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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1217 - (show annotations)
Fri Nov 9 17:07:01 2012 UTC (6 years, 9 months ago) by ph10
File size: 239879 byte(s)
Preparations for 8.32-RC1.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.32 11-November-2012
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 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
142 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
143 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
144
145 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
146 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
147 the "old" RFC 2279). Instead, it ended up passing a negative length to
148 pcre_exec().
149
150 31. Add support for GCC's visibility feature to hide internal functions.
151
152 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
153 "unknown -C option" after outputting 0 or 1.
154
155 33. There is now support for generating a code coverage report for the test
156 suite in environments where gcc is the compiler and lcov is installed. This
157 is mainly for the benefit of the developers.
158
159 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
160 unaddressable using valgrind annotations, allowing valgrind to detect
161 invalid memory accesses. This is mainly for the benefit of the developers.
162
163
164 Version 8.31 06-July-2012
165 -------------------------
166
167 1. Fixing a wrong JIT test case and some compiler warnings.
168
169 2. Removed a bashism from the RunTest script.
170
171 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
172 to unsigned type, result still unsigned" that was given by an MS compiler
173 on encountering the code "-sizeof(xxx)".
174
175 4. Partial matching support is added to the JIT compiler.
176
177 5. Fixed several bugs concerned with partial matching of items that consist
178 of more than one character:
179
180 (a) /^(..)\1/ did not partially match "aba" because checking references was
181 done on an "all or nothing" basis. This also applied to repeated
182 references.
183
184 (b) \R did not give a hard partial match if \r was found at the end of the
185 subject.
186
187 (c) \X did not give a hard partial match after matching one or more
188 characters at the end of the subject.
189
190 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
191 a partial match for the string "\r".
192
193 (e) When newline was set to CRLF, the metacharacter "." did not recognize
194 a partial match for a CR character at the end of the subject string.
195
196 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
197 running pcretest, the text "(JIT)" added to the output whenever JIT is
198 actually used to run the match.
199
200 7. Individual JIT compile options can be set in pcretest by following -s+[+]
201 or /S+[+] with a digit between 1 and 7.
202
203 8. OP_NOT now supports any UTF character not just single-byte ones.
204
205 9. (*MARK) control verb is now supported by the JIT compiler.
206
207 10. The command "./RunTest list" lists the available tests without actually
208 running any of them. (Because I keep forgetting what they all are.)
209
210 11. Add PCRE_INFO_MAXLOOKBEHIND.
211
212 12. Applied a (slightly modified) user-supplied patch that improves performance
213 when the heap is used for recursion (compiled with --disable-stack-for-
214 recursion). Instead of malloc and free for each heap frame each time a
215 logical recursion happens, frames are retained on a chain and re-used where
216 possible. This sometimes gives as much as 30% improvement.
217
218 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
219 call.
220
221 14. As documented, (*COMMIT) is now confined to within a positive assertion.
222
223 15. It is now possible to link pcretest with libedit as an alternative to
224 libreadline.
225
226 16. (*COMMIT) control verb is now supported by the JIT compiler.
227
228 17. The Unicode data tables have been updated to Unicode 6.1.0.
229
230 18. Added --file-list option to pcregrep.
231
232 19. Added binary file support to pcregrep, including the -a, --binary-files,
233 -I, and --text options.
234
235 20. The madvise function is renamed for posix_madvise for QNX compatibility
236 reasons. Fixed by Giuseppe D'Angelo.
237
238 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
239
240 22. Changed the option for creating symbolic links for 16-bit man pages from
241 -s to -sf so that re-installing does not cause issues.
242
243 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
244
245 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
246 matches in certain environments (the workspace was not being correctly
247 retained). Also added to pcre_dfa_exec() a simple plausibility check on
248 some of the workspace data at the beginning of a restart.
249
250 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
251 was not doing so when it should - probably a typo introduced by SVN 528
252 (change 8.10/14).
253
254 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
255 \w+ when the character tables indicated that \x{c4} was a word character.
256 There were several related cases, all because the tests for doing a table
257 lookup were testing for characters less than 127 instead of 255.
258
259 27. If a pattern contains capturing parentheses that are not used in a match,
260 their slots in the ovector are set to -1. For those that are higher than
261 any matched groups, this happens at the end of processing. In the case when
262 there were back references that the ovector was too small to contain
263 (causing temporary malloc'd memory to be used during matching), and the
264 highest capturing number was not used, memory off the end of the ovector
265 was incorrectly being set to -1. (It was using the size of the temporary
266 memory instead of the true size.)
267
268 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
269 ovector size, it uses memory at the end of the block that it has got.
270
271 29. Check for an overlong MARK name and give an error at compile time. The
272 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
273
274 30. JIT compiler update.
275
276 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
277 Rill for the patch.
278
279 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
280
281 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
282
283 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
284 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
285 was enabled.
286
287 35. Improve JIT code generation for greedy plus quantifier.
288
289 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
290 1 to "aa" instead of to an empty string. The bug affected repeated groups
291 that could potentially match an empty string.
292
293 37. Optimizing single character iterators in JIT.
294
295 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
296 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
297 codepoints that are too big for the mode are faulted, and in a UTF mode,
298 disallowed codepoints are also faulted.
299
300 39. If PCRE was compiled with UTF support, in three places in the DFA
301 matcher there was code that should only have been obeyed in UTF mode, but
302 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
303 processing when bytes with values greater than 127 were present. In 16-bit
304 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
305 both cases the values are those that cannot be the first data item in a UTF
306 character. The three items that might have provoked this were recursions,
307 possessively repeated groups, and atomic groups.
308
309 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
310 and pcregrep, because some OS require shared objects to be explicitly
311 passed to ld, causing the link step to fail if they are not.
312
313 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
314 mode, patterns that started with \h* or \R* might be incorrectly matched.
315
316
317 Version 8.30 04-February-2012
318 -----------------------------
319
320 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
321 name is defined in ctype.h.
