1589
1590 4. Caseless matching was not working for non-ASCII characters in back
1591 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1592 It now works when Unicode Property Support is available.
1593
1594 5. In pcretest, an escape such as \x{de} in the data was always generating
1595 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1596 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1597 truncation.
1598
1599 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1600
1601 7. Added two (int) casts to pcregrep when printing the difference of two
1602 pointers, in case they are 64-bit values.
1603
1604 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1605 test 2 if it fails.
1606
1607 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1608 and a #define of that name to empty if it is not externally set. This is to
1609 allow users of MSVC to set it if necessary.
1610
1611 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1612 the convenience functions in the pcre_get.c source file.
1613
1614 11. An option change at the start of a pattern that had top-level alternatives
1615 could cause overwriting and/or a crash. This command provoked a crash in
1616 some environments:
1617
1618 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1619
1620 This potential security problem was recorded as CVE-2008-2371.
1621
1622 12. For a pattern where the match had to start at the beginning or immediately
1623 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1624 pcre_dfa_exec() could read past the end of the passed subject if there was
1625 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1626 pcretest so that it places the subject at the end of its malloc-ed buffer.
1627
1628 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1629 exec() might read past the end of the data buffer in UTF-8 mode.
1630
1631 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1632 the data contained the byte 0x85 as part of a UTF-8 character within its
1633 first line. This applied both to normal and DFA matching.
1634
1635 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1636 /^[^d]*?$/8 failed to match "abc".
1637
1638 16. Added a missing copyright notice to pcrecpp_internal.h.
1639
1640 17. Make it more clear in the documentation that values returned from
1641 pcre_exec() in ovector are byte offsets, not character counts.
1642
1643 18. Tidied a few places to stop certain compilers from issuing warnings.
1644
1645 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1646 supplied by Stefan Weber. I made a further small update for 7.8 because
1647 there is a change of source arrangements: the pcre_searchfuncs.c module is
1648 replaced by pcre_ucd.c.
1649
1650
1651 Version 7.7 07-May-08
1652 ---------------------
1653
1654 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1655 a string to a long long, pretend we don't even have a long long." This is
1656 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1657
1658 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1659 pre-7.6 versions, which defined a global no_arg variable instead of putting
1660 it in the RE class. (See also #8 below.)
1661
1662 3. Remove a line of dead code, identified by coverity and reported by Nuno
1663 Lopes.
1664
1665 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1666
1667 (1) The include/exclude patterns were being applied to the whole pathnames
1668 of files, instead of just to the final components.
1669
1670 (2) If there was more than one level of directory, the subdirectories were
1671 skipped unless they satisfied the include/exclude conditions. This is
1672 inconsistent with GNU grep (and could even be seen as contrary to the
1673 pcregrep specification - which I improved to make it absolutely clear).
1674 The action now is always to scan all levels of directory, and just
1675 apply the include/exclude patterns to regular files.
1676
1677 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1678 --exclude_dir in the tests to avoid scanning .svn directories.
1679
1680 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1681 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1682 doesn't support NULs in patterns.
1683
1684 7. Added some missing "const"s to declarations of static tables in
1685 pcre_compile.c and pcre_dfa_exec.c.
1686
1687 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1688 caused by fix #2 above. (Subsequently also a second patch to fix the
1689 first patch. And a third patch - this was a messy problem.)
1690
1691 9. Applied Craig's patch to remove the use of push_back().
1692
1693 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1694 matching function regexec().
1695
1696 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1697 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1698 references. PCRE supports relative numbers with this syntax (I don't think
1699 Oniguruma does).
1700
1701 12. Previously, a group with a zero repeat such as (...){0} was completely
1702 omitted from the compiled regex. However, this means that if the group
1703 was called as a subroutine from elsewhere in the pattern, things went wrong
1704 (an internal error was given). Such groups are now left in the compiled
1705 pattern, with a new opcode that causes them to be skipped at execution
1706 time.
1707
1708 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1709 to the way PCRE behaves:
1710
1711 (a) A lone ] character is dis-allowed (Perl treats it as data).
1712
1713 (b) A back reference to an unmatched subpattern matches an empty string
1714 (Perl fails the current match path).
1715
1716 (c) A data ] in a character class must be notated as \] because if the
1717 first data character in a class is ], it defines an empty class. (In
1718 Perl it is not possible to have an empty class.) The empty class []
1719 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1720 The negative empty class [^] matches any one character, independently
1721 of the DOTALL setting.
1722
1723 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1724 non-existent subpattern following a character class starting with ']' and
1725 containing () gave an internal compiling error instead of "reference to
1726 non-existent subpattern". Fortunately, when the pattern did exist, the
1727 compiled code was correct. (When scanning forwards to check for the
1728 existencd of the subpattern, it was treating the data ']' as terminating
1729 the class, so got the count wrong. When actually compiling, the reference
1730 was subsequently set up correctly.)
1731
1732 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1733 it was being rejected as not supported by pcre_dfa_exec(), even though
1734 other assertions are supported. I have made pcre_dfa_exec() support
1735 (*FAIL).
1736
1737 16. The implementation of 13c above involved the invention of a new opcode,
1738 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1739 cannot be changed at match time, I realized I could make a small
1740 improvement to matching performance by compiling OP_ALLANY instead of
1741 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1742 on the OP_ANY path.
1743
1744 17. Compiling pcretest on Windows with readline support failed without the
1745 following two fixes: (1) Make the unistd.h include conditional on
1746 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1747
1748 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1749 ncurses library to be included for pcretest when ReadLine support is
1750 requested, but also to allow for it to be overridden. This patch came from
1751 Daniel Bergström.
1752
1753 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1754 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1755 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1756 spotting this.
1757
1758
1759 Version 7.6 28-Jan-08
1760 ---------------------
1761
1762 1. A character class containing a very large number of characters with
1763 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1764 overflow.
1765
1766 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1767 HAVE_LONG_LONG is not defined.
1768
1769 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1770 bring them up to date and include new features. This patch includes:
1771
1772 - Fixed PH's badly added libz and libbz2 support.
1773 - Fixed a problem with static linking.
1774 - Added pcredemo. [But later removed - see 7 below.]
1775 - Fixed dftables problem and added an option.
1776 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1777 HAVE_LONG_LONG.
1778 - Added readline support for pcretest.
1779 - Added an listing of the option settings after cmake has run.
1780
1781 4. A user submitted a patch to Makefile that makes it easy to create
1782 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1783 Makefile.am that cause it to include this special target, without
1784 affecting anything else. Note that the same mingw target plus all
1785 the other distribution libraries and programs are now supported
1786 when configuring with CMake (see 6 below) instead of with
1787 Configure/Make.
1788
1789 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1790 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1791 exported in the Windows port". It has not yet been confirmed that the patch
1792 solves the problem, but it does no harm.
1793
1794 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1795 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1796 with CMake, and also correct the comment about stack recursion.
1797
1798 7. Remove the automatic building of pcredemo from the ./configure system and
1799 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1800 of a program that users should build themselves after PCRE is installed, so
1801 building it automatically is not really right. What is more, it gave
1802 trouble in some build environments.
1803
1804 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1805
1806
1807 Version 7.5 10-Jan-08
1808 ---------------------
1809
1810 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1811 values in parens when parsing an RE using the C++ wrapper."
1812
1813 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1814 Characters greater than 255 were excluded from the class instead of being
1815 included.
1816
1817 3. The same bug as (2) above applied to negated POSIX classes such as
1818 [:^space:].
1819
1820 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1821 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1822 I have changed it.
1823
1824 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1825 first named subpattern) and a construct such as (?&a) would reference the
1826 first named subpattern whose name started with "a" (in other words, the
1827 length check was missing). Both these problems are fixed. "Subpattern name
1828 expected" is now given for (?&) (a zero-length name), and this patch also
1829 makes it give the same error for \k'' (previously it complained that that
1830 was a reference to a non-existent subpattern).
1831
1832 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1833 this is right because (?- can be followed by option settings as well as by
1834 digits. I have, however, made the messages clearer.
1835
1836 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1837 than the number used in the conditional) now cause a compile-time error.
1838 This is actually not compatible with Perl, which accepts such patterns, but
1839 treats the conditional as always being FALSE (as PCRE used to), but it
1840 seems to me that giving a diagnostic is better.
1841
1842 8. Change "alphameric" to the more common word "alphanumeric" in comments
1843 and messages.
1844
1845 9. Fix two occurrences of "backslash" in comments that should have been
1846 "backspace".
1847
1848 10. Remove two redundant lines of code that can never be obeyed (their function
1849 was moved elsewhere).
1850
1851 11. The program that makes PCRE's Unicode character property table had a bug
1852 which caused it to generate incorrect table entries for sequences of
1853 characters that have the same character type, but are in different scripts.
1854 It amalgamated them into a single range, with the script of the first of
1855 them. In other words, some characters were in the wrong script. There were
1856 thirteen such cases, affecting characters in the following ranges:
1857
1858 U+002b0 - U+002c1
1859 U+0060c - U+0060d
1860 U+0061e - U+00612
1861 U+0064b - U+0065e
1862 U+0074d - U+0076d
1863 U+01800 - U+01805
1864 U+01d00 - U+01d77
1865 U+01d9b - U+01dbf
1866 U+0200b - U+0200f
1867 U+030fc - U+030fe
1868 U+03260 - U+0327f
1869 U+0fb46 - U+0fbb1
1870 U+10450 - U+1049d
1871
1872 12. The -o option (show only the matching part of a line) for pcregrep was not
1873 compatible with GNU grep in that, if there was more than one match in a
1874 line, it showed only the first of them. It now behaves in the same way as
1875 GNU grep.
1876
1877 13. If the -o and -v options were combined for pcregrep, it printed a blank
1878 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1879 does the same. The return code can be used to tell if there were any
1880 non-matching lines.
1881
1882 14. Added --file-offsets and --line-offsets to pcregrep.
1883
1884 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1885 infinitely looping recursion. The bug was that positive lookaheads were not
1886 being skipped when checking for a possible empty match (negative lookaheads
1887 and both kinds of lookbehind were skipped).
1888
1889 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1890 inclusion of <windows.h> to before rather than after the definition of
1891 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1892
1893 17. Specifying a possessive quantifier with a specific limit for a Unicode
1894 character property caused pcre_compile() to compile bad code, which led at
1895 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1896 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1897 caused the error; without that there was no problem.
1898
1899 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1900
1901 19. Added --enable-pcretest-libreadline.
1902
1903 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1904 RE::GlobalReplace(). As a result, the number of replacements returned was
1905 double what it should be. I removed one of the increments, but Craig sent a
1906 later patch that removed the other one (the right fix) and added unit tests
1907 that check the return values (which was not done before).
1908
1909 21. Several CMake things:
1910
1911 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1912 the names libpcre and libpcreposix, not just pcre and pcreposix.
1913
1914 (2) The above change means that pcretest and pcregrep are now correctly
1915 linked with the newly-built libraries, not previously installed ones.
1916
1917 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1918
1919 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1920 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1921 UTF-8 newline character). The key issue is that the pattern starts .*;
1922 this means that the match must be either at the beginning, or after a
1923 newline. The bug was in the code for advancing after a failed match and
1924 checking that the new position followed a newline. It was not taking
1925 account of UTF-8 characters correctly.
1926
1927 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1928 character classes. PCRE was not treating the sequence [:...:] as a
1929 character class unless the ... were all letters. Perl, however, seems to
1930 allow any characters between [: and :], though of course it rejects as
1931 unknown any "names" that contain non-letters, because all the known class
1932 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1933 for example, whereas PCRE did not - it did not recognize a POSIX character
1934 class. This seemed a bit dangerous, so the code has been changed to be
1935 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1936 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1937 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1938 Perl does, and where it didn't before.
1939
1940 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1941 Windows environments %n is disabled by default.
1942
1943
1944 Version 7.4 21-Sep-07
1945 ---------------------
1946
1947 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1948 means that a class such as [\s] counted as "explicit reference to CR or
1949 LF". That isn't really right - the whole point of the change was to try to
1950 help when there was an actual mention of one of the two characters. So now
1951 the change happens only if \r or \n (or a literal CR or LF) character is
1952 encountered.
1953
1954 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1955 of both had grown to the point where there were only 3 bits left.
1956 Fortunately, there was spare space in the data structure, and so I have
1957 moved the internal flags into a new 16-bit field to free up more option
1958 bits.
1959
1960 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1961 but did not set the internal JCHANGED flag - either of these is enough to
1962 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1963 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1964 start sets both bits.
1965
1966 4. Added options (at build time, compile time, exec time) to change \R from
1967 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1968
1969 5. doc/pcresyntax.html was missing from the distribution.
1970
1971 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1972 compatibility, even though it is no longer used.
1973
1974 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1975 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1976 windows.h file is present (where different names are used). [This was
1977 reversed later after testing - see 16 below.]
1978
1979 8. Changed all #include <config.h> to #include "config.h". There were also
1980 some further <pcre.h> cases that I changed to "pcre.h".
1981
1982 9. When pcregrep was used with the --colour option, it missed the line ending
1983 sequence off the lines that it output.
1984
1985 10. It was pointed out to me that arrays of string pointers cause lots of
1986 relocations when a shared library is dynamically loaded. A technique of
1987 using a single long string with a table of offsets can drastically reduce
1988 these. I have refactored PCRE in four places to do this. The result is
1989 dramatic:
1990
1991 Originally: 290
1992 After changing UCP table: 187
1993 After changing error message table: 43
1994 After changing table of "verbs" 36
1995 After changing table of Posix names 22
1996
1997 Thanks to the folks working on Gregex for glib for this insight.
1998
1999 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2000 unicode-properties was also set.
2001
2002 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2003
2004 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2005 checked only for CRLF.
2006
2007 14. Added casts to pcretest.c to avoid compiler warnings.
2008
2009 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2010
2011 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2012 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2013 entirely. This removes changes made in 7 above.
2014
2015 17. The CMake files have been updated, and there is now more information about
2016 building with CMake in the NON-UNIX-USE document.
2017
2018
2019 Version 7.3 28-Aug-07
2020 ---------------------
2021
2022 1. In the rejigging of the build system that eventually resulted in 7.1, the
2023 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2024 brackets there is not right, since it causes compilers to look for an
2025 installed pcre.h, not the version that is in the source that is being
2026 compiled (which of course may be different). I have changed it back to:
2027
2028 #include "pcre.h"
2029
2030 I have a vague recollection that the change was concerned with compiling in
2031 different directories, but in the new build system, that is taken care of
2032 by the VPATH setting the Makefile.