322
323 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
324 only in quite long subpatterns.
325
326 3. Removed the function pcre_info(), which has been obsolete and deprecated
327 since it was replaced by pcre_fullinfo() in February 2000.
328
329 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
330 match a (*MARK), and the match failed at the start of the subject, a
331 reference to memory before the start of the subject could occur. This bug
332 was introduced by fix 17 of release 8.21.
333
334 5. A reference to an unset group with zero minimum repetition was giving
335 totally wrong answers (in non-JavaScript-compatibility mode). For example,
336 /(another)?(\1?)test/ matched against "hello world test". This bug was
337 introduced in release 8.13.
338
339 6. Add support for 16-bit character strings (a large amount of work involving
340 many changes and refactorings).
341
342 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
343 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
344 from a file.
345
346 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
347 rounding is not applied in this particular case).
348
349 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
350 if they appear, or are escaped, in patterns.
351
352 10. Get rid of a number of -Wunused-but-set-variable warnings.
353
354 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
355 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
356 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
357 also returns the mark "x". This bug applied to capturing parentheses,
358 non-capturing parentheses, and atomic parentheses. It also applied to some
359 assertions.
360
361 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
362 information out of configure.ac instead of relying on pcre.h.generic, which
363 is not stored in the repository.
364
365 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
366 -lreadline.
367
368 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
369
370 15. Applied Graycode's patch to put the top-level frame on the stack rather
371 than the heap when not using the stack for recursion. This gives a
372 performance improvement in many cases when recursion is not deep.
373
374 16. Experimental code added to "pcretest -C" to output the stack frame size.
375
376
377 Version 8.21 12-Dec-2011
378 ------------------------
379
380 1. Updating the JIT compiler.
381
382 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
383 are added as well.
384
385 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
386 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
387 calling _pcre_jit_exec. Some extra comments are added.
388
389 4. (*MARK) settings inside atomic groups that do not contain any capturing
390 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
391 was introduced by change 18 for 8.20.
392
393 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
394 ECMA-262 standard.
395
396 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
397 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
398 This bug was probably introduced by change 9 of 8.13.
399
400 7. While fixing 6 above, I noticed that a number of other items were being
401 incorrectly rejected as "not fixed length". This arose partly because newer
402 opcodes had not been added to the fixed-length checking code. I have (a)
403 corrected the bug and added tests for these items, and (b) arranged for an
404 error to occur if an unknown opcode is encountered while checking for fixed
405 length instead of just assuming "not fixed length". The items that were
406 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
407 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
408 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
409
410 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
411 being incorrectly compiled and would have given unpredicatble results.
412
413 9. A possessively repeated subpattern with minimum repeat count greater than
414 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
415 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
416 the first (A) could occur when it should not.
417
418 10. Add a cast and remove a redundant test from the code.
419
420 11. JIT should use pcre_malloc/pcre_free for allocation.
421
422 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
423 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
424 is anything other than /usr, -L is still shown).
425
426 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
427
428 14. Perl does not support \N without a following name in a [] class; PCRE now
429 also gives an error.
430
431 15. If a forward reference was repeated with an upper limit of around 2000,
432 it caused the error "internal error: overran compiling workspace". The
433 maximum number of forward references (including repeats) was limited by the
434 internal workspace, and dependent on the LINK_SIZE. The code has been
435 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
436 the default depends on LINK_SIZE. There is a new upper limit (for safety)
437 of around 200,000 forward references. While doing this, I also speeded up
438 the filling in of repeated forward references.
439
440 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
441 incorrectly expecting the subject to contain another "a" after the start.
442
443 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
444 in the match, the SKIP should be ignored. This was not happening; instead
445 the SKIP was being treated as NOMATCH. For patterns such as
446 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
447 tested.
448
449 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
450 now much more compatible with Perl, in particular in cases where the result
451 is a non-match for a non-anchored pattern. For example, if
452 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
453 "m", where previously it did not return a name. A side effect of this
454 change is that for partial matches, the last encountered mark name is
455 returned, as for non matches. A number of tests that were previously not
456 Perl-compatible have been moved into the Perl-compatible test files. The
457 refactoring has had the pleasing side effect of removing one argument from
458 the match() function, thus reducing its stack requirements.
459
460 19. If the /S+ option was used in pcretest to study a pattern using JIT,
461 subsequent uses of /S (without +) incorrectly behaved like /S+.
462
463 21. Retrieve executable code size support for the JIT compiler and fixing
464 some warnings.
465
466 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
467 not work when the shorter character appeared right at the end of the
468 subject string.
469
470 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
471 systems.
472
473 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
474 output it when the /M option is used in pcretest.
475
476 25. The CheckMan script was not being included in the distribution. Also, added
477 an explicit "perl" to run Perl scripts from the PrepareRelease script
478 because this is reportedly needed in Windows.
479
480 26. If study data was being save in a file and studying had not found a set of
481 "starts with" bytes for the pattern, the data written to the file (though
482 never used) was taken from uninitialized memory and so caused valgrind to
483 complain.
484
485 27. Updated RunTest.bat as provided by Sheri Pierce.
486
487 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
488
489 29. Computation of memory usage for the table of capturing group names was
490 giving an unnecessarily large value.
491
492
493 Version 8.20 21-Oct-2011
494 ------------------------
495
496 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
497 a POSIX class. After further experiments with Perl, which convinced me that
498 Perl has bugs and confusions, a closing square bracket is no longer allowed
499 in a POSIX name. This bug also affected patterns with classes that started
500 with full stops.
501
502 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
503 captured substring, but while checking the failing first alternative,
504 substring 1 is temporarily captured. If the output vector supplied to
505 pcre_exec() was not big enough for this capture, the yield of the function
506 was still zero ("insufficient space for captured substrings"). This cannot
507 be totally fixed without adding another stack variable, which seems a lot
508 of expense for a edge case. However, I have improved the situation in cases
509 such as /(a)(b)x|abc/ matched against "abc", where the return code
510 indicates that fewer than the maximum number of slots in the ovector have
511 been set.