2033
2034 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2035 when the subject happened to end in the byte 0x85 (e.g. if the last
2036 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2037 characters but of course it shouldn't be taken as a newline when it is part
2038 of another character. The bug was that, for an unlimited repeat of . in
2039 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2040 characters when looking for a newline.
2041
2042 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2043
2044 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2045 in debug output.
2046
2047 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2048 long printing in the pcrecpp unittest when running under MinGW.
2049
2050 6. ESC_K was left out of the EBCDIC table.
2051
2052 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2053 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2054 limit also applies to "virtual nesting" when a pattern is recursive, and in
2055 this case 1000 isn't so big. I have been able to remove this limit at the
2056 expense of backing off one optimization in certain circumstances. Normally,
2057 when pcre_exec() would call its internal match() function recursively and
2058 immediately return the result unconditionally, it uses a "tail recursion"
2059 feature to save stack. However, when a subpattern that can match an empty
2060 string has an unlimited repetition quantifier, it no longer makes this
2061 optimization. That gives it a stack frame in which to save the data for
2062 checking that an empty string has been matched. Previously this was taken
2063 from the 1000-entry workspace that had been reserved. So now there is no
2064 explicit limit, but more stack is used.
2065
2066 8. Applied Daniel's patches to solve problems with the import/export magic
2067 syntax that is required for Windows, and which was going wrong for the
2068 pcreposix and pcrecpp parts of the library. These were overlooked when this
2069 problem was solved for the main library.
2070
2071 9. There were some crude static tests to avoid integer overflow when computing
2072 the size of patterns that contain repeated groups with explicit upper
2073 limits. As the maximum quantifier is 65535, the maximum group length was
2074 set at 30,000 so that the product of these two numbers did not overflow a
2075 32-bit integer. However, it turns out that people want to use groups that
2076 are longer than 30,000 bytes (though not repeat them that many times).
2077 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2078 made it possible to implement the integer overflow checks in a much more
2079 dynamic way, which I have now done. The artificial limitation on group
2080 length has been removed - we now have only the limit on the total length of
2081 the compiled pattern, which depends on the LINK_SIZE setting.
2082
2083 10. Fixed a bug in the documentation for get/copy named substring when
2084 duplicate names are permitted. If none of the named substrings are set, the
2085 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2086 empty string.
2087
2088 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2089 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2090 because the ] is interpreted as the first data character and the
2091 terminating ] is not found. PCRE has been made compatible with Perl in this
2092 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2093 cause memory overwriting.
2094
2095 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2096 string has been matched (to stop an infinite loop). It was not recognizing
2097 a conditional subpattern that could match an empty string if that
2098 subpattern was within another subpattern. For example, it looped when
2099 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2100 condition was not nested. This bug has been fixed.
2101
2102 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2103 past the start of the subject in the presence of bytes with the top bit
2104 set, for example "\x8aBCD".
2105
2106 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2107 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2108
2109 14. Optimized (?!) to (*FAIL).
2110
2111 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2112 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2113 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2114 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2115 does: it's just the validity check that is more restrictive.
2116
2117 16. Inserted checks for integer overflows during escape sequence (backslash)
2118 processing, and also fixed erroneous offset values for syntax errors during
2119 backslash processing.
2120
2121 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2122 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2123
2124 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2125 caused an overrun.
2126
2127 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2128 something other than just ASCII characters) inside a group that had an
2129 unlimited repeat caused a loop at compile time (while checking to see
2130 whether the group could match an empty string).
2131
2132 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2133 [\P{Any}] did so. (Error in the code for printing property names.)
2134
2135 21. An orphan \E inside a character class could cause a crash.
2136
2137 22. A repeated capturing bracket such as (A)? could cause a wild memory
2138 reference during compilation.
2139
2140 23. There are several functions in pcre_compile() that scan along a compiled
2141 expression for various reasons (e.g. to see if it's fixed length for look
2142 behind). There were bugs in these functions when a repeated \p or \P was
2143 present in the pattern. These operators have additional parameters compared
2144 with \d, etc, and these were not being taken into account when moving along
2145 the compiled data. Specifically:
2146
2147 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2148 length.
2149
2150 (b) An item such as \pL+ within a repeated group could cause crashes or
2151 loops.
2152
2153 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2154 "reference to non-existent subpattern" error.
2155
2156 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2157
2158 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2159 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2160
2161 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2162
2163 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2164 character were causing crashes (broken optimization).
2165
2166 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2167 \p or \P) caused a compile-time loop.
2168
2169 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2170 break. For example, the unstudied pattern [\r\n]A does not match the string
2171 "\r\nA" because change 7.0/46 below moves the current point on by two
2172 characters after failing to match at the start. However, the pattern \nA
2173 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2174 the same is true. There doesn't seem any very clean way out of this, but
2175 what I have chosen to do makes the common cases work: PCRE now takes note
2176 of whether there can be an explicit match for \r or \n anywhere in the
2177 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2178 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2179 pattern has explicit CR or LF references.
2180
2181 29. Added (*CR) etc for changing newline setting at start of pattern.
2182
2183
2184 Version 7.2 19-Jun-07
2185 ---------------------
2186
2187 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2188 which is apparently normally available under Windows.
2189
2190 2. Re-jig the pcregrep tests with different newline settings in an attempt
2191 to make them independent of the local environment's newline setting.
2192
2193 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2194
2195 4. Some of the "internals" tests were previously cut out when the link size
2196 was not 2, because the output contained actual offsets. The recent new
2197 "Z" feature of pcretest means that these can be cut out, making the tests
2198 usable with all link sizes.
2199
2200 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2201 stack recursion. This gives a massive performance boost under BSD, but just
2202 a small improvement under Linux. However, it saves one field in the frame
2203 in all cases.
2204
2205 6. Added more features from the forthcoming Perl 5.10:
2206
2207 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2208 recursion call. It refers to the nth most recently opened parentheses.
2209
2210 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2211 to be opened parentheses.
2212
2213 (c) Conditions that refer to capturing parentheses can be specified
2214 relatively, for example, (?(-2)... or (?(+3)...
2215
2216 (d) \K resets the start of the current match so that everything before
2217 is not part of it.
2218
2219 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2220
2221 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2222 reference syntax.
2223
2224 (g) (?| introduces a group in which the numbering of parentheses in each
2225 alternative starts with the same number.
2226
2227 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2228
2229 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2230 PCRE_INFO_JCHANGED.
2231
2232 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2233 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2234 for detecting groups that can match an empty string.
2235
2236 9. A pattern with a very large number of alternatives (more than several
2237 hundred) was running out of internal workspace during the pre-compile
2238 phase, where pcre_compile() figures out how much memory will be needed. A
2239 bit of new cunning has reduced the workspace needed for groups with
2240 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2241 workspace instead of running out of the 4096 that are available.
2242
2243 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2244
2245 11. Applied patch from Google to remove an optimization that didn't quite work.
2246 The report of the bug said:
2247
2248 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2249 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2250 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2251
2252 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2253 it matched the wrong number of bytes.
2254
2255
2256 Version 7.1 24-Apr-07
2257 ---------------------
2258
2259 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2260 that is more "standard", making use of automake and other Autotools. There
2261 is some re-arrangement of the files and adjustment of comments consequent
2262 on this.
2263
2264 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2265 for recursive directory scanning broke on some systems because the files
2266 are not scanned in any specific order and on different systems the order
2267 was different. A call to "sort" has been inserted into RunGrepTest for the
2268 approprate test as a short-term fix. In the longer term there may be an
2269 alternative.
2270
2271 3. I had an email from Eric Raymond about problems translating some of PCRE's
2272 man pages to HTML (despite the fact that I distribute HTML pages, some
2273 people do their own conversions for various reasons). The problems
2274 concerned the use of low-level troff macros .br and .in. I have therefore
2275 removed all such uses from the man pages (some were redundant, some could
2276 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2277 HTML has been updated to handle .nf/.fi and to complain if it encounters
2278 .br or .in.
2279
2280 4. Updated comments in configure.ac that get placed in config.h.in and also
2281 arranged for config.h to be included in the distribution, with the name
2282 config.h.generic, for the benefit of those who have to compile without
2283 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2284
2285 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2286 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2287 makevp.bat for the current PCRE, using the additional files
2288 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2289
2290 6. A Windows user reported a minor discrepancy with test 2, which turned out
2291 to be caused by a trailing space on an input line that had got lost in his
2292 copy. The trailing space was an accident, so I've just removed it.
2293
2294 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2295 that is needed.
2296
2297 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2298 as "const" (a) because they are and (b) because it helps the PHP
2299 maintainers who have recently made a script to detect big data structures
2300 in the php code that should be moved to the .rodata section. I remembered
2301 to update Builducptable as well, so it won't revert if ucptable.h is ever
2302 re-created.
2303
2304 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2305 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2306 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2307 support is not required. This saves 1.5-2K of code, which is important in
2308 some applications.
2309
2310 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2311 so as not to refer to the tables, even though these functions will never be
2312 called when UTF-8 support is disabled. Otherwise there are problems with a
2313 shared library.
2314
2315 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2316
2317 (a) It was defining its arguments as char * instead of void *.
2318
2319 (b) It was assuming that all moves were upwards in memory; this was true
2320 a long time ago when I wrote it, but is no longer the case.
2321
2322 The emulated memove() is provided for those environments that have neither
2323 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2324 is clearly not the case, as these two bugs were recently reported.
2325
2326 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2327 and Detrail to create the HTML documentation, the .txt form of the man
2328 pages, and it removes trailing spaces from listed files. It also creates
2329 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2330 case, it wraps all the #defines with #ifndefs. This script should be run
2331 before "make dist".
2332
2333 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2334 with Unicode property support.
2335
2336 (a) For a maximizing quantifier, if the two different cases of the
2337 character were of different lengths in their UTF-8 codings (there are
2338 some cases like this - I found 11), and the matching function had to
2339 back up over a mixture of the two cases, it incorrectly assumed they
2340 were both the same length.
2341
2342 (b) When PCRE was configured to use the heap rather than the stack for
2343 recursion during matching, it was not correctly preserving the data for
2344 the other case of a UTF-8 character when checking ahead for a match
2345 while processing a minimizing repeat. If the check also involved
2346 matching a wide character, but failed, corruption could cause an
2347 erroneous result when trying to check for a repeat of the original
2348 character.
2349
2350 13. Some tidying changes to the testing mechanism:
2351
2352 (a) The RunTest script now detects the internal link size and whether there
2353 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2354 values substituted by "configure". (The RunGrepTest script already did
2355 this for UTF-8.) The configure.ac script no longer substitutes the
2356 relevant variables.
2357
2358 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2359 with length and offset values. This means that the output is different
2360 for different internal link sizes. Test 2 is skipped for link sizes
2361 other than 2 because of this, bypassing the problem. Unfortunately,
2362 there was also a test in test 3 (the locale tests) that used /B and
2363 failed for link sizes other than 2. Rather than cut the whole test out,
2364 I have added a new /Z option to pcretest that replaces the length and
2365 offset values with spaces. This is now used to make test 3 independent
2366 of link size. (Test 2 will be tidied up later.)
2367
2368 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2369 segmentation fault instead of returning the appropriate error message.
2370
2371 15. In multiline mode when the newline sequence was set to "any", the pattern
2372 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2373 This doesn't seem right; it now treats the CRLF combination as the line
2374 ending, and so does not match in that case. It's only a pattern such as ^$
2375 that would hit this one: something like ^ABC$ would have failed after \r
2376 and then tried again after \r\n.
2377
2378 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2379 in an attempt to make files that differ only in their line terminators
2380 compare equal. This works on Linux.
2381
2382 17. Under certain error circumstances pcregrep might try to free random memory
2383 as it exited. This is now fixed, thanks to valgrind.
2384
2385 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2386 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2387 was because its rules for how to advance for /g after matching an empty
2388 string at the end of a line did not allow for this case. They now check for
2389 it specially.
2390
2391 20. pcretest is supposed to handle patterns and data of any length, by
2392 extending its buffers when necessary. It was getting this wrong when the
2393 buffer for a data line had to be extended.
2394
2395 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2396 CRLF as a newline sequence.
2397
2398 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2399 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2400 I have nevertheless tidied it up.
2401
2402 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2403
2404 24. Added a man page for pcre-config.
2405
2406
2407 Version 7.0 19-Dec-06
2408 ---------------------
2409
2410 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2411 moving to gcc 4.1.1.
2412
2413 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2414 sys/time.h, which is documented as needed for this function. It doesn't
2415 seem to matter on Linux, but it showed up on some releases of OS X.
2416
2417 3. It seems that there are systems where bytes whose values are greater than
2418 127 match isprint() in the "C" locale. The "C" locale should be the
2419 default when a C program starts up. In most systems, only ASCII printing
2420 characters match isprint(). This difference caused the output from pcretest
2421 to vary, making some of the tests fail. I have changed pcretest so that:
2422
2423 (a) When it is outputting text in the compiled version of a pattern, bytes
2424 other than 32-126 are always shown as hex escapes.
2425
2426 (b) When it is outputting text that is a matched part of a subject string,
2427 it does the same, unless a different locale has been set for the match
2428 (using the /L modifier). In this case, it uses isprint() to decide.
2429
2430 4. Fixed a major bug that caused incorrect computation of the amount of memory
2431 required for a compiled pattern when options that changed within the
2432 pattern affected the logic of the preliminary scan that determines the
2433 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2434 that the computed length was too small. The symptoms of this bug were
2435 either the PCRE error "internal error: code overflow" from pcre_compile(),
2436 or a glibc crash with a message such as "pcretest: free(): invalid next
2437 size (fast)". Examples of patterns that provoked this bug (shown in
2438 pcretest format) are:
2439
2440 /(?-x: )/x
2441 /(?x)(?-x: \s*#\s*)/
2442 /((?i)[\x{c0}])/8
2443 /(?i:[\x{c0}])/8
2444
2445 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2446 is now done differently.
2447
2448 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2449 wrapper classes; (b) implement a new function in the C++ scanner that is
2450 more efficient than the old way of doing things because it avoids levels of
2451 recursion in the regex matching; (c) add a paragraph to the documentation
2452 for the FullMatch() function.
2453
2454 6. The escape sequence \n was being treated as whatever was defined as
2455 "newline". Not only was this contrary to the documentation, which states
2456 that \n is character 10 (hex 0A), but it also went horribly wrong when
2457 "newline" was defined as CRLF. This has been fixed.
2458
2459 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2460 was being set to -1 for the "end of line" case (supposedly a value that no
2461 character can have). Though this value is never used (the check for end of
2462 line is "zero bytes in current character"), it caused compiler complaints.
2463 I've changed it to 0xffffffff.
2464
2465 8. In pcre_version.c, the version string was being built by a sequence of
2466 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2467 string (as it is for production releases) called a macro with an empty
2468 argument. The C standard says the result of this is undefined. The gcc
2469 compiler treats it as an empty string (which was what was wanted) but it is
2470 reported that Visual C gives an error. The source has been hacked around to
2471 avoid this problem.
2472
2473 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2474 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2475 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2476 of them did).
2477
2478 10. Originally, pcretest opened its input and output without "b"; then I was
2479 told that "b" was needed in some environments, so it was added for release
2480 5.0 to both the input and output. (It makes no difference on Unix-like
2481 systems.) Later I was told that it is wrong for the input on Windows. I've
2482 now abstracted the modes into two macros, to make it easier to fiddle with
2483 them, and removed "b" from the input mode under Windows.
2484
2485 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2486
2487 12. Added -help and --help to pcretest as an official way of being reminded
2488 of the options.
2489
2490 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2491 and pcrecpp.cc because they annoy compilers at high warning levels.
2492
2493 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2494
2495 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2496 scripts are not C programs :-) and which was not noticed because it works
2497 on Linux.
2498
2499 16. pcretest is supposed to handle any length of pattern and data line (as one
2500 line or as a continued sequence of lines) by extending its input buffer if
2501 necessary. This feature was broken for very long pattern lines, leading to
2502 a string of junk being passed to pcre_compile() if the pattern was longer
2503 than about 50K.
2504
2505 17. I have done a major re-factoring of the way pcre_compile() computes the
2506 amount of memory needed for a compiled pattern. Previously, there was code
2507 that made a preliminary scan of the pattern in order to do this. That was
2508 OK when PCRE was new, but as the facilities have expanded, it has become
2509 harder and harder to keep it in step with the real compile phase, and there
2510 have been a number of bugs (see for example, 4 above). I have now found a
2511 cunning way of running the real compile function in a "fake" mode that
2512 enables it to compute how much memory it would need, while actually only
2513 ever using a few hundred bytes of working memory and without too many
2514 tests of the mode. This should make future maintenance and development
2515 easier. A side effect of this work is that the limit of 200 on the nesting
2516 depth of parentheses has been removed (though this was never a serious
2517 limitation, I suspect). However, there is a downside: pcre_compile() now
2518 runs more slowly than before (30% or more, depending on the pattern). I
2519 hope this isn't a big issue. There is no effect on runtime performance.
2520
2521 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2522 newline (only possible for the last line of a file) and it was a
2523 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2524
2525 19. Added additional timing features to pcretest. (1) The -tm option now times
2526 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2527 separate command line item, by a number that specifies the number of
2528 repeats to use when timing. The default is 50000; this gives better
2529 precision, but takes uncomfortably long for very large patterns.
2530
2531 20. Extended pcre_study() to be more clever in cases where a branch of a
2532 subpattern has no definite first character. For example, (a*|b*)[cd] would
2533 previously give no result from pcre_study(). Now it recognizes that the
2534 first character must be a, b, c, or d.
2535
2536 21. There was an incorrect error "recursive call could loop indefinitely" if
2537 a subpattern (or the entire pattern) that was being tested for matching an
2538 empty string contained only one non-empty item after a nested subpattern.
2539 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2540 incorrectly, because the \d was being skipped in the check.
2541
2542 22. The pcretest program now has a new pattern option /B and a command line
2543 option -b, which is equivalent to adding /B to every pattern. This causes
2544 it to show the compiled bytecode, without the additional information that
2545 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2546 is the same as /B/I).
2547
2548 23. A new optimization is now able automatically to treat some sequences such
2549 as a*b as a*+b. More specifically, if something simple (such as a character
2550 or a simple class like \d) has an unlimited quantifier, and is followed by
2551 something that cannot possibly match the quantified thing, the quantifier
2552 is automatically "possessified".
2553
2554 24. A recursive reference to a subpattern whose number was greater than 39
2555 went wrong under certain circumstances in UTF-8 mode. This bug could also
2556 have affected the operation of pcre_study().
2557
2558 25. Realized that a little bit of performance could be had by replacing
2559 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2560
2561 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2562
2563 27. Possessive quantifiers such as a++ were previously implemented by turning
2564 them into atomic groups such as ($>a+). Now they have their own opcodes,
2565 which improves performance. This includes the automatically created ones
2566 from 23 above.
2567
2568 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2569 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2570 the first matched character to be a colon. This applied both to named and
2571 numbered groups.
2572
2573 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2574
2575 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2576 building PCRE on Windows easier, so I have included it in the distribution.
2577
2578 31. There is now a check in pcretest against a ridiculously large number being
2579 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2580 loop, the loop is abandoned.
2581
2582 32. Forward references to subpatterns in conditions such as (?(2)...) where
2583 subpattern 2 is defined later cause pcre_compile() to search forwards in
2584 the pattern for the relevant set of parentheses. This search went wrong
2585 when there were unescaped parentheses in a character class, parentheses
2586 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2587
2588 33. "Subroutine" calls and backreferences were previously restricted to
2589 referencing subpatterns earlier in the regex. This restriction has now
2590 been removed.
2591
2592 34. Added a number of extra features that are going to be in Perl 5.10. On the
2593 whole, these are just syntactic alternatives for features that PCRE had
2594 previously implemented using the Python syntax or my own invention. The
2595 other formats are all retained for compatibility.
2596
2597 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2598 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2599 also .NET compatible.
2600
2601 (b) A recursion or subroutine call to a named group can now be defined as
2602 (?&name) as well as (?P>name).
2603
2604 (c) A backreference to a named group can now be defined as \k<name> or
2605 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2606 5.10, are also .NET compatible.
2607
2608 (d) A conditional reference to a named group can now use the syntax
2609 (?(<name>) or (?('name') as well as (?(name).
2610
2611 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2612 groups (named and numbered) that are never evaluated inline, but can be
2613 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2614 is always false. There may be only one alternative in such a group.
2615
2616 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2617 as the simple (?(R). The condition is true only if the most recent
2618 recursion is that of the given number or name. It does not search out
2619 through the entire recursion stack.
2620
2621 (g) The escape \gN or \g{N} has been added, where N is a positive or
2622 negative number, specifying an absolute or relative reference.
2623
2624 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2625 some "unreachable code" warnings.
2626
2627 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2628 things, this adds five new scripts.
2629
2630 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2631 There were also incompatibilities regarding the handling of \Q..\E inside
2632 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2633 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2634
2635 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2636 matches an empty string, and forcibly breaks the loop. There were bugs in
2637 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2638 against aaaa the result was just "a" rather than "aaaa", for example. Two
2639 separate and independent bugs (that affected different cases) have been
2640 fixed.
2641
2642 39. Refactored the code to abolish the use of different opcodes for small
2643 capturing bracket numbers. This is a tidy that I avoided doing when I
2644 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2645 The new approach is not only tidier, it makes it possible to reduce the
2646 memory needed to fix the previous bug (38).
2647
2648 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2649 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2650 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2651 mode.
2652
2653 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2654 report.
2655
2656 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2657 copy construction and assignment in the C++ wrapper.
2658
2659 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2660 couple of bugs that could have given wrong results in the "--newline=crlf"
2661 case.
2662
2663 44. Added a number of casts and did some reorganization of signed/unsigned int
2664 variables following suggestions from Dair Grant. Also renamed the variable
2665 "this" as "item" because it is a C++ keyword.
2666
2667 45. Arranged for dftables to add
2668
2669 #include "pcre_internal.h"
2670
2671 to pcre_chartables.c because without it, gcc 4.x may remove the array
2672 definition from the final binary if PCRE is built into a static library and
2673 dead code stripping is activated.
2674
2675 46. For an unanchored pattern, if a match attempt fails at the start of a
2676 newline sequence, and the newline setting is CRLF or ANY, and the next two
2677 characters are CRLF, advance by two characters instead of one.
2678
2679
2680 Version 6.7 04-Jul-06
2681 ---------------------
2682
2683 1. In order to handle tests when input lines are enormously long, pcretest has
2684 been re-factored so that it automatically extends its buffers when
2685 necessary. The code is crude, but this _is_ just a test program. The
2686 default size has been increased from 32K to 50K.
2687
2688 2. The code in pcre_study() was using the value of the re argument before
2689 testing it for NULL. (Of course, in any sensible call of the function, it
2690 won't be NULL.)
2691
2692 3. The memmove() emulation function in pcre_internal.h, which is used on
2693 systems that lack both memmove() and bcopy() - that is, hardly ever -
2694 was missing a "static" storage class specifier.
2695
2696 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2697 containing an extended class (one that cannot be represented by a bitmap
2698 because it contains high-valued characters or Unicode property items, e.g.
2699 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2700 pattern, but PCRE should not loop if you do not (it no longer does).
2701 [Detail: two cases were found: (a) a repeated subpattern containing an
2702 extended class; (b) a recursive reference to a subpattern that followed a
2703 previous extended class. It wasn't skipping over the extended class
2704 correctly when UTF-8 mode was not set.]
2705
2706 5. A negated single-character class was not being recognized as fixed-length
2707 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2708 compile error "lookbehind assertion is not fixed length".
2709
2710 6. The RunPerlTest auxiliary script was showing an unexpected difference
2711 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2712 write a Perl script that can interpret lines of an input file either as
2713 byte characters or as UTF-8, which is what "perltest" was being required to
2714 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2715 can't do is switch easily at run time between having the "use utf8;" pragma
2716 or not. In the end, I fudged it by using the RunPerlTest script to insert
2717 "use utf8;" explicitly for the UTF-8 tests.
2718
2719 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2720 the end of the subject string, contrary to the documentation and to what
2721 Perl does. This was true of both matching functions. Now it matches only at
2722 the start of the subject and immediately after *internal* newlines.
2723
2724 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2725 a pointer to an int instead of a pointer to an unsigned long int. This
2726 caused problems on 64-bit systems.
2727
2728 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2729 instance of the 'standard' template library not being so standard".
2730
2731 10. There was no check on the number of named subpatterns nor the maximum
2732 length of a subpattern name. The product of these values is used to compute
2733 the size of the memory block for a compiled pattern. By supplying a very
2734 long subpattern name and a large number of named subpatterns, the size
2735 computation could be caused to overflow. This is now prevented by limiting
2736 the length of names to 32 characters, and the number of named subpatterns
2737 to 10,000.
2738
2739 11. Subpatterns that are repeated with specific counts have to be replicated in
2740 the compiled pattern. The size of memory for this was computed from the
2741 length of the subpattern and the repeat count. The latter is limited to
2742 65535, but there was no limit on the former, meaning that integer overflow
2743 could in principle occur. The compiled length of a repeated subpattern is
2744 now limited to 30,000 bytes in order to prevent this.
2745
2746 12. Added the optional facility to have named substrings with the same name.
2747
2748 13. Added the ability to use a named substring as a condition, using the
2749 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2750 are numbers (not recommended). Forward references are permitted.
2751
2752 14. Added forward references in named backreferences (if you see what I mean).
2753
2754 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2755 pattern could run off the end of the subject. For example, the pattern
2756 "(?s)(.{1,5})"8 did this with the subject "ab".
2757
2758 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2759 PCRE_CASELESS was set when matching characters that were quantified with ?
2760 or *.
2761
2762 17. A character class other than a single negated character that had a minimum
2763 but no maximum quantifier - for example [ab]{6,} - was not handled
2764 correctly by pce_dfa_exec(). It would match only one character.
2765
2766 18. A valid (though odd) pattern that looked like a POSIX character
2767 class but used an invalid character after [ (for example [[,abc,]]) caused
2768 pcre_compile() to give the error "Failed: internal error: code overflow" or
2769 in some cases to crash with a glibc free() error. This could even happen if
2770 the pattern terminated after [[ but there just happened to be a sequence of
2771 letters, a binary zero, and a closing ] in the memory that followed.
2772
2773 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2774 over the years. Originally (before any Unicode support), just the bottom 8
2775 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2776 output from "man perlunicode" includes this:
2777
2778 The regular expression compiler produces polymorphic opcodes. That
2779 is, the pattern adapts to the data and automatically switches to
2780 the Unicode character scheme when presented with Unicode data--or
2781 instead uses a traditional byte scheme when presented with byte
2782 data.
2783
2784 Sadly, a wide octal escape does not cause a switch, and in a string with
2785 no other multibyte characters, these octal escapes are treated as before.
2786 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2787 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2788 Unicode string.
2789
2790 I have not perpetrated such confusion in PCRE. Up till now, it took just
2791 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2792 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2793 translate to the appropriate multibyte character.
2794
2795 29. Applied some refactoring to reduce the number of warnings from Microsoft
2796 and Borland compilers. This has included removing the fudge introduced
2797 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2798 a warning about an unused variable.
2799
2800 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2801 characters since release 4.0, because Perl (from release 5.004) does not.
2802 [Or at least, is documented not to: some releases seem to be in conflict
2803 with the documentation.] However, when a pattern was studied with
2804 pcre_study() and all its branches started with \s, PCRE still included VT
2805 as a possible starting character. Of course, this did no harm; it just
2806 caused an unnecessary match attempt.
2807
2808 22. Removed a now-redundant internal flag bit that recorded the fact that case
2809 dependency changed within the pattern. This was once needed for "required
2810 byte" processing, but is no longer used. This recovers a now-scarce options
2811 bit. Also moved the least significant internal flag bit to the most-
2812 significant bit of the word, which was not previously used (hangover from
2813 the days when it was an int rather than a uint) to free up another bit for
2814 the future.
2815
2816 23. Added support for CRLF line endings as well as CR and LF. As well as the
2817 default being selectable at build time, it can now be changed at runtime
2818 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2819 specify that it is scanning data with non-default line endings.