512
513 3. Related to (2) above: when there are more back references in a pattern than
514 slots in the output vector, pcre_exec() uses temporary memory during
515 matching, and copies in the captures as far as possible afterwards. It was
516 using the entire output vector, but this conflicts with the specification
517 that only 2/3 is used for passing back captured substrings. Now it uses
518 only the first 2/3, for compatibility. This is, of course, another edge
519 case.
520
521 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
522 main code base, and can be used by building with --enable-jit. When this is
523 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
524 runtime --no-jit option is given.
525
526 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
527 ovector, the return from the function was zero, implying that there were
528 other matches that did not fit. The correct "exactly full" value is now
529 returned.
530
531 6. If a subpattern that was called recursively or as a subroutine contained
532 (*PRUNE) or any other control that caused it to give a non-standard return,
533 invalid errors such as "Error -26 (nested recursion at the same subject
534 position)" or even infinite loops could occur.
535
536 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
537 computing the minimum length on reaching *ACCEPT, and so ended up with the
538 wrong value of 1 rather than 0. Further investigation indicates that
539 computing a minimum subject length in the presence of *ACCEPT is difficult
540 (think back references, subroutine calls), and so I have changed the code
541 so that no minimum is registered for a pattern that contains *ACCEPT.
542
543 8. If (*THEN) was present in the first (true) branch of a conditional group,
544 it was not handled as intended. [But see 16 below.]
545
546 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
547 Sheri Pierce.
548
549 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
550 the first byte in a match must be "a".
551
552 11. Change 17 for 8.13 increased the recursion depth for patterns like
553 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
554 pattern contains any instances of (*THEN). If it does not, the old
555 optimizations are restored. It would be nice to do this on a per-group
556 basis, but at the moment that is not feasible.
557
558 12. In some environments, the output of pcretest -C is CRLF terminated. This
559 broke RunTest's code that checks for the link size. A single white space
560 character after the value is now allowed for.
561
562 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
563 For "fr", it uses the Windows-specific input and output files.
564
565 14. If (*THEN) appeared in a group that was called recursively or as a
566 subroutine, it did not work as intended. [But see next item.]
567
568 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
569 pattern fragments (but not containing any | characters). If A and B are
570 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
571 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
572 D. In other words, Perl considers parentheses that do not contain any |
573 characters to be part of a surrounding alternative, whereas PCRE was
574 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
575 differently. PCRE now behaves in the same way as Perl, except in the case
576 of subroutine/recursion calls such as (?1) which have in any case always
577 been different (but PCRE had them first :-).
578
579 16. Related to 15 above: Perl does not treat the | in a conditional group as
580 creating alternatives. Such a group is treated in the same way as an
581 ordinary group without any | characters when processing (*THEN). PCRE has
582 been changed to match Perl's behaviour.
583
584 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
585 RunGrepTest script failed.
586
587 18. Change 22 for version 13 caused atomic groups to use more stack. This is
588 inevitable for groups that contain captures, but it can lead to a lot of
589 stack use in large patterns. The old behaviour has been restored for atomic
590 groups that do not contain any capturing parentheses.
591
592 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
593 suppress the check for a minimum subject length at run time. (If it was
594 given to pcre_exec() or pcre_dfa_exec() it did work.)
595
596 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
597 fail to work when decoding hex characters in data strings in EBCDIC
598 environments.
599
600 21. It appears that in at least one Mac OS environment, the isxdigit() function
601 is implemented as a macro that evaluates to its argument more than once,
602 contravening the C 90 Standard (I haven't checked a later standard). There
603 was an instance in pcretest which caused it to go wrong when processing
604 \x{...} escapes in subject strings. The has been rewritten to avoid using
605 things like p++ in the argument of isxdigit().
606
607
608 Version 8.13 16-Aug-2011
609 ------------------------
610
611 1. The Unicode data tables have been updated to Unicode 6.0.0.
612
613 2. Two minor typos in pcre_internal.h have been fixed.
614
615 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
616 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
617 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
618
619 4. There were a number of related bugs in the code for matching backrefences
620 caselessly in UTF-8 mode when codes for the characters concerned were
621 different numbers of bytes. For example, U+023A and U+2C65 are an upper
622 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
623 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
624 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
625 2-byte code at the end of the subject (it thought there wasn't enough data
626 left).
627
628 5. Comprehensive information about what went wrong is now returned by
629 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
630 as the output vector has at least 2 elements. The offset of the start of
631 the failing character and a reason code are placed in the vector.
632
633 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
634 now returned is for the first byte of the failing character, instead of the
635 last byte inspected. This is an incompatible change, but I hope it is small
636 enough not to be a problem. It makes the returned offset consistent with
637 pcre_exec() and pcre_dfa_exec().
638
639 7. pcretest now gives a text phrase as well as the error number when
640 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
641 failure, the offset and reason code are output.
642
643 8. When \R was used with a maximizing quantifier it failed to skip backwards
644 over a \r\n pair if the subsequent match failed. Instead, it just skipped
645 back over a single character (\n). This seems wrong (because it treated the
646 two characters as a single entity when going forwards), conflicts with the
647 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
648 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
649 has been changed.
650
651 9. Some internal refactoring has changed the processing so that the handling
652 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
653 time (the PCRE_DOTALL option was changed this way some time ago: version
654 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
655 which was always a bit of a fudge. It also means that there is one less
656 argument for the match() function, which reduces its stack requirements
657 slightly. This change also fixes an incompatibility with Perl: the pattern
658 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
659
660 10. More internal refactoring has drastically reduced the number of recursive
661 calls to match() for possessively repeated groups such as (abc)++ when
662 using pcre_exec().
663
664 11. While implementing 10, a number of bugs in the handling of groups were
665 discovered and fixed:
666
667 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
668 (a|)*(?1) gave a compile-time internal error.