2820
2821 24. Changed the definition of CXXLINK to make it agree with the definition of
2822 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2823
2824 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2825 recursions. This makes a big different to stack usage for some patterns.
2826
2827 26. If a subpattern containing a named recursion or subroutine reference such
2828 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2829 the space required for the compiled pattern went wrong and gave too small a
2830 value. Depending on the environment, this could lead to "Failed: internal
2831 error: code overflow at offset 49" or "glibc detected double free or
2832 corruption" errors.
2833
2834 27. Applied patches from Google (a) to support the new newline modes and (b) to
2835 advance over multibyte UTF-8 characters in GlobalReplace.
2836
2837 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2838 difference for some implementation of PCRE in some Windows version.
2839
2840 29. Added some extra testing facilities to pcretest:
2841
2842 \q<number> in a data line sets the "match limit" value
2843 \Q<number> in a data line sets the "match recursion limt" value
2844 -S <number> sets the stack size, where <number> is in megabytes
2845
2846 The -S option isn't available for Windows.
2847
2848
2849 Version 6.6 06-Feb-06
2850 ---------------------
2851
2852 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2853 in pcreposix.h. I have copied the definition from pcre.h.
2854
2855 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2856 because pcre.h is no longer a built file.
2857
2858 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2859 not normally included in the compiled code.
2860
2861
2862 Version 6.5 01-Feb-06
2863 ---------------------
2864
2865 1. When using the partial match feature with pcre_dfa_exec(), it was not
2866 anchoring the second and subsequent partial matches at the new starting
2867 point. This could lead to incorrect results. For example, with the pattern
2868 /1234/, partially matching against "123" and then "a4" gave a match.
2869
2870 2. Changes to pcregrep:
2871
2872 (a) All non-match returns from pcre_exec() were being treated as failures
2873 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2874 error message is output. Some extra information is given for the
2875 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2876 probably the only errors that are likely to be caused by users (by
2877 specifying a regex that has nested indefinite repeats, for instance).
2878 If there are more than 20 of these errors, pcregrep is abandoned.
2879
2880 (b) A binary zero was treated as data while matching, but terminated the
2881 output line if it was written out. This has been fixed: binary zeroes
2882 are now no different to any other data bytes.
2883
2884 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2885 used to set a locale for matching. The --locale=xxxx long option has
2886 been added (no short equivalent) to specify a locale explicitly on the
2887 pcregrep command, overriding the environment variables.
2888
2889 (d) When -B was used with -n, some line numbers in the output were one less
2890 than they should have been.
2891
2892 (e) Added the -o (--only-matching) option.
2893
2894 (f) If -A or -C was used with -c (count only), some lines of context were
2895 accidentally printed for the final match.
2896
2897 (g) Added the -H (--with-filename) option.
2898
2899 (h) The combination of options -rh failed to suppress file names for files
2900 that were found from directory arguments.
2901
2902 (i) Added the -D (--devices) and -d (--directories) options.
2903
2904 (j) Added the -F (--fixed-strings) option.
2905
2906 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2907
2908 (l) Added the --colo(u)r option.
2909
2910 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2911 is not present by default.
2912
2913 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2914 items such as (?R) or (?1), when the recursion could match a number of
2915 alternatives. If it matched one of the alternatives, but subsequently,
2916 outside the recursion, there was a failure, the code tried to back up into
2917 the recursion. However, because of the way PCRE is implemented, this is not
2918 possible, and the result was an incorrect result from the match.
2919
2920 In order to prevent this happening, the specification of recursion has
2921 been changed so that all such subpatterns are automatically treated as
2922 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2923
2924 4. I had overlooked the fact that, in some locales, there are characters for
2925 which isalpha() is true but neither isupper() nor islower() are true. In
2926 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2927 and ordfeminine) are like this. This affected the treatment of \w and \W
2928 when they appeared in character classes, but not when they appeared outside
2929 a character class. The bit map for "word" characters is now created
2930 separately from the results of isalnum() instead of just taking it from the
2931 upper, lower, and digit maps. (Plus the underscore character, of course.)
2932
2933 5. The above bug also affected the handling of POSIX character classes such as
2934 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2935 permanent tables. Instead, the bit maps for such a class were previously
2936 created as the appropriate unions of the upper, lower, and digit bitmaps.
2937 Now they are created by subtraction from the [[:word:]] class, which has
2938 its own bitmap.
2939
2940 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2941 It is created by subtracting the vertical space characters (\x09, \x0a,
2942 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2943 subtraction was done in the overall bitmap for a character class, meaning
2944 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2945 be recognized. This bug has been fixed.
2946
2947 7. Patches from the folks at Google:
2948
2949 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2950 real life, but is still worth protecting against".
2951
2952 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2953 regular expressions".
2954
2955 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2956 have it.
2957
2958 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2959 "configure" and the latter not, in order to fix a problem somebody had
2960 with compiling the Arg class on HP-UX.
2961
2962 (e) Improve the error-handling of the C++ wrapper a little bit.
2963
2964 (f) New tests for checking recursion limiting.
2965
2966 8. The pcre_memmove() function, which is used only if the environment does not
2967 have a standard memmove() function (and is therefore rarely compiled),
2968 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2969 returning a result (though PCRE never actually uses the result).
2970
2971 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2972 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2973 returned instead of calling malloc() with an overflowing number that would
2974 most likely cause subsequent chaos.
2975
2976 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2977
2978 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2979 with this option is matched, the nmatch and pmatch options of regexec() are
2980 ignored.
2981
2982 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2983 provided in case anyone wants to the the POSIX interface with UTF-8
2984 strings.
2985
2986 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2987 C++ linking (needed for some HP-UX environments).
2988
2989 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2990 (unused parameter) and in the pcre_printint() function (omitted "default"
2991 switch label when the default is to do nothing).
2992
2993 15. Added some code to make it possible, when PCRE is compiled as a C++
2994 library, to replace subject pointers for pcre_exec() with a smart pointer
2995 class, thus making it possible to process discontinuous strings.
2996
2997 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2998 much the same function. They were added by different people who were trying
2999 to make PCRE easy to compile on non-Unix systems. It has been suggested
3000 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3001 for compiling on Windows systems. I have therefore replaced it with
3002 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3003 defaults to "extern" for C or "extern C" for C++, which works fine on
3004 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3005 SCOPE with something explicit in config.h. In addition:
3006
3007 (a) pcreposix.h still had just "extern" instead of either of these macros;
3008 I have replaced it with PCRE_DATA_SCOPE.
3009
3010 (b) Functions such as _pcre_xclass(), which are internal to the library,
3011 but external in the C sense, all had PCRE_EXPORT in their definitions.
3012 This is apparently wrong for the Windows case, so I have removed it.
3013 (It makes no difference on Unix-like systems.)
3014
3015 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3016 of recursive calls to match(). This is different to MATCH_LIMIT because
3017 that limits the total number of calls to match(), not all of which increase
3018 the depth of recursion. Limiting the recursion depth limits the amount of
3019 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3020 when PCRE is compiled, and changed at run time. A patch from Google adds
3021 this functionality to the C++ interface.
3022
3023 18. Changes to the handling of Unicode character properties:
3024
3025 (a) Updated the table to Unicode 4.1.0.
3026
3027 (b) Recognize characters that are not in the table as "Cn" (undefined).
3028
3029 (c) I revised the way the table is implemented to a much improved format
3030 which includes recognition of ranges. It now supports the ranges that
3031 are defined in UnicodeData.txt, and it also amalgamates other
3032 characters into ranges. This has reduced the number of entries in the
3033 table from around 16,000 to around 3,000, thus reducing its size
3034 considerably. I realized I did not need to use a tree structure after
3035 all - a binary chop search is just as efficient. Having reduced the
3036 number of entries, I extended their size from 6 bytes to 8 bytes to
3037 allow for more data.
3038
3039 (d) Added support for Unicode script names via properties such as \p{Han}.
3040
3041 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3042 matching that character.
3043
3044 20. When matching a repeated Unicode property with a minimum greater than zero,
3045 (for example \pL{2,}), PCRE could look past the end of the subject if it
3046 reached it while seeking the minimum number of characters. This could
3047 happen only if some of the characters were more than one byte long, because
3048 there is a check for at least the minimum number of bytes.
3049
3050 21. Refactored the implementation of \p and \P so as to be more general, to
3051 allow for more different types of property in future. This has changed the
3052 compiled form incompatibly. Anybody with saved compiled patterns that use
3053 \p or \P will have to recompile them.
3054
3055 22. Added "Any" and "L&" to the supported property types.
3056
3057 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3058 but give a compile time error if the value is greater than 0xff.
3059
3060 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3061 accidentally not being installed or uninstalled.
3062
3063 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3064 made were to insert the current release number. This seemed silly, because
3065 it made things harder for people building PCRE on systems that don't run
3066 "configure". I have turned pcre.h into a distributed file, no longer built
3067 by "configure", with the version identification directly included. There is
3068 no longer a pcre.h.in file.
3069
3070 However, this change necessitated a change to the pcre-config script as
3071 well. It is built from pcre-config.in, and one of the substitutions was the
3072 release number. I have updated configure.ac so that ./configure now finds
3073 the release number by grepping pcre.h.
3074
3075 26. Added the ability to run the tests under valgrind.
3076
3077
3078 Version 6.4 05-Sep-05
3079 ---------------------
3080
3081 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3082 "--" to be printed when multiple files were scanned, even when none of the
3083 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3084 consider it to be a bug, and have restored the previous behaviour.
3085
3086 2. A couple of code tidies to get rid of compiler warnings.
3087
3088 3. The pcretest program used to cheat by referring to symbols in the library
3089 whose names begin with _pcre_. These are internal symbols that are not
3090 really supposed to be visible externally, and in some environments it is
3091 possible to suppress them. The cheating is now confined to including
3092 certain files from the library's source, which is a bit cleaner.
3093
3094 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3095 file's purpose clearer.
3096
3097 5. Reorganized pcre_ucp_findchar().
3098
3099
3100 Version 6.3 15-Aug-05
3101 ---------------------
3102
3103 1. The file libpcre.pc.in did not have general read permission in the tarball.
3104
3105 2. There were some problems when building without C++ support:
3106
3107 (a) If C++ support was not built, "make install" and "make test" still
3108 tried to test it.
3109
3110 (b) There were problems when the value of CXX was explicitly set. Some
3111 changes have been made to try to fix these, and ...
3112
3113 (c) --disable-cpp can now be used to explicitly disable C++ support.
3114
3115 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3116 backslash in a target when C++ was disabled. This confuses some
3117 versions of "make", apparently. Using an intermediate variable solves
3118 this. (Same for CPP_LOBJ.)
3119
3120 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3121 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3122 necessary on certain architectures.
3123
3124 4. Added a setting of -export-symbols-regex to the link command to remove
3125 those symbols that are exported in the C sense, but actually are local
3126 within the library, and not documented. Their names all begin with
3127 "_pcre_". This is not a perfect job, because (a) we have to except some
3128 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3129 available (and never for static libraries). I have made a note to try to
3130 find a way round (a) in the future.
3131
3132
3133 Version 6.2 01-Aug-05
3134 ---------------------
3135
3136 1. There was no test for integer overflow of quantifier values. A construction
3137 such as {1111111111111111} would give undefined results. What is worse, if
3138 a minimum quantifier for a parenthesized subpattern overflowed and became
3139 negative, the calculation of the memory size went wrong. This could have
3140 led to memory overwriting.
3141
3142 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3143
3144 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3145 operating environments where this matters.
3146
3147 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3148 PCRE options from within the C++ wrapper.
3149
3150 5. Named capturing subpatterns were not being correctly counted when a pattern
3151 was compiled. This caused two problems: (a) If there were more than 100
3152 such subpatterns, the calculation of the memory needed for the whole
3153 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3154 back references of the form \12, where the number was greater than 9, were
3155 not recognized as back references, even though there were sufficient
3156 previous subpatterns.
3157
3158 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3159 versions of gcc, e.g. 2.95.4.
3160
3161
3162 Version 6.1 21-Jun-05
3163 ---------------------
3164
3165 1. There was one reference to the variable "posix" in pcretest.c that was not
3166 surrounded by "#if !defined NOPOSIX".
3167
3168 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3169 the cross-check on the old pcre_info() function, for the benefit of the
3170 cut-down version of PCRE that is currently imported into Exim.
3171
3172 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3173 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3174 patterns that start (?i) but I don't think that matters. The use of (?i) is
3175 just an example; this all applies to the other options as well.
3176
3177 4. Since libtool seems to echo the compile commands it is issuing, the output
3178 from "make" can be reduced a bit by putting "@" in front of each libtool
3179 compile command.
3180
3181 5. Patch from the folks at Google for configure.in to be a bit more thorough
3182 in checking for a suitable C++ installation before trying to compile the
3183 C++ stuff. This should fix a reported problem when a compiler was present,
3184 but no suitable headers.
3185
3186 6. The man pages all had just "PCRE" as their title. I have changed them to
3187 be the relevant file name. I have also arranged that these names are
3188 retained in the file doc/pcre.txt, which is a concatenation in text format
3189 of all the man pages except the little individual ones for each function.
3190
3191 7. The NON-UNIX-USE file had not been updated for the different set of source
3192 files that come with release 6. I also added a few comments about the C++
3193 wrapper.
3194
3195
3196 Version 6.0 07-Jun-05
3197 ---------------------
3198
3199 1. Some minor internal re-organization to help with my DFA experiments.
3200
3201 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3202 didn't matter for the library itself when fully configured, but did matter
3203 when compiling without UCP support, or within Exim, where the ucp files are
3204 not imported.
3205
3206 3. Refactoring of the library code to split up the various functions into
3207 different source modules. The addition of the new DFA matching code (see
3208 below) to a single monolithic source would have made it really too
3209 unwieldy, quite apart from causing all the code to be include in a
3210 statically linked application, when only some functions are used. This is
3211 relevant even without the DFA addition now that patterns can be compiled in
3212 one application and matched in another.
3213
3214 The downside of splitting up is that there have to be some external
3215 functions and data tables that are used internally in different modules of
3216 the library but which are not part of the API. These have all had their
3217 names changed to start with "_pcre_" so that they are unlikely to clash
3218 with other external names.
3219
3220 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3221 a different (DFA) algorithm. Although it is slower than the original
3222 function, it does have some advantages for certain types of matching
3223 problem.
3224
3225 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3226 including restarting after a partial match.
3227
3228 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3229 defined when compiling for Windows was sent to me. I have put it into the
3230 code, though I have no means of testing or verifying it.
3231
3232 7. Added the pcre_refcount() auxiliary function.
3233
3234 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3235 match before or at the first newline in the subject string. In pcretest,
3236 the /f option on a pattern can be used to set this.
3237
3238 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3239 would behave wrongly. This has been present in PCRE since release 4.0.
3240
3241 10. A number of changes to the pcregrep command:
3242
3243 (a) Refactored how -x works; insert ^(...)$ instead of setting
3244 PCRE_ANCHORED and checking the length, in preparation for adding
3245 something similar for -w.
3246
3247 (b) Added the -w (match as a word) option.
3248
3249 (c) Refactored the way lines are read and buffered so as to have more
3250 than one at a time available.
3251
3252 (d) Implemented a pcregrep test script.
3253
3254 (e) Added the -M (multiline match) option. This allows patterns to match
3255 over several lines of the subject. The buffering ensures that at least
3256 8K, or the rest of the document (whichever is the shorter) is available
3257 for matching (and similarly the previous 8K for lookbehind assertions).
3258
3259 (f) Changed the --help output so that it now says
3260
3261 -w, --word-regex(p)
3262
3263 instead of two lines, one with "regex" and the other with "regexp"
3264 because that confused at least one person since the short forms are the
3265 same. (This required a bit of code, as the output is generated
3266 automatically from a table. It wasn't just a text change.)
3267
3268 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3269 option but starts with a hyphen. Could be a pattern or a path name
3270 starting with a hyphen, for instance.
3271
3272 (h) "-" can be given as a file name to represent stdin.
3273
3274 (i) When file names are being printed, "(standard input)" is used for
3275 the standard input, for compatibility with GNU grep. Previously
3276 "<stdin>" was used.
3277
3278 (j) The option --label=xxx can be used to supply a name to be used for
3279 stdin when file names are being printed. There is no short form.
3280
3281 (k) Re-factored the options decoding logic because we are going to add
3282 two more options that take data. Such options can now be given in four
3283 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3284
3285 (l) Added the -A, -B, and -C options for requesting that lines of context
3286 around matches be printed.
3287
3288 (m) Added the -L option to print the names of files that do not contain
3289 any matching lines, that is, the complement of -l.
3290
3291 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3292 continue to scan other files.
3293
3294 (o) The -s option was incorrectly implemented. For compatibility with other
3295 greps, it now suppresses the error message for a non-existent or non-
3296 accessible file (but not the return code). There is a new option called
3297 -q that suppresses the output of matching lines, which was what -s was
3298 previously doing.
3299
3300 (p) Added --include and --exclude options to specify files for inclusion
3301 and exclusion when recursing.
3302
3303 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3304 Hopefully, it now does.
3305
3306 12. Missing cast in pcre_study().
3307
3308 13. Added an "uninstall" target to the makefile.
3309
3310 14. Replaced "extern" in the function prototypes in Makefile.in with
3311 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3312 world, but is set differently for Windows.
3313
3314 15. Added a second compiling function called pcre_compile2(). The only
3315 difference is that it has an extra argument, which is a pointer to an
3316 integer error code. When there is a compile-time failure, this is set
3317 non-zero, in addition to the error test pointer being set to point to an
3318 error message. The new argument may be NULL if no error number is required
3319 (but then you may as well call pcre_compile(), which is now just a
3320 wrapper). This facility is provided because some applications need a
3321 numeric error indication, but it has also enabled me to tidy up the way
3322 compile-time errors are handled in the POSIX wrapper.
3323
3324 16. Added VPATH=.libs to the makefile; this should help when building with one
3325 prefix path and installing with another. (Or so I'm told by someone who
3326 knows more about this stuff than I do.)
3327
3328 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3329 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3330 match everything, including newlines. This is not POSIX-compatible, but
3331 somebody wanted the feature. From pcretest it can be activated by using
3332 both the P and the s flags.
3333
3334 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3335
3336 19. libpcre.pc was being incorrectly installed as executable.
3337
3338 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3339 it now also looks for '\r' so that it will work unmodified on Windows.
3340
3341 21. Added Google's contributed C++ wrapper to the distribution.
3342
3343 22. Added some untidy missing memory free() calls in pcretest, to keep
3344 Electric Fence happy when testing.
3345
3346
3347
3348 Version 5.0 13-Sep-04
3349 ---------------------
3350
3351 1. Internal change: literal characters are no longer packed up into items
3352 containing multiple characters in a single byte-string. Each character
3353 is now matched using a separate opcode. However, there may be more than one
3354 byte in the character in UTF-8 mode.
3355
3356 2. The pcre_callout_block structure has two new fields: pattern_position and
3357 next_item_length. These contain the offset in the pattern to the next match
3358 item, and its length, respectively.
3359
3360 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3361 insertion of callouts before each pattern item. Added the /C option to
3362 pcretest to make use of this.
3363
3364 4. On the advice of a Windows user, the lines
3365
3366 #if defined(_WIN32) || defined(WIN32)
3367 _setmode( _fileno( stdout ), 0x8000 );
3368 #endif /* defined(_WIN32) || defined(WIN32) */
3369
3370 have been added to the source of pcretest. This apparently does useful
3371 magic in relation to line terminators.
3372
3373 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3374 for the benefit of those environments where the "b" makes a difference.
3375
3376 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3377 to know about it. I have put a hack into configure.in that adds in code
3378 to set GCC=yes if CC=icc. This seems to end up at a point in the
3379 generated configure script that is early enough to affect the setting of
3380 compiler options, which is what is needed, but I have no means of testing
3381 whether it really works. (The user who reported this had patched the
3382 generated configure script, which of course I cannot do.)
3383
3384 LATER: After change 22 below (new libtool files), the configure script
3385 seems to know about icc (and also ecc). Therefore, I have commented out
3386 this hack in configure.in.
3387
3388 7. Added support for pkg-config (2 patches were sent in).
3389
3390 8. Negated POSIX character classes that used a combination of internal tables
3391 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3392 [[:^ascii]]. Typically, they would match almost any characters. The other
3393 POSIX classes were not broken in this way.
3394
3395 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3396 to find the match, as PCRE was deluded into thinking that the match had to
3397 start at the start point or following a newline. The same bug applied to
3398 patterns with negative forward assertions or any backward assertions
3399 preceding ".*" at the start, unless the pattern required a fixed first
3400 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3401
3402 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3403 starting at the last subject character, bytes beyond the end of the subject
3404 string were read.
3405
3406 11. Renamed the variable "class" as "classbits" to make life easier for C++
3407 users. (Previously there was a macro definition, but it apparently wasn't
3408 enough.)
3409
3410 12. Added the new field "tables" to the extra data so that tables can be passed
3411 in at exec time, or the internal tables can be re-selected. This allows
3412 a compiled regex to be saved and re-used at a later time by a different
3413 program that might have everything at different addresses.
3414
3415 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3416 -R library as well as a -L library.
3417
3418 14. The debugging options of pcretest (-d on the command line or D on a
3419 pattern) showed incorrect output for anything following an extended class
3420 that contained multibyte characters and which was followed by a quantifier.
3421
3422 15. Added optional support for general category Unicode character properties
3423 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3424 support. It adds about 90K to the size of the library. The meanings of the
3425 inbuilt class escapes such as \d and \s have NOT been changed.
3426
3427 16. Updated pcredemo.c to include calls to free() to release the memory for the
3428 compiled pattern.
3429
3430 17. The generated file chartables.c was being created in the source directory
3431 instead of in the building directory. This caused the build to fail if the
3432 source directory was different from the building directory, and was
3433 read-only.
3434
3435 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3436 file. No doubt somebody will tell me if they don't make sense... Also added
3437 Dan Mooney's comments about building on OpenVMS.
3438
3439 19. Added support for partial matching via the PCRE_PARTIAL option for
3440 pcre_exec() and the \P data escape in pcretest.
3441
3442 20. Extended pcretest with 3 new pattern features:
3443
3444 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3445 write the compiled pattern to the file whose name is "rest-of-line".
3446 This is a straight binary dump of the data, with the saved pointer to
3447 the character tables forced to be NULL. The study data, if any, is
3448 written too. After writing, pcretest reads a new pattern.
3449
3450 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3451 compiled pattern from the given file. There must not be any
3452 occurrences of "<" in the file name (pretty unlikely); if there are,
3453 pcretest will instead treat the initial "<" as a pattern delimiter.
3454 After reading in the pattern, pcretest goes on to read data lines as
3455 usual.
3456
3457 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3458 and 16-bit fields in a compiled pattern, to simulate a pattern that
3459 was compiled on a host of opposite endianness.
3460
3461 21. The pcre-exec() function can now cope with patterns that were compiled on
3462 hosts of opposite endianness, with this restriction:
3463
3464 As for any compiled expression that is saved and used later, the tables
3465 pointer field cannot be preserved; the extra_data field in the arguments
3466 to pcre_exec() should be used to pass in a tables address if a value
3467 other than the default internal tables were used at compile time.
3468
3469 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3470 now diagnosed as an error. Previously, most of the time, a negative number
3471 would have been treated as zero, but if in addition "ovector" was passed as
3472 NULL, a crash could occur.
3473
3474 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3475 new versions from the libtool 1.5 distribution (the last one is a copy of
3476 a file called libtool.m4). This seems to have fixed the need to patch
3477 "configure" to support Darwin 1.3 (which I used to do). However, I still
3478 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3479 workstation).
3480
3481 24. Changed the PCRE licence to be the more standard "BSD" licence.
3482
3483
3484 Version 4.5 01-Dec-03
3485 ---------------------
3486
3487 1. There has been some re-arrangement of the code for the match() function so
3488 that it can be compiled in a version that does not call itself recursively.
3489 Instead, it keeps those local variables that need separate instances for
3490 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3491 needs to "recurse". Keeping track of where control must go is done by means
3492 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3493 hide most of the details from the main code, and operates only if
3494 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3495 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3496 operating.
3497
3498 To make it easier for callers to provide specially tailored get/free
3499 functions for this usage, two new functions, pcre_stack_malloc, and
3500 pcre_stack_free, are used. They are always called in strict stacking order,
3501 and the size of block requested is always the same.
3502
3503 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3504 PCRE has been compiled to use the stack or the heap for recursion. The
3505 -C option of pcretest uses this to show which version is compiled.
3506
3507 A new data escape \S, is added to pcretest; it causes the amounts of store
3508 obtained and freed by both kinds of malloc/free at match time to be added
3509 to the output.
3510
3511 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3512 what's available on my current Linux desktop machine.
3513
3514 3. When matching a UTF-8 string, the test for a valid string at the start has
3515 been extended. If start_offset is not zero, PCRE now checks that it points
3516 to a byte that is the start of a UTF-8 character. If not, it returns
3517 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3518 this is necessary because there may be backward assertions in the pattern.
3519 When matching the same subject several times, it may save resources to use
3520 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3521
3522 4. The code for checking the validity of UTF-8 strings has been tightened so
3523 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3524 containing "overlong sequences".
3525
3526 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3527 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3528 should have been "&", but it just so happened that all the cases this let
3529 through by mistake were picked up later in the function.
3530
3531 6. I had used a variable called "isblank" - this is a C99 function, causing
3532 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3533
3534 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3535 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3536 so the version is shown; (c) stop "make test" echoing "./RunTest".
3537
3538 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3539
3540 9. The prototype for memmove() for systems that don't have it was using
3541 size_t, but the inclusion of the header that defines size_t was later. I've
3542 moved the #includes for the C headers earlier to avoid this.
3543
3544 10. Added some adjustments to the code to make it easier to compiler on certain
3545 special systems:
3546
3547 (a) Some "const" qualifiers were missing.
3548 (b) Added the macro EXPORT before all exported functions; by default this
3549 is defined to be empty.
3550 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3551 that it reads its output file name as an argument instead of writing
3552 to the standard output and assuming this can be redirected.
3553
3554 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3555 class containing characters with values greater than 255, PCRE compilation
3556 went into a loop.
3557
3558 12. A recursive reference to a subpattern that was within another subpattern
3559 that had a minimum quantifier of zero caused PCRE to crash. For example,
3560 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3561 recursion. If the recursively-called subpattern itself had a zero repeat,
3562 that was OK.
3563
3564 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3565 buffer into which it was copied (for escape processing) was still set at
3566 1024, so long lines caused crashes.
3567
3568 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3569 "internal error: code overflow...". This applied to any character class
3570 that was followed by a possessive quantifier.
3571
3572 15. Modified the Makefile to add libpcre.la as a prerequisite for
3573 libpcreposix.la because I was told this is needed for a parallel build to
3574 work.
3575
3576 16. If a pattern that contained .* following optional items at the start was
3577 studied, the wrong optimizing data was generated, leading to matching
3578 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3579 matching string must start with a or b or c. The correct conclusion for
3580 this pattern is that a match can start with any character.
3581
3582
3583 Version 4.4 13-Aug-03
3584 ---------------------
3585
3586 1. In UTF-8 mode, a character class containing characters with values between
3587 127 and 255 was not handled correctly if the compiled pattern was studied.
3588 In fixing this, I have also improved the studying algorithm for such
3589 classes (slightly).
3590
3591 2. Three internal functions had redundant arguments passed to them. Removal
3592 might give a very teeny performance improvement.
3593
3594 3. Documentation bug: the value of the capture_top field in a callout is *one
3595 more than* the number of the hightest numbered captured substring.
3596
3597 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3598 in incorrectly linking with a previously installed version. They now link
3599 explicitly with libpcre.la.
3600
3601 5. configure.in no longer needs to recognize Cygwin specially.
3602
3603 6. A problem in pcre.in for Windows platforms is fixed.
3604
3605 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3606 pcretest, it used to include the size of the study block as part of its
3607 output. Unfortunately, the structure contains a field that has a different
3608 size on different hardware architectures. This meant that the tests that
3609 showed this size failed. As the block is currently always of a fixed size,
3610 this information isn't actually particularly useful in pcretest output, so
3611 I have just removed it.
3612
3613 8. Three pre-processor statements accidentally did not start in column 1.