669 ((a|)+)+ did not notice that the outer group could match an empty string.
670 (^a|^)+ was not marked as anchored.
671 (.*a|.*)+ was not marked as matching at start or after a newline.
672
673 12. Yet more internal refactoring has removed another argument from the match()
674 function. Special calls to this function are now indicated by setting a
675 value in a variable in the "match data" data block.
676
677 13. Be more explicit in pcre_study() instead of relying on "default" for
678 opcodes that mean there is no starting character; this means that when new
679 ones are added and accidentally left out of pcre_study(), testing should
680 pick them up.
681
682 14. The -s option of pcretest has been documented for ages as being an old
683 synonym of -m (show memory usage). I have changed it to mean "force study
684 for every regex", that is, assume /S for every regex. This is similar to -i
685 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
686 using it. It makes it easier to run collections of tests with and without
687 study enabled, and thereby test pcre_study() more easily. All the standard
688 tests are now run with and without -s (but some patterns can be marked as
689 "never study" - see 20 below).
690
691 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
692 restoration of the capturing data to the outer values was not happening
693 correctly.
694
695 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
696 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
697 pattern had matched an empty string, and so incorrectly returned a no
698 match.
699
700 17. There was optimizing code for the last branch of non-capturing parentheses,
701 and also for the obeyed branch of a conditional subexpression, which used
702 tail recursion to cut down on stack usage. Unfortunately, now that there is
703 the possibility of (*THEN) occurring in these branches, tail recursion is
704 no longer possible because the return has to be checked for (*THEN). These
705 two optimizations have therefore been removed. [But see 8.20/11 above.]
706
707 18. If a pattern containing \R was studied, it was assumed that \R always
708 matched two bytes, thus causing the minimum subject length to be
709 incorrectly computed because \R can also match just one byte.
710
711 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
712 was incorrectly computed.
713
714 20. If /S is present twice on a test pattern in pcretest input, it now
715 *disables* studying, thereby overriding the use of -s on the command line
716 (see 14 above). This is necessary for one or two tests to keep the output
717 identical in both cases.
718
719 21. When (*ACCEPT) was used in an assertion that matched an empty string and
720 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
721
722 22. When an atomic group that contained a capturing parenthesis was
723 successfully matched, but the branch in which it appeared failed, the
724 capturing was not being forgotten if a higher numbered group was later
725 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
726 group 1 to "a", when in fact it should be unset. This applied to multi-
727 branched capturing and non-capturing groups, repeated or not, and also to
728 positive assertions (capturing in negative assertions does not happen
729 in PCRE) and also to nested atomic groups.
730
731 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
732 subject after a captured substring, to make it easier to tell which of a
733 number of identical substrings has been captured.
734
735 24. The way atomic groups are processed by pcre_exec() has been changed so that
736 if they are repeated, backtracking one repetition now resets captured
737 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
738 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
739 "aaa". Previously, it would have been "a". As part of this code
740 refactoring, the way recursive calls are handled has also been changed.
741
742 25. If an assertion condition captured any substrings, they were not passed
743 back unless some other capturing happened later. For example, if
744 (?(?=(a))a) was matched against "a", no capturing was returned.
745
746 26. When studying a pattern that contained subroutine calls or assertions,
747 the code for finding the minimum length of a possible match was handling
748 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
749 group 1 called group 2 while simultaneously a separate group 2 called group
750 1). A stack overflow occurred in this case. I have fixed this by limiting
751 the recursion depth to 10.
752
753 27. Updated RunTest.bat in the distribution to the version supplied by Tom
754 Fortmann. This supports explicit test numbers on the command line, and has
755 argument validation and error reporting.
756
757 28. An instance of \X with an unlimited repeat could fail if at any point the
758 first character it looked at was a mark character.
759
760 29. Some minor code refactoring concerning Unicode properties and scripts
761 should reduce the stack requirement of match() slightly.
762
763 30. Added the '=' option to pcretest to check the setting of unused capturing
764 slots at the end of the pattern, which are documented as being -1, but are
765 not included in the return count.
766
767 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
768 compiled something random. Now it gives a compile-time error (as does
769 Perl).
770
771 32. A *MARK encountered during the processing of a positive assertion is now
772 recorded and passed back (compatible with Perl).
773
774 33. If --only-matching or --colour was set on a pcregrep call whose pattern
775 had alternative anchored branches, the search for a second match in a line
776 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
777 matched the line "0102" twice. The same bug affected patterns that started
778 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
779 twice.
780
781 34. Previously, PCRE did not allow quantification of assertions. However, Perl
782 does, and because of capturing effects, quantifying parenthesized
783 assertions may at times be useful. Quantifiers are now allowed for
784 parenthesized assertions.
785
786 35. A minor code tidy in pcre_compile() when checking options for \R usage.
787
788 36. \g was being checked for fancy things in a character class, when it should
789 just be a literal "g".
790
791 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
792 appearance of a nested POSIX class supersedes an apparent external class.
793 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
794 unescaped square brackets may also appear as part of class names. For
795 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
796 more like Perl. (But see 8.20/1 above.)
797
798 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
799 was because it thought it was \N{name}, which is not supported).
800
801 39. Add minix to OS list not supporting the -S option in pcretest.
802
803 40. PCRE tries to detect cases of infinite recursion at compile time, but it
804 cannot analyze patterns in sufficient detail to catch mutual recursions
805 such as ((?1))((?2)). There is now a runtime test that gives an error if a
806 subgroup is called recursively as a subpattern for a second time at the
807 same position in the subject string. In previous releases this might have
808 been caught by the recursion limit, or it might have run out of stack.
809
810 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
811 happen only once. PCRE was, however incorrectly giving a compile time error
812 "recursive call could loop indefinitely" because it cannot analyze the
813 pattern in sufficient detail. The compile time test no longer happens when
814 PCRE is compiling a conditional subpattern, but actual runaway loops are
815 now caught at runtime (see 40 above).
816
817 42. It seems that Perl allows any characters other than a closing parenthesis
818 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
819 has been changed to be the same.
820
821 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
822 as not to get warnings when autogen.sh is called. Also changed
823 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
824
825 44. To help people who use pcregrep to scan files containing exceedingly long
826 lines, the following changes have been made:
827
828 (a) The default value of the buffer size parameter has been increased from
829 8K to 20K. (The actual buffer used is three times this size.)
830
831 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
832 PCRE is built.
833
834 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
835 to be set at run time.
836
837 (d) Numerical values in pcregrep options can be followed by K or M, for
838 example --buffer-size=50K.
839
840 (e) If a line being scanned overflows pcregrep's buffer, an error is now
841 given and the return code is set to 2.
842
843 45. Add a pointer to the latest mark to the callout data block.
844
845 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
846 partial match of an empty string instead of no match. This was specific to
847 the use of ".".
848
849 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
850 complete match instead of a partial match. This bug was dependent on both
851 the PCRE_UTF8 and PCRE_DOTALL options being set.
852
853 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
854 starting byte set, because \b was not being ignored.
855
856
857 Version 8.12 15-Jan-2011
858 ------------------------
859
860 1. Fixed some typos in the markup of the man pages, and wrote a script that
861 checks for such things as part of the documentation building process.
862
863 2. On a big-endian 64-bit system, pcregrep did not correctly process the
864 --match-limit and --recursion-limit options (added for 8.11). In
865 particular, this made one of the standard tests fail. (The integer value
866 went into the wrong half of a long int.)
867
868 3. If the --colour option was given to pcregrep with -v (invert match), it
869 did strange things, either producing crazy output, or crashing. It should,
870 of course, ignore a request for colour when reporting lines that do not
871 match.
872
873 4. Another pcregrep bug caused similar problems if --colour was specified with
874 -M (multiline) and the pattern match finished with a line ending.
875
876 5. In pcregrep, when a pattern that ended with a literal newline sequence was
877 matched in multiline mode, the following line was shown as part of the
878 match. This seems wrong, so I have changed it.
879
880 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
881 the check for further matches in the same line (so they could be coloured)
882 to overrun the end of the current line. If another match was found, it was
883 incorrectly shown (and then shown again when found in the next line).
884
885 7. If pcregrep was compiled under Windows, there was a reference to the
886 function pcregrep_exit() before it was defined. I am assuming this was
887 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
888 reported by a user. I've moved the definition above the reference.
889
890
891 Version 8.11 10-Dec-2010
892 ------------------------
893
894 1. (*THEN) was not working properly if there were untried alternatives prior
895 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
896 backtracked to try for "b" instead of moving to the next alternative branch
897 at the same level (in this case, to look for "c"). The Perl documentation
898 is clear that when (*THEN) is backtracked onto, it goes to the "next
899 alternative in the innermost enclosing group".
900
901 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
902 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
903 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
904 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
905 (*THEN).
906
907 3. If \s appeared in a character class, it removed the VT character from
908 the class, even if it had been included by some previous item, for example
909 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
910 of \s, but is part of the POSIX "space" class.)
911
912 4. A partial match never returns an empty string (because you can always
913 match an empty string at the end of the subject); however the checking for
914 an empty string was starting at the "start of match" point. This has been
915 changed to the "earliest inspected character" point, because the returned
916 data for a partial match starts at this character. This means that, for
917 example, /(?<=abc)def/ gives a partial match for the subject "abc"
918 (previously it gave "no match").
919
920 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
921 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
922 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
923 has an implication that the given string is incomplete (because a partial
924 match is preferred over a full match). For this reason, these items now
925 give a partial match in this situation. [Aside: previously, the one case
926 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
927 match rather than a full match, which was wrong by the old rules, but is
928 now correct.]
929
930 6. There was a bug in the handling of #-introduced comments, recognized when
931 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
932 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
933 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
934 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
935 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
936 places in pcre_compile().
937
938 7. Related to (6) above, when pcre_compile() was skipping #-introduced
939 comments when looking ahead for named forward references to subpatterns,
940 the only newline sequence it recognized was NL. It now handles newlines
941 according to the set newline convention.
942
943 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
944 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
945 cater for a lack of strerror(). These oversights have been fixed.
946
947 9. Added --match-limit and --recursion-limit to pcregrep.
948
949 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
950
951 11. When the -o option was used, pcregrep was setting a return code of 1, even
952 when matches were found, and --line-buffered was not being honoured.
953
954 12. Added an optional parentheses number to the -o and --only-matching options
955 of pcregrep.
956
957 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
958 can match an empty string. The code to do it in pcretest and pcredemo
959 needed fixing:
960
961 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
962 only one byte after an empty string match just before CRLF (this case
963 just got forgotten; "any" and "anycrlf" were OK).
964
965 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
966 mode when an empty string match preceded an ASCII character followed by
967 a non-ASCII character. (The code for advancing by one character rather
968 than one byte was nonsense.)
969
970 (c) The pcredemo.c sample program did not have any code at all to handle
971 the cases when CRLF is a valid newline sequence.
972
973 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
974 as a starting offset was within the subject string. There is now a new
975 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
976 negative or greater than the length of the string. In order to test this,
977 pcretest is extended to allow the setting of negative starting offsets.
978
979 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
980 starting offset points to the beginning of a UTF-8 character was
981 unnecessarily clumsy. I tidied it up.
982
983 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
984 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
985
986 17. Nobody had reported that the --include_dir option, which was added in
987 release 7.7 should have been called --include-dir (hyphen, not underscore)
988 for compatibility with GNU grep. I have changed it to --include-dir, but
989 left --include_dir as an undocumented synonym, and the same for
990 --exclude-dir, though that is not available in GNU grep, at least as of
991 release 2.5.4.