3614 Sadly, there are *still* compilers around that complain, even though
3615 standard C has not required this for well over a decade. Sigh.
3616
3617 9. In pcretest, the code for checking callouts passed small integers in the
3618 callout_data field, which is a void * field. However, some picky compilers
3619 complained about the casts involved for this on 64-bit systems. Now
3620 pcretest passes the address of the small integer instead, which should get
3621 rid of the warnings.
3622
3623 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3624 both compile and run time, and gives an error if an invalid UTF-8 sequence
3625 is found. There is a option for disabling this check in cases where the
3626 string is known to be correct and/or the maximum performance is wanted.
3627
3628 11. In response to a bug report, I changed one line in Makefile.in from
3629
3630 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3631 to
3632 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3633
3634 to look similar to other lines, but I have no way of telling whether this
3635 is the right thing to do, as I do not use Windows. No doubt I'll get told
3636 if it's wrong...
3637
3638
3639 Version 4.3 21-May-03
3640 ---------------------
3641
3642 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3643 Makefile.
3644
3645 2. Some refactoring to improve the quality of the code:
3646
3647 (i) The utf8_table... variables are now declared "const".
3648
3649 (ii) The code for \cx, which used the "case flipping" table to upper case
3650 lower case letters, now just substracts 32. This is ASCII-specific,
3651 but the whole concept of \cx is ASCII-specific, so it seems
3652 reasonable.
3653
3654 (iii) PCRE was using its character types table to recognize decimal and
3655 hexadecimal digits in the pattern. This is silly, because it handles
3656 only 0-9, a-f, and A-F, but the character types table is locale-
3657 specific, which means strange things might happen. A private
3658 table is now used for this - though it costs 256 bytes, a table is
3659 much faster than multiple explicit tests. Of course, the standard
3660 character types table is still used for matching digits in subject
3661 strings against \d.
3662
3663 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3664 ending in _t are). So I've renamed it as ESC_tee.
3665
3666 3. The first argument for regexec() in the POSIX wrapper should have been
3667 defined as "const".
3668
3669 4. Changed pcretest to use malloc() for its buffers so that they can be
3670 Electric Fenced for debugging.
3671
3672 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3673 to read one or more bytes before the start of the subject string. Often this
3674 had no effect on PCRE's behaviour, but in some circumstances it could
3675 provoke a segmentation fault.
3676
3677 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3678 to try to read one or more bytes before the start of the subject string.
3679
3680 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3681 UTF-8 support could misbehave in various ways if the subject string
3682 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3683 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3684 back over UTF-8 characters.)
3685
3686
3687 Version 4.2 14-Apr-03
3688 ---------------------
3689
3690 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3691
3692 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3693 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3694 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3695 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3696 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3697 and BUILD_EXEEXT
3698 Note: automatic setting of the BUILD variables is not yet working
3699 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3700 compile-time but not at link-time
3701 [LINK]: use for linking executables only
3702 make different versions for Windows and non-Windows
3703 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3704 libraries
3705 [LINK_FOR_BUILD]: new variable
3706 [OBJEXT]: use throughout
3707 [EXEEXT]: use throughout
3708 <winshared>: new target
3709 <wininstall>: new target
3710 <dftables.o>: use native compiler
3711 <dftables>: use native linker
3712 <install>: handle Windows platform correctly
3713 <clean>: ditto
3714 <check>: ditto
3715 copy DLL to top builddir before testing
3716
3717 As part of these changes, -no-undefined was removed again. This was reported
3718 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3719 in any case.
3720
3721 3. Some tidies to get rid of compiler warnings:
3722
3723 . In the match_data structure, match_limit was an unsigned long int, whereas
3724 match_call_count was an int. I've made them both unsigned long ints.
3725
3726 . In pcretest the fact that a const uschar * doesn't automatically cast to
3727 a void * provoked a warning.
3728
3729 . Turning on some more compiler warnings threw up some "shadow" variables
3730 and a few more missing casts.
3731
3732 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3733 option, a class that contained a single character with a value between 128
3734 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3735
3736 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3737 option, a class that contained several characters, but with at least one
3738 whose value was between 128 and 255 caused PCRE to crash.
3739
3740
3741 Version 4.1 12-Mar-03
3742 ---------------------
3743
3744 1. Compiling with gcc -pedantic found a couple of places where casts were
3745 needed, and a string in dftables.c that was longer than standard compilers are
3746 required to support.
3747
3748 2. Compiling with Sun's compiler found a few more places where the code could
3749 be tidied up in order to avoid warnings.
3750
3751 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3752 first of these names is deprecated in the latest Autoconf in favour of the name
3753 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3754 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3755 analogy I have changed it to CFLAGS_FOR_BUILD.
3756
3757 4. Added -no-undefined to the linking command in the Makefile, because this is
3758 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3759 linking step for the pcreposix library.
3760
3761 5. PCRE was failing to diagnose the case of two named groups with the same
3762 name.
3763
3764 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3765 literal character that is needed in the subject for a match, and scans along to
3766 ensure that it is present before embarking on the full matching process. This
3767 saves time in cases of nested unlimited repeats that are never going to match.
3768 Problem: the scan can take a lot of time if the subject is very long (e.g.
3769 megabytes), thus penalizing straightforward matches. It is now done only if the
3770 amount of subject to be scanned is less than 1000 bytes.
3771
3772 7. A lesser problem with the same optimization is that it was recording the
3773 first character of an anchored pattern as "needed", thus provoking a search
3774 right along the subject, even when the first match of the pattern was going to
3775 fail. The "needed" character is now not set for anchored patterns, unless it
3776 follows something in the pattern that is of non-fixed length. Thus, it still
3777 fulfils its original purpose of finding quick non-matches in cases of nested
3778 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3779
3780
3781 Version 4.0 17-Feb-03
3782 ---------------------
3783
3784 1. If a comment in an extended regex that started immediately after a meta-item
3785 extended to the end of string, PCRE compiled incorrect data. This could lead to
3786 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3787
3788 2. Moved to autoconf 2.53 and libtool 1.4.2.
3789
3790 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3791 the special perltest8 script is no longer needed - all the tests can be run
3792 from a single perltest script.
3793
3794 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3795 by \s. It has now been removed in PCRE. This means it isn't recognized as
3796 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3797 class [:space:] *does* include VT, thereby creating a mess.
3798
3799 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3800 space and tab.
3801
3802 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3803 its new features into the main test script, reducing the number of scripts.
3804
3805 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3806 were backward compatible, and made the (?i) apply to the whole pattern, as if
3807 /i were given. Now it behaves more logically, and applies the option setting
3808 only to what follows. PCRE has been changed to follow suit. However, if it
3809 finds options settings right at the start of the pattern, it extracts them into
3810 the global options, as before. Thus, they show up in the info data.
3811
3812 8. Added support for the \Q...\E escape sequence. Characters in between are
3813 treated as literals. This is slightly different from Perl in that $ and @ are
3814 also handled as literals inside the quotes. In Perl, they will cause variable
3815 interpolation. Note the following examples:
3816
3817 Pattern PCRE matches Perl matches
3818
3819 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3820 \Qabc\$xyz\E abc\$xyz abc\$xyz
3821 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3822
3823 For compatibility with Perl, \Q...\E sequences are recognized inside character
3824 classes as well as outside them.
3825
3826 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3827 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3828 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3829 signed/unsigned warnings.
3830
3831 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3832 option for pcretest, so I've replaced it by a simple function that does just
3833 that job.
3834
3835 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3836 "pcregrep -".
3837
3838 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3839 Java package. This provides some syntactic sugar for simple cases of what my
3840 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3841 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3842 item, you can use this simplified notation. Note that only makes sense with
3843 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3844 greediness, whatever the setting of the PCRE_UNGREEDY option.
3845
3846 13. A change of greediness default within a pattern was not taking effect at
3847 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3848 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3849 was abstracted outside.
3850
3851 14. PCRE now supports the \G assertion. It is true when the current matching
3852 position is at the start point of the match. This differs from \A when the
3853 starting offset is non-zero. Used with the /g option of pcretest (or similar
3854 code), it works in the same way as it does for Perl's /g option. If all
3855 alternatives of a regex begin with \G, the expression is anchored to the start
3856 match position, and the "anchored" flag is set in the compiled expression.
3857
3858 15. Some bugs concerning the handling of certain option changes within patterns
3859 have been fixed. These applied to options other than (?ims). For example,
3860 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3861 been the other way round. Some of this was related to change 7 above.
3862
3863 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3864 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3865 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3866 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3867
3868 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3869 mode. Unlike ".", it always matches newline, whatever the setting of
3870 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3871 assertions. Perl allows it, but it doesn't (in general) work because it can't
3872 calculate the length of the lookbehind. At least, that's the case for Perl
3873 5.8.0 - I've been told they are going to document that it doesn't work in
3874 future.
3875
3876 18. Added an error diagnosis for escapes that PCRE does not support: these are
3877 \L, \l, \N, \P, \p, \U, \u, and \X.
3878
3879 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3880 reading past the end of the pattern in cases such as /[abcd/.
3881
3882 20. PCRE was getting more memory than necessary for patterns with classes that
3883 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3884
3885 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3886 compiling PCRE for use with Virtual Pascal.
3887
3888 22. Small fix to the Makefile to make it work properly if the build is done
3889 outside the source tree.
3890
3891 23. Added a new extension: a condition to go with recursion. If a conditional
3892 subpattern starts with (?(R) the "true" branch is used if recursion has
3893 happened, whereas the "false" branch is used only at the top level.
3894
3895 24. When there was a very long string of literal characters (over 255 bytes
3896 without UTF support, over 250 bytes with UTF support), the computation of how
3897 much memory was required could be incorrect, leading to segfaults or other
3898 strange effects.
3899
3900 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3901 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3902 there was a subsequent back reference to those brackets. This meant that, for
3903 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3904 possible to check for precisely this case. All we can do is abandon the
3905 optimization if .* occurs inside capturing brackets when there are any back
3906 references whatsoever. (See below for a better fix that came later.)
3907
3908 26. The handling of the optimization for finding the first character of a
3909 non-anchored pattern, and for finding a character that is required later in the
3910 match were failing in some cases. This didn't break the matching; it just
3911 failed to optimize when it could. The way this is done has been re-implemented.
3912
3913 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3914
3915 28. Added a new feature that provides some of the functionality that Perl
3916 provides with (?{...}). The facility is termed a "callout". The way it is done
3917 in PCRE is for the caller to provide an optional function, by setting
3918 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3919 global variable. By default it is unset, which disables all calling out. To get
3920 the function called, the regex must include (?C) at appropriate points. This
3921 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3922 This provides a means of identifying different callout points. When PCRE
3923 reaches such a point in the regex, if pcre_callout has been set, the external
3924 function is called. It is provided with data in a structure called
3925 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3926 matching continues; if it returns a non-zero value, the match at the current
3927 point fails. However, backtracking will occur if possible. [This was changed
3928 later and other features added - see item 49 below.]
3929
3930 29. pcretest is upgraded to test the callout functionality. It provides a
3931 callout function that displays information. By default, it shows the start of
3932 the match and the current position in the text. There are some new data escapes
3933 to vary what happens:
3934
3935 \C+ in addition, show current contents of captured substrings
3936 \C- do not supply a callout function
3937 \C!n return 1 when callout number n is reached
3938 \C!n!m return 1 when callout number n is reached for the mth time
3939
3940 30. If pcregrep was called with the -l option and just a single file name, it
3941 output "<stdin>" if a match was found, instead of the file name.
3942
3943 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3944 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3945 pcre_exec(). This saves a malloc/free per call. The default value of
3946 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3947 when configuring.
3948
3949 32. The default maximum size of a compiled pattern is 64K. There have been a
3950 few cases of people hitting this limit. The code now uses macros to handle the
3951 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3952 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3953 configuring. Tests 2 and 5 work only with 2-byte links because they output
3954 debugging information about compiled patterns.
3955
3956 33. Internal code re-arrangements:
3957
3958 (a) Moved the debugging function for printing out a compiled regex into
3959 its own source file (printint.c) and used #include to pull it into
3960 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3961 separate copies.
3962
3963 (b) Defined the list of op-code names for debugging as a macro in
3964 internal.h so that it is next to the definition of the opcodes.
3965
3966 (c) Defined a table of op-code lengths for simpler skipping along compiled
3967 code. This is again a macro in internal.h so that it is next to the
3968 definition of the opcodes.
3969
3970 34. Added support for recursive calls to individual subpatterns, along the
3971 lines of Robin Houston's patch (but implemented somewhat differently).
3972
3973 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3974 allow it to read and process whole directories in Win32. This code was
3975 contributed by Lionel Fourquaux; it has not been tested by me.
3976
3977 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3978 used to name a group. Names consist of alphanumerics and underscores, and must
3979 be unique. Back references use the syntax (?P=name) and recursive calls use
3980 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3981 numbers. The function pcre_fullinfo() can be used after compilation to extract
3982 a name/number map. There are three relevant calls:
3983
3984 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3985 PCRE_INFO_NAMECOUNT yields the number of entries
3986 PCRE_INFO_NAMETABLE yields a pointer to the map.
3987
3988 The map is a vector of fixed-size entries. The size of each entry depends on
3989 the length of the longest name used. The first two bytes of each entry are the
3990 group number, most significant byte first. There follows the corresponding
3991 name, zero terminated. The names are in alphabetical order.
3992
3993 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3994 case instead of 255. Making it the same both with and without UTF-8 support
3995 means that the same test output works with both.
3996
3997 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3998 calling malloc() with a zero argument.
3999
4000 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4001 optimization. I've improved things by keeping a bitmap of backreferences with
4002 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4003 fact referenced, the optimization can be applied. It is unlikely that a
4004 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4005 the match to follow \n) will appear inside brackets with a number greater than
4006 31, but if it does, any back reference > 31 suppresses the optimization.
4007
4008 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4009 of disabling numbered capturing parentheses. Any opening parenthesis that is
4010 not followed by ? behaves as if it were followed by ?: but named parentheses
4011 can still be used for capturing (and they will acquire numbers in the usual
4012 way).
4013
4014 41. Redesigned the return codes from the match() function into yes/no/error so
4015 that errors can be passed back from deep inside the nested calls. A malloc
4016 failure while inside a recursive subpattern call now causes the
4017 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4018
4019 42. It is now possible to set a limit on the number of times the match()
4020 function is called in a call to pcre_exec(). This facility makes it possible to
4021 limit the amount of recursion and backtracking, though not in a directly
4022 obvious way, because the match() function is used in a number of different
4023 circumstances. The count starts from zero for each position in the subject
4024 string (for non-anchored patterns). The default limit is, for compatibility, a
4025 large number, namely 10 000 000. You can change this in two ways:
4026
4027 (a) When configuring PCRE before making, you can use --with-match-limit=n
4028 to set a default value for the compiled library.
4029
4030 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4031 a different value is set. See 45 below.
4032
4033 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4034
4035 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4036 of things that can be changed at compile time. The first argument specifies
4037 what is wanted and the second points to where the information is to be placed.
4038 The current list of available information is:
4039
4040 PCRE_CONFIG_UTF8
4041
4042 The output is an integer that is set to one if UTF-8 support is available;
4043 otherwise it is set to zero.
4044
4045 PCRE_CONFIG_NEWLINE
4046
4047 The output is an integer that it set to the value of the code that is used for
4048 newline. It is either LF (10) or CR (13).
4049
4050 PCRE_CONFIG_LINK_SIZE
4051
4052 The output is an integer that contains the number of bytes used for internal
4053 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4054
4055 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4056
4057 The output is an integer that contains the threshold above which the POSIX
4058 interface uses malloc() for output vectors. See item 31 above.
4059
4060 PCRE_CONFIG_MATCH_LIMIT
4061
4062 The output is an unsigned integer that contains the default limit of the number
4063 of match() calls in a pcre_exec() execution. See 42 above.
4064
4065 44. pcretest has been upgraded by the addition of the -C option. This causes it
4066 to extract all the available output from the new pcre_config() function, and to
4067 output it. The program then exits immediately.
4068
4069 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4070 order to support additional features. One way would have been to define
4071 pcre_exec2() (for example) with extra arguments, but this would not have been
4072 extensible, and would also have required all calls to the original function to
4073 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4074 is used for passing in "extra" data from pcre_study().
4075
4076 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4077 contains the following fields:
4078
4079 flags a bitmap indicating which of the following fields are set
4080 study_data opaque data from pcre_study()
4081 match_limit a way of specifying a limit on match() calls for a specific
4082 call to pcre_exec()
4083 callout_data data for callouts (see 49 below)
4084
4085 The flag bits are also defined in pcre.h, and are
4086
4087 PCRE_EXTRA_STUDY_DATA
4088 PCRE_EXTRA_MATCH_LIMIT
4089 PCRE_EXTRA_CALLOUT_DATA
4090
4091 The pcre_study() function now returns one of these new pcre_extra blocks, with
4092 the actual study data pointed to by the study_data field, and the
4093 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4094 before. That is, this change is entirely upwards-compatible and requires no
4095 change to existing code.
4096
4097 If you want to pass in additional data to pcre_exec(), you can either place it
4098 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4099 block.
4100
4101 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4102 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4103 times with different match limits, until it finds the minimum value needed for
4104 pcre_exec() to complete. The value is then output. This can be instructive; for
4105 most simple matches the number is quite small, but for pathological cases it
4106 gets very large very quickly.
4107
4108 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4109 returns the size of the data block pointed to by the study_data field in a
4110 pcre_extra block, that is, the value that was passed as the argument to
4111 pcre_malloc() when PCRE was getting memory in which to place the information
4112 created by pcre_study(). The fourth argument should point to a size_t variable.
4113 pcretest has been extended so that this information is shown after a successful
4114 pcre_study() call when information about the compiled regex is being displayed.
4115
4116 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4117 because what follows is always an absolute path. (Later: it turns out that this
4118 is more than cosmetic for MinGW, because it doesn't like empty path
4119 components.)
4120
4121 49. Some changes have been made to the callout feature (see 28 above):
4122
4123 (i) A callout function now has three choices for what it returns:
4124
4125 0 => success, carry on matching
4126 > 0 => failure at this point, but backtrack if possible
4127 < 0 => serious error, return this value from pcre_exec()
4128
4129 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4130 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4131 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4132 use by callout functions. It will never be used by PCRE itself.
4133
4134 (ii) The pcre_extra structure (see 45 above) has a void * field called
4135 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4136 pcre_callout_block structure has a field of the same name. The contents of
4137 the field passed in the pcre_extra structure are passed to the callout
4138 function in the corresponding field in the callout block. This makes it
4139 easier to use the same callout-containing regex from multiple threads. For
4140 testing, the pcretest program has a new data escape
4141
4142 \C*n pass the number n (may be negative) as callout_data
4143
4144 If the callout function in pcretest receives a non-zero value as
4145 callout_data, it returns that value.
4146
4147 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4148 there were some redundant $(CFLAGS) in commands that are now specified as
4149 $(LINK), which already includes $(CFLAGS).
4150
4151 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4152 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4153 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4154 one-byte characters throughout. Note that case-insensitive matching applies
4155 only to characters whose values are less than 256. PCRE doesn't support the
4156 notion of cases for higher-valued characters.
4157
4158 (i) A character class whose characters are all within 0-255 is handled as
4159 a bit map, and the map is inverted for negative classes. Previously, a
4160 character > 255 always failed to match such a class; however it should
4161 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4162
4163 (ii) A negated character class with a single character < 255 is coded as
4164 "not this character" (OP_NOT). This wasn't working properly when the test
4165 character was multibyte, either singly or repeated.
4166
4167 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4168 mode, for example: \x{100}{2,3}.
4169
4170 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4171 singly or repeated) now correctly test multibyte characters. However,
4172 PCRE doesn't recognize any characters with values greater than 255 as
4173 digits, spaces, or word characters. Such characters always match \D, \S,
4174 and \W, and never match \d, \s, or \w.
4175
4176 (v) Classes may now contain characters and character ranges with values
4177 greater than 255. For example: [ab\x{100}-\x{400}].
4178
4179 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4180 PCRE in UTF-8 mode.
4181
4182 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4183 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4184 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4185 value.)
4186
4187 53. The single man page has become too large. I have therefore split it up into
4188 a number of separate man pages. These also give rise to individual HTML pages;
4189 these are now put in a separate directory, and there is an index.html page that
4190 lists them all. Some hyperlinking between the pages has been installed.
4191
4192 54. Added convenience functions for handling named capturing parentheses.
4193
4194 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4195 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4196 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4197 are faulted.
4198
4199 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4200 calling configure. These values are used when compiling the dftables.c program
4201 which is run to generate the source of the default character tables. They
4202 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4203 you will need to set these values.
4204
4205 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4206
4207
4208 Version 3.9 02-Jan-02
4209 ---------------------
4210
4211 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4212
4213 2. If --disable-static was given, the building process failed when trying to
4214 build pcretest and pcregrep. (For some reason it was using libtool to compile
4215 them, which is not right, as they aren't part of the library.)
4216
4217
4218 Version 3.8 18-Dec-01
4219 ---------------------
4220
4221 1. The experimental UTF-8 code was completely screwed up. It was packing the
4222 bytes in the wrong order. How dumb can you get?
4223
4224
4225 Version 3.7 29-Oct-01
4226 ---------------------
4227
4228 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4229 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4230 this didn't happen under Solaris 8, where I normally test things.
4231
4232 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4233 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4234 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4235 this fix an hour or so after the initial 3.7 release.)
4236
4237
4238 Version 3.6 23-Oct-01
4239 ---------------------
4240
4241 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4242 offsets passed as NULL with zero offset count.
4243
4244 2. The config.guess and config.sub files had not been updated when I moved to
4245 the latest autoconf.
4246
4247
4248 Version 3.5 15-Aug-01
4249 ---------------------
4250
4251 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4252 had been forgotten.
4253
4254 2. By using declared but undefined structures, we can avoid using "void"
4255 definitions in pcre.h while keeping the internal definitions of the structures
4256 private.
4257
4258 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4259 user point of view, this means that both static and shared libraries are built
4260 by default, but this can be individually controlled. More of the work of
4261 handling this static/shared cases is now inside libtool instead of PCRE's make
4262 file.
4263
4264 4. The pcretest utility is now installed along with pcregrep because it is
4265 useful for users (to test regexs) and by doing this, it automatically gets
4266 relinked by libtool. The documentation has been turned into a man page, so
4267 there are now .1, .txt, and .html versions in /doc.
4268
4269 5. Upgrades to pcregrep:
4270 (i) Added long-form option names like gnu grep.
4271 (ii) Added --help to list all options with an explanatory phrase.
4272 (iii) Added -r, --recursive to recurse into sub-directories.
4273 (iv) Added -f, --file to read patterns from a file.
4274
4275 6. pcre_exec() was referring to its "code" argument before testing that
4276 argument for NULL (and giving an error if it was NULL).
4277
4278 7. Upgraded Makefile.in to allow for compiling in a different directory from
4279 the source directory.
4280
4281 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4282 options bits, the pointer it was passed was to an int instead of to an unsigned
4283 long int. This mattered only on 64-bit systems.
4284
4285 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4286 generated) instead of pcre.in, which it its source. Also made the same change
4287 in several of the .c files.
4288
4289 10. A new release of gcc defines printf() as a macro, which broke pcretest
4290 because it had an ifdef in the middle of a string argument for printf(). Fixed
4291 by using separate calls to printf().
4292
4293 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4294 script, to force use of CR or LF instead of \n in the source. On non-Unix
4295 systems, the value can be set in config.h.
4296
4297 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4298 absolute limit. Changed the text of the error message to make this clear, and
4299 likewise updated the man page.
4300
4301 13. The limit of 99 on the number of capturing subpatterns has been removed.
4302 The new limit is 65535, which I hope will not be a "real" limit.
4303
4304
4305 Version 3.4 22-Aug-00
4306 ---------------------
4307
4308 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4309
4310 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4311
4312
4313 Version 3.3 01-Aug-00
4314 ---------------------
4315
4316 1. If an octal character was given, but the value was greater than \377, it
4317 was not getting masked to the least significant bits, as documented. This could
4318 lead to crashes in some systems.
4319
4320 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4321 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4322
4323 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4324 These just pass their arguments on to (pcre_free)(), but they are provided
4325 because some uses of PCRE bind it to non-C systems that can call its functions,
4326 but cannot call free() or pcre_free() directly.
4327
4328 4. Add "make test" as a synonym for "make check". Corrected some comments in
4329 the Makefile.
4330
4331 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4332 Makefile.
4333
4334 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4335 command called pgrep for grepping around the active processes.
4336
4337 7. Added the beginnings of support for UTF-8 character strings.
4338
4339 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4340 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4341 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4342 out for the ar command.)
4343
4344
4345 Version 3.2 12-May-00
4346 ---------------------
4347
4348 This is purely a bug fixing release.
4349
4350 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4351 of ZA. This was just one example of several cases that could provoke this bug,
4352 which was introduced by change 9 of version 2.00. The code for breaking
4353 infinite loops after an iteration that matches an empty string was't working
4354 correctly.
4355
4356 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4357 when matched against abbab (for example). After matching an empty string, it
4358 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4359 caused it to match further down the string than it should.
4360
4361 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4362 was there because it doesn't seem to be needed, and it causes trouble on some
4363 systems, as it is not a Standard C header. It has been removed.
4364
4365 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4366 were reported on the Macintosh. The changes were from
4367
4368 while ((c = *(++ptr)) != 0 && c != '\n');
4369 to
4370 while ((c = *(++ptr)) != 0 && c != '\n') ;
4371
4372 Totally extraordinary, but if that's what it takes...
4373
4374 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4375 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4376 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4377 assumes the way PCRE uses memmove() (always moving upwards).
4378
4379 6. PCRE is being used in one environment where strchr() is not available. There
4380 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4381 faster code anyway.
4382
4383
4384 Version 3.1 09-Feb-00
4385 ---------------------
4386
4387 The only change in this release is the fixing of some bugs in Makefile.in for
4388 the "install" target:
4389
4390 (1) It was failing to install pcreposix.h.
4391
4392 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4393
4394
4395 Version 3.0 01-Feb-00
4396 ---------------------
4397
4398 1. Add support for the /+ modifier to perltest (to output $` like it does in
4399 pcretest).
4400
4401 2. Add support for the /g modifier to perltest.
4402
4403 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4404 matches null strings.
4405
4406 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4407 pattern. Perl treats empty patterns specially - it reuses the most recent
4408 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4409 effect.
4410
4411 5. The POSIX interface was broken in that it was just handing over the POSIX
4412 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4413 required a bigger vector, with some working space on the end. This means that
4414 the POSIX wrapper now has to get and free some memory, and copy the results.
4415
4416 6. Added some simple autoconf support, placing the test data and the
4417 documentation in separate directories, re-organizing some of the
4418 information files, and making it build pcre-config (a GNU standard). Also added
4419 libtool support for building PCRE as a shared library, which is now the
4420 default.
4421
4422 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4423 09 are not valid octal constants. Single digits will be used for minor values
4424 less than 10.
4425
4426 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4427 existing programs that set these in the POSIX interface can use PCRE without
4428 modification.
4429
4430 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4431 return all that pcre_info() returns, plus additional data. The pcre_info()
4432 function is retained for compatibility, but is considered to be obsolete.
4433
4434 10. Added experimental recursion feature (?R) to handle one common case that
4435 Perl 5.6 will be able to do with (?p{...}).
4436
4437 11. Added support for POSIX character classes like [:alpha:], which Perl is
4438 adopting.
4439
4440
4441 Version 2.08 31-Aug-99
4442 ----------------------
4443
4444 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4445 trying to match at the startoffset position, but instead was moving forward to
4446 the next newline as if a previous match had failed.
4447
4448 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4449 and could get into a loop if a null string was matched other than at the start
4450 of the subject.
4451
4452 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4453 be distinguished at compile time, and for completeness also added PCRE_DATE.
4454
4455 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4456 in GnuWin32 environments.
4457
4458
4459 Version 2.07 29-Jul-99
4460 ----------------------
4461
4462 1. The documentation is now supplied in plain text form and HTML as well as in
4463 the form of man page sources.
4464
4465 2. C++ compilers don't like assigning (void *) values to other pointer types.
4466 In particular this affects malloc(). Although there is no problem in Standard
4467 C, I've put in casts to keep C++ compilers happy.
4468
4469 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4470 should be (const char *).