992
993 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
994 characters from a string of bytes) have been redefined so as not to use
995 loops, in order to improve performance in some environments. At the same
996 time, I abstracted some of the common code into auxiliary macros to save
997 repetition (this should not affect the compiled code).
998
999 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1000 compile-time error is now given if \c is not followed by an ASCII
1001 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1002 different, and any byte value is allowed.)
1003
1004 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1005 START_OPTIMIZE option, which is now allowed at compile time - but just
1006 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1007 to pcregrep and other applications that have no direct access to PCRE
1008 options. The new /Y option in pcretest sets this option when calling
1009 pcre_compile().
1010
1011 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1012 back references. Groups containing recursive back references were forced to
1013 be atomic by that change, but in the case of named groups, the amount of
1014 memory required was incorrectly computed, leading to "Failed: internal
1015 error: code overflow". This has been fixed.
1016
1017 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1018 pcretest.c, to avoid build problems in some Borland environments.
1019
1020
1021 Version 8.10 25-Jun-2010
1022 ------------------------
1023
1024 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1025 THEN.
1026
1027 2. (*ACCEPT) was not working when inside an atomic group.
1028
1029 3. Inside a character class, \B is treated as a literal by default, but
1030 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1031 causes the error). The code is unchanged, but I tidied the documentation.
1032
1033 4. Inside a character class, PCRE always treated \R and \X as literals,
1034 whereas Perl faults them if its -w option is set. I have changed PCRE so
1035 that it faults them when PCRE_EXTRA is set.
1036
1037 5. Added support for \N, which always matches any character other than
1038 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1039
1040 6. When compiling pcregrep with newer versions of gcc which may have
1041 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1042 declared with attribute warn_unused_result" were given. Just casting the
1043 result to (void) does not stop the warnings; a more elaborate fudge is
1044 needed. I've used a macro to implement this.
1045
1046 7. Minor change to pcretest.c to avoid a compiler warning.
1047
1048 8. Added four artifical Unicode properties to help with an option to make
1049 \s etc use properties (see next item). The new properties are: Xan
1050 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1051
1052 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1053 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1054 this option. Modified pcretest to add /W to test this facility. Added
1055 REG_UCP to make it available via the POSIX interface.
1056
1057 10. Added --line-buffered to pcregrep.
1058
1059 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1060 studied, and the match started with a letter with a code point greater than
1061 127 whose first byte was different to the first byte of the other case of
1062 the letter, the other case of this starting letter was not recognized
1063 (#976).
1064
1065 12. If a pattern that was studied started with a repeated Unicode property
1066 test, for example, \p{Nd}+, there was the theoretical possibility of
1067 setting up an incorrect bitmap of starting bytes, but fortunately it could
1068 not have actually happened in practice until change 8 above was made (it
1069 added property types that matched character-matching opcodes).
1070
1071 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1072 possible starting bytes for non-anchored patterns.
1073
1074 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1075 \R, and also a number of cases that involve Unicode properties, both
1076 explicit and implicit when PCRE_UCP is set.
1077
1078 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1079 input, it could crash or give wrong results if characters with values
1080 greater than 0xc0 were present in the subject string. (Detail: it assumed
1081 UTF-8 input when processing these items.)
1082
1083 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1084 size_t is 64-bit (#991).
1085
1086 17. Added a check for running out of memory when PCRE is compiled with
1087 --disable-stack-for-recursion (#990).
1088
1089 18. If the last data line in a file for pcretest does not have a newline on
1090 the end, a newline was missing in the output.
1091
1092 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1093 less than 128) in its various bitmaps. However, there is a facility for
1094 generating tables according to the current locale when PCRE is compiled. It
1095 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1096 characters, are recognized by isspace() and therefore were getting set in
1097 these tables, and indeed these tables seem to approximate to ISO 8859. This
1098 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1099 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1100 which of course cannot start UTF-8 characters. I have changed the code so
1101 that only real ASCII characters (less than 128) and the correct starting
1102 bytes for UTF-8 encodings are set for characters greater than 127 when in
1103 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1104 altogether.)
1105
1106 20. Added the /T option to pcretest so as to be able to run tests with non-
1107 standard character tables, thus making it possible to include the tests
1108 used for 19 above in the standard set of tests.
1109
1110 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1111 reference to a subpattern the other side of a comment that contains an
1112 opening parenthesis caused either an internal compiling error, or a
1113 reference to the wrong subpattern.
1114
1115
1116 Version 8.02 19-Mar-2010
1117 ------------------------
1118
1119 1. The Unicode data tables have been updated to Unicode 5.2.0.
1120
1121 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1122 configured.
1123
1124 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1125 original author of that file, following a query about its status.
1126
1127 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1128 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1129
1130 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1131 quantifier applied to a forward-referencing subroutine call, could compile
1132 incorrect code or give the error "internal error: previously-checked
1133 referenced subpattern not found".
1134
1135 6. Both MS Visual Studio and Symbian OS have problems with initializing
1136 variables to point to external functions. For these systems, therefore,
1137 pcre_malloc etc. are now initialized to local functions that call the
1138 relevant global functions.
1139
1140 7. There were two entries missing in the vectors called coptable and poptable
1141 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1142 I've fixed the data, and added a kludgy way of testing at compile time that
1143 the lengths are correct (equal to the number of opcodes).
1144
1145 8. Following on from 7, I added a similar kludge to check the length of the
1146 eint vector in pcreposix.c.
1147
1148 9. Error texts for pcre_compile() are held as one long string to avoid too
1149 much relocation at load time. To find a text, the string is searched,
1150 counting zeros. There was no check for running off the end of the string,
1151 which could happen if a new error number was added without updating the
1152 string.
1153
1154 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1155
1156 11. \K was not working if it appeared in an atomic group or in a group that
1157 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1158 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1159 the assertion is positive, but not if it is negative.
1160
1161 12. Change 11 fortuitously reduced the size of the stack frame used in the
1162 "match()" function of pcre_exec.c by one pointer. Forthcoming
1163 implementation of support for (*MARK) will need an extra pointer on the
1164 stack; I have reserved it now, so that the stack frame size does not
1165 decrease.
1166
1167 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1168 item in branch that calls a recursion is a subroutine call - as in the
1169 second branch in the above example - was incorrectly given the compile-
1170 time error "recursive call could loop indefinitely" because pcre_compile()
1171 was not correctly checking the subroutine for matching a non-empty string.
1172
1173 14. The checks for overrunning compiling workspace could trigger after an
1174 overrun had occurred. This is a "should never occur" error, but it can be
1175 triggered by pathological patterns such as hundreds of nested parentheses.
1176 The checks now trigger 100 bytes before the end of the workspace.
1177
1178 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1179
1180
1181 Version 8.01 19-Jan-2010
1182 ------------------------
1183
1184 1. If a pattern contained a conditional subpattern with only one branch (in
1185 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1186 computed the wrong minimum data length (which is of course zero for such
1187 subpatterns). This could cause incorrect "no match" results.
1188
1189 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1190 the pattern is reset in the first branch, pcre_compile() failed with
1191 "internal error: code overflow at offset...". This happened only when
1192 the reset was to the original external option setting. (An optimization
1193 abstracts leading options settings into an external setting, which was the
1194 cause of this.)
1195
1196 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1197 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1198 assertion pattern did not match (meaning that the assertion was true), it
1199 was incorrectly treated as false if the SKIP had been reached during the
1200 matching. This also applied to assertions used as conditions.
1201
1202 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1203 assertion subpattern, including such a pattern used as a condition,
1204 unpredictable results occurred, instead of the error return
1205 PCRE_ERROR_DFA_UITEM.
1206
1207 5. The C++ GlobalReplace function was not working like Perl for the special
1208 situation when an empty string is matched. It now does the fancy magic
1209 stuff that is necessary.
1210
1211 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1212 removed. (These were left over from very, very early versions of PCRE.)
1213
1214 7. Some cosmetic changes to the code to make life easier when compiling it
1215 as part of something else:
1216
1217 (a) Change DEBUG to PCRE_DEBUG.
1218
1219 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1220 called "current" as "current_branch", to prevent a collision with the
1221 Linux macro when compiled as a kernel module.
1222
1223 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1224 prevent a collision with the Linux macro when compiled as a kernel
1225 module.
1226
1227 8. In pcre_compile() there are some checks for integer overflows that used to
1228 cast potentially large values to (double). This has been changed to that
1229 when building, a check for int64_t is made, and if it is found, it is used
1230 instead, thus avoiding the use of floating point arithmetic. (There is no
1231 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1232 double.
1233
1234 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1235 2005 (difference between two addresses compared to an unsigned value).
1236
1237 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1238 custom one, because of the following reported problem in Windows:
1239
1240 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1241 under Win32.
1242 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1243 therefore missing the function definition.
1244 - The compiler thus generates a "C" signature for the test function.
1245 - The linker fails to find the "C" function.
1246 - PCRE fails to configure if asked to do so against libbz2.
1247
1248 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1249 messages were output:
1250
1251 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1252 rerunning libtoolize, to keep the correct libtool macros in-tree.
1253 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1254
1255 I have done both of these things.
1256
1257 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1258 most of the time, it *can* run out if it is given a pattern that contains a
1259 runaway infinite recursion. I updated the discussion in the pcrestack man
1260 page.
1261
1262 13. Now that we have gone to the x.xx style of version numbers, the minor
1263 version may start with zero. Using 08 or 09 is a bad idea because users
1264 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1265 interpreted as invalid octal numbers. I've updated the previous comment in
1266 configure.ac, and also added a check that gives an error if 08 or 09 are
1267 used.
1268
1269 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1270 causing partial matching to fail when the end of the subject matched \W
1271 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1272
1273 15. There were some discrepancies between the declarations in pcre_internal.h
1274 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1275 their definitions. The declarations used "const uschar *" and the
1276 definitions used USPTR. Even though USPTR is normally defined as "const
1277 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1278 reported that: "This difference in casting confuses some C++ compilers, for
1279 example, SunCC recognizes above declarations as different functions and
1280 generates broken code for hbpcre." I have changed the declarations to use
1281 USPTR.
1282
1283 16. GNU libtool is named differently on some systems. The autogen.sh script now
1284 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1285 (FreeBSD).
1286
1287 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1288 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1289 comment: "Figure out how to create a longlong from a string: strtoll and
1290 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1291 instance, but it only takes 2 args instead of 3!"
1292
1293 18. A subtle bug concerned with back references has been fixed by a change of
1294 specification, with a corresponding code fix. A pattern such as
1295 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1296 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1297 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1298 same bug. Such groups have to be quantified to be useful, or contained
1299 inside another quantified group. (If there's no repetition, the reference
1300 can never match.) The problem arises because, having left the group and
1301 moved on to the rest of the pattern, a later failure that backtracks into
1302 the group uses the captured value from the final iteration of the group
1303 rather than the correct earlier one. I have fixed this in PCRE by forcing
1304 any group that contains a reference to itself to be an atomic group; that
1305 is, there cannot be any backtracking into it once it has completed. This is
1306 similar to recursive and subroutine calls.
1307
1308
1309 Version 8.00 19-Oct-09
1310 ----------------------
1311
1312 1. The table for translating pcre_compile() error codes into POSIX error codes
1313 was out-of-date, and there was no check on the pcre_compile() error code
1314 being within the table. This could lead to an OK return being given in
1315 error.
1316
1317 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1318 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1319 in a Windows environment.
1320
1321 3. The pcregrep --count option prints the count for each file even when it is
1322 zero, as does GNU grep. However, pcregrep was also printing all files when
1323 --files-with-matches was added. Now, when both options are given, it prints
1324 counts only for those files that have at least one match. (GNU grep just
1325 prints the file name in this circumstance, but including the count seems
1326 more useful - otherwise, why use --count?) Also ensured that the
1327 combination -clh just lists non-zero counts, with no names.
1328
1329 4. The long form of the pcregrep -F option was incorrectly implemented as
1330 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1331 but it seems right to fix it, and I didn't think it was worth preserving
1332 the old behaviour.
1333
1334 5. The command line items --regex=pattern and --regexp=pattern were not
1335 recognized by pcregrep, which required --regex pattern or --regexp pattern
1336 (with a space rather than an '='). The man page documented the '=' forms,
1337 which are compatible with GNU grep; these now work.
1338
1339 6. No libpcreposix.pc file was created for pkg-config; there was just
1340 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1341
1342 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1343 when UCP support is not needed, by modifying the Python script that
1344 generates it from Unicode data files. This should not matter if the module
1345 is correctly used as a library, but I received one complaint about 50K of
1346 unwanted data. My guess is that the person linked everything into his
1347 program rather than using a library. Anyway, it does no harm.
1348
1349 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1350 was a minimum greater than 1 for a wide character in a possessive
1351 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1352 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1353 character. Chaos in the form of incorrect output or a compiling loop could
1354 result.
1355
1356 9. The restrictions on what a pattern can contain when partial matching is
1357 requested for pcre_exec() have been removed. All patterns can now be
1358 partially matched by this function. In addition, if there are at least two
1359 slots in the offset vector, the offset of the earliest inspected character
1360 for the match and the offset of the end of the subject are set in them when
1361 PCRE_ERROR_PARTIAL is returned.
1362
1363 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1364 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1365 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1366 and may be more useful for multi-segment matching.
1367
1368 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1369 used to be given if ever the end of the subject was reached; now it is
1370 given only if matching could not proceed because another character was
1371 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1372 string "Z", which now yields "no match" instead of "partial match". In the
1373 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1374 final character ended with (*FAIL).
1375
1376 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1377 if the pattern had a "must contain" character that was already found in the
1378 earlier partial match, unless partial matching was again requested. For
1379 example, with the pattern /dog.(body)?/, the "must contain" character is
1380 "g". If the first part-match was for the string "dog", restarting with
1381 "sbody" failed. This bug has been fixed.
1382
1383 13. The string returned by pcre_dfa_exec() after a partial match has been
1384 changed so that it starts at the first inspected character rather than the
1385 first character of the match. This makes a difference only if the pattern
1386 starts with a lookbehind assertion or \b or \B (\K is not supported by
1387 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1388 matching functions compatible, and I think it's the right thing to do.
1389
1390 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1391 so that the demonstration program is easily available in environments where
1392 PCRE has not been installed from source.
1393
1394 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1395 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1396 library.
1397
1398 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1399 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1400 is not the first non-POSIX option to be added. Clearly some people find
1401 these options useful.
1402
1403 17. If a caller to the POSIX matching function regexec() passes a non-zero
1404 value for nmatch with a NULL value for pmatch, the value of
1405 nmatch is forced to zero.
1406
1407 18. RunGrepTest did not have a test for the availability of the -u option of
1408 the diff command, as RunTest does. It now checks in the same way as
1409 RunTest, and also checks for the -b option.
1410
1411 19. If an odd number of negated classes containing just a single character
1412 interposed, within parentheses, between a forward reference to a named
1413 subpattern and the definition of the subpattern, compilation crashed with
1414 an internal error, complaining that it could not find the referenced
1415 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1416 [The bug was that it was starting one character too far in when skipping
1417 over the character class, thus treating the ] as data rather than
1418 terminating the class. This meant it could skip too much.]
1419
1420 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1421 /g option in pcretest when the pattern contains \K, which makes it possible
1422 to have an empty string match not at the start, even when the pattern is
1423 anchored. Updated pcretest and pcredemo to use this option.
1424
1425 21. If the maximum number of capturing subpatterns in a recursion was greater
1426 than the maximum at the outer level, the higher number was returned, but
1427 with unset values at the outer level. The correct (outer level) value is
1428 now given.
1429
1430 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1431 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1432 make it do so. The string so far is captured, making this feature
1433 compatible with Perl.
1434
1435 23. The tests have been re-organized, adding tests 11 and 12, to make it
1436 possible to check the Perl 5.10 features against Perl 5.10.
1437
1438 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1439 pattern matches a fixed length string. PCRE did not allow this; now it
1440 does. Neither allows recursion.
1441
1442 25. I finally figured out how to implement a request to provide the minimum
1443 length of subject string that was needed in order to match a given pattern.
1444 (It was back references and recursion that I had previously got hung up
1445 on.) This code has now been added to pcre_study(); it finds a lower bound
1446 to the length of subject needed. It is not necessarily the greatest lower
1447 bound, but using it to avoid searching strings that are too short does give
1448 some useful speed-ups. The value is available to calling programs via
1449 pcre_fullinfo().
1450
1451 26. While implementing 25, I discovered to my embarrassment that pcretest had
1452 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1453 study optimizations had never been tested with that matching function.
1454 Oops. What is worse, even when it was passed study data, there was a bug in
1455 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1456 were also very few tests of studied patterns with pcre_dfa_exec().
1457
1458 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1459 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1460 on the other side of the coin, they are no longer allowed to have different
1461 names, because these cannot be distinguished in PCRE, and this has caused
1462 confusion. (This is a difference from Perl.)
1463
1464 28. When duplicate subpattern names are present (necessarily with different
1465 numbers, as required by 27 above), and a test is made by name in a
1466 conditional pattern, either for a subpattern having been matched, or for
1467 recursion in such a pattern, all the associated numbered subpatterns are
1468 tested, and the overall condition is true if the condition is true for any
1469 one of them. This is the way Perl works, and is also more like the way
1470 testing by number works.
1471
1472
1473 Version 7.9 11-Apr-09
1474 ---------------------
1475
1476 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1477 (pcretest), all