4471
4472 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4473 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4474 However, I haven't made this a standard facility. The documentation doesn't
4475 mention it, and the Makefile doesn't support it.
4476
4477 5. The Makefile now contains an "install" target, with editable destinations at
4478 the top of the file. The pcretest program is not installed.
4479
4480 6. pgrep -V now gives the PCRE version number and date.
4481
4482 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4483 causing the entire string to be ignored, instead of just the last character.
4484
4485 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4486 non-matching string, it can take a very, very long time, even for strings of
4487 quite modest length, because of the nested recursion. PCRE now does better in
4488 some of these cases. It does this by remembering the last required literal
4489 character in the pattern, and pre-searching the subject to ensure it is present
4490 before running the real match. In other words, it applies a heuristic to detect
4491 some types of certain failure quickly, and in the above example, if presented
4492 with a string that has no trailing " it gives "no match" very quickly.
4493
4494 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4495 other alternatives are tried instead.
4496
4497
4498 Version 2.06 09-Jun-99
4499 ----------------------
4500
4501 1. Change pcretest's output for amount of store used to show just the code
4502 space, because the remainder (the data block) varies in size between 32-bit and
4503 64-bit systems.
4504
4505 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4506 start matching at. This allows lookbehinds to work when searching for multiple
4507 occurrences in a string.
4508
4509 3. Added additional options to pcretest for testing multiple occurrences:
4510
4511 /+ outputs the rest of the string that follows a match
4512 /g loops for multiple occurrences, using the new startoffset argument
4513 /G loops for multiple occurrences by passing an incremented pointer
4514
4515 4. PCRE wasn't doing the "first character" optimization for patterns starting
4516 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4517 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4518 the letter 'x'. On long subject strings, this gives a significant speed-up.
4519
4520
4521 Version 2.05 21-Apr-99
4522 ----------------------
4523
4524 1. Changed the type of magic_number from int to long int so that it works
4525 properly on 16-bit systems.
4526
4527 2. Fixed a bug which caused patterns starting with .* not to work correctly
4528 when the subject string contained newline characters. PCRE was assuming
4529 anchoring for such patterns in all cases, which is not correct because .* will
4530 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4531 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4532 must be retried after every newline in the subject.
4533
4534
4535 Version 2.04 18-Feb-99
4536 ----------------------
4537
4538 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4539 computation of the store needed to hold the pattern was incorrect (too large).
4540 If such patterns were nested a few deep, this could multiply and become a real
4541 problem.
4542
4543 2. Added /M option to pcretest to show the memory requirement of a specific
4544 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4545
4546 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4547 compiled in such a way that the backtracking after subsequent failure was
4548 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4549 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4550
4551
4552 Version 2.03 02-Feb-99
4553 ----------------------
4554
4555 1. Fixed typo and small mistake in man page.
4556
4557 2. Added 4th condition (GPL supersedes if conflict) and created separate
4558 LICENCE file containing the conditions.
4559
4560 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4561 Perl, that is the internal \ allows the delimiter to be included in the
4562 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4563 the final delimiter, add \ to the end of the pattern (to test the error).
4564
4565 4. Added the convenience functions for extracting substrings after a successful
4566 match. Updated pcretest to make it able to test these functions.
4567
4568
4569 Version 2.02 14-Jan-99
4570 ----------------------
4571
4572 1. Initialized the working variables associated with each extraction so that
4573 their saving and restoring doesn't refer to uninitialized store.
4574
4575 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4576 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4577 fix the problem.
4578
4579 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4580 calls, and wasn't printing the correct value for compiling calls. Increased the
4581 default value of LOOPREPEAT, and the number of significant figures in the
4582 times.
4583
4584 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4585
4586 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4587 a building problem on Windows NT with a FAT file system.
4588
4589
4590 Version 2.01 21-Oct-98
4591 ----------------------
4592
4593 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4594 to character tables built by pcre_maketables() in the current locale. If NULL
4595 is passed, the default tables are used.
4596
4597
4598 Version 2.00 24-Sep-98
4599 ----------------------
4600
4601 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4602 it any more.
4603
4604 2. Allow quantification of (?>) groups, and make it work correctly.
4605
4606 3. The first character computation wasn't working for (?>) groups.
4607
4608 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4609 end of the subject) and add 5.005's \z, which really does match only at the
4610 very end of the subject.
4611
4612 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4613
4614 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4615 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4616 localized options. All options to pcre_study() were also removed.
4617
4618 7. Add other new features from 5.005:
4619
4620 $(?<= positive lookbehind
4621 $(?<! negative lookbehind
4622 (?imsx-imsx) added the unsetting capability
4623 such a setting is global if at outer level; local otherwise
4624 (?imsx-imsx:) non-capturing groups with option setting
4625 (?(cond)re|re) conditional pattern matching
4626
4627 A backreference to itself in a repeated group matches the previous
4628 captured string.
4629
4630 8. General tidying up of studying (both automatic and via "study")
4631 consequential on the addition of new assertions.
4632
4633 9. As in 5.005, unlimited repeated groups that could match an empty substring
4634 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4635 runtime if any iteration does actually match an empty substring.
4636
4637 10. Include the RunTest script in the distribution.
4638
4639 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4640 discrepancies, some of which were old and were also with respect to 5.004. They
4641 have now been fixed.
4642
4643
4644 Version 1.09 28-Apr-98
4645 ----------------------
4646
4647 1. A negated single character class followed by a quantifier with a minimum
4648 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4649 program crashes, or just wrong answers. This did not apply to negated classes
4650 containing more than one character, or to minima other than one.
4651
4652
4653 Version 1.08 27-Mar-98
4654 ----------------------
4655
4656 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4657
4658 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4659 latter must appear before anything that relies on it in the pattern.
4660
4661
4662 Version 1.07 16-Feb-98
4663 ----------------------
4664
4665 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4666 repeat of a potentially empty string).
4667
4668
4669 Version 1.06 23-Jan-98
4670 ----------------------
4671
4672 1. Added Markus Oberhumer's little patches for C++.
4673
4674 2. Literal strings longer than 255 characters were broken.
4675
4676
4677 Version 1.05 23-Dec-97
4678 ----------------------
4679
4680 1. Negated character classes containing more than one character were failing if
4681 PCRE_CASELESS was set at run time.
4682
4683
4684 Version 1.04 19-Dec-97
4685 ----------------------
4686
4687 1. Corrected the man page, where some "const" qualifiers had been omitted.
4688
4689 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4690 input syntax.
4691
4692 3. Fixed memory leak which occurred when a regex with back references was
4693 matched with an offsets vector that wasn't big enough. The temporary memory
4694 that is used in this case wasn't being freed if the match failed.
4695
4696 4. Tidied pcretest to ensure it frees memory that it gets.
4697
4698 5. Temporary memory was being obtained in the case where the passed offsets
4699 vector was exactly big enough.
4700
4701 6. Corrected definition of offsetof() from change 5 below.
4702
4703 7. I had screwed up change 6 below and broken the rules for the use of
4704 setjmp(). Now fixed.
4705
4706
4707 Version 1.03 18-Dec-97
4708 ----------------------
4709
4710 1. A erroneous regex with a missing opening parenthesis was correctly
4711 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4712 on some systems.
4713
4714 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4715 it was reported that one broken compiler failed on the former because "code" is
4716 also an independent variable.
4717
4718 3. The erroneous regex a[]b caused an array overrun reference.
4719
4720 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4721 fail on data ending with that character. (It was going on too far, and checking
4722 the next character, typically a binary zero.) This was specific to the
4723 optimized code for single-character negative classes.
4724
4725 5. Added a contributed patch from the TIN world which does the following:
4726
4727 + Add an undef for memmove, in case the the system defines a macro for it.
4728
4729 + Add a definition of offsetof(), in case there isn't one. (I don't know
4730 the reason behind this - offsetof() is part of the ANSI standard - but
4731 it does no harm).
4732
4733 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4734 most of the places where whitespace preceded '#'. I have given up and
4735 allowed the remaining 2 cases to be at the margin.
4736
4737 + Rename some variables in pcre to eliminate shadowing. This seems very
4738 pedantic, but does no harm, of course.
4739
4740 6. Moved the call to setjmp() into its own function, to get rid of warnings
4741 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4742
4743 7. Constructs such as \d{8,} were compiling into the equivalent of
4744 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4745 outcome, but in this particular case used more store than had been allocated,
4746 which caused the bug to be discovered because it threw up an internal error.
4747
4748 8. The debugging code in both pcre and pcretest for outputting the compiled
4749 form of a regex was going wrong in the case of back references followed by
4750 curly-bracketed repeats.
4751
4752
4753 Version 1.02 12-Dec-97
4754 ----------------------
4755
4756 1. Typos in pcre.3 and comments in the source fixed.
4757
4758 2. Applied a contributed patch to get rid of places where it used to remove
4759 'const' from variables, and fixed some signed/unsigned and uninitialized
4760 variable warnings.
4761
4762 3. Added the "runtest" target to Makefile.
4763
4764 4. Set default compiler flag to -O2 rather than just -O.
4765
4766
4767 Version 1.01 19-Nov-97
4768 ----------------------
4769
4770 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4771 like /([ab]*)*/, that is, for classes with more than one character in them.
4772
4773 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4774 as /((?>a*))*/ (a PCRE_EXTRA facility).
4775
4776
4777 Version 1.00 18-Nov-97
4778 ----------------------
4779
4780 1. Added compile-time macros to support systems such as SunOS4 which don't have
4781 memmove() or strerror() but have other things that can be used instead.
4782
4783 2. Arranged that "make clean" removes the executables.
4784
4785
4786 Version 0.99 27-Oct-97
4787 ----------------------
4788
4789 1. Fixed bug in code for optimizing classes with only one character. It was
4790 initializing a 32-byte map regardless, which could cause it to run off the end
4791 of the memory it had got.
4792
4793 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4794
4795
4796 Version 0.98 22-Oct-97
4797 ----------------------
4798
4799 1. Fixed bug in code for handling temporary memory usage when there are more
4800 back references than supplied space in the ovector. This could cause segfaults.
4801
4802
4803 Version 0.97 21-Oct-97
4804 ----------------------
4805
4806 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4807
4808 2. Optimized negated single characters not to use a bit map.
4809
4810 3. Brought error texts together as macro definitions; clarified some of them;
4811 fixed one that was wrong - it said "range out of order" when it meant "invalid
4812 escape sequence".
4813
4814 4. Changed some char * arguments to const char *.
4815
4816 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4817
4818 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4819 pcretest.
4820
4821
4822 Version 0.96 16-Oct-97
4823 ----------------------
4824
4825 1. Added a simple "pgrep" utility to the distribution.
4826
4827 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4828 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4829 where "ddd" means "one or more decimal digits".
4830
4831 3. Fixed serious bug. If a pattern had a back reference, but the call to
4832 pcre_exec() didn't supply a large enough ovector to record the related
4833 identifying subpattern, the match always failed. PCRE now remembers the number
4834 of the largest back reference, and gets some temporary memory in which to save
4835 the offsets during matching if necessary, in order to ensure that
4836 backreferences always work.
4837
4838 4. Increased the compatibility with Perl in a number of ways:
4839
4840 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4841 to request this handling. The option can be set at compile or exec time.
4842
4843 (b) $ matches before a terminating newline by default; an option
4844 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4845 mode). The option can be set at compile or exec time.
4846
4847 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4848 the same as Perl's. If the decimal number it represents is less than 10
4849 or there aren't that many previous left capturing parentheses, an octal
4850 escape is read. Inside a character class, it's always an octal escape,
4851 even if it is a single digit.
4852
4853 (d) An escaped but undefined alphabetic character is taken as a literal,
4854 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4855 escapes.
4856
4857 (e) {0} is now permitted. (The previous item is removed from the compiled
4858 pattern).
4859
4860 5. Changed all the names of code files so that the basic parts are no longer
4861 than 10 characters, and abolished the teeny "globals.c" file.
4862
4863 6. Changed the handling of character classes; they are now done with a 32-byte
4864 bit map always.
4865
4866 7. Added the -d and /D options to pcretest to make it possible to look at the
4867 internals of compilation without having to recompile pcre.
4868
4869
4870 Version 0.95 23-Sep-97
4871 ----------------------
4872
4873 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4874 \x20 at the start of a run of normal characters. These were being treated as
4875 real characters, instead of the source characters being re-checked.
4876
4877
4878 Version 0.94 18-Sep-97
4879 ----------------------
4880
4881 1. The functions are now thread-safe, with the caveat that the global variables
4882 containing pointers to malloc() and free() or alternative functions are the
4883 same for all threads.
4884
4885 2. Get pcre_study() to generate a bitmap of initial characters for non-
4886 anchored patterns when this is possible, and use it if passed to pcre_exec().
4887
4888
4889 Version 0.93 15-Sep-97
4890 ----------------------
4891
4892 1. /(b)|(:+)/ was computing an incorrect first character.
4893
4894 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4895 but not actually doing anything yet.
4896
4897 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4898 as Perl does (e.g. [-az] or [az-]).
4899
4900 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4901 all possible positions.
4902
4903 5. Split up into different modules to avoid including unneeded functions in a
4904 compiled binary. However, compile and exec are still in one module. The "study"
4905 function is split off.
4906
4907 6. The character tables are now in a separate module whose source is generated
4908 by an auxiliary program - but can then be edited by hand if required. There are
4909 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4910 toupper() in the code.
4911
4912 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4913 make them global. Abolish the function for setting them, as the caller can now
4914 set them directly.
4915
4916
4917 Version 0.92 11-Sep-97
4918 ----------------------
4919
4920 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4921 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4922
4923 2. Caseless matching was not working in character classes if the characters in
4924 the pattern were in upper case.
4925
4926 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4927
4928 4. Make PCRE_ANCHORED public and accept as a compile option.
4929
4930 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4931 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4932 pass them.
4933
4934 6. Give an error if bad option bits passed at compile or run time.
4935
4936 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4937 pcretest to cause it to pass that flag.
4938
4939 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4940 options, and the first character, if set.
4941
4942 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4943
4944
4945 Version 0.91 10-Sep-97
4946 ----------------------
4947
4948 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4949 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4950
4951 2. PCRE was looping on encountering an indefinitely repeated back reference to
4952 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4953 Perl does - treats the match as successful.
4954
4955 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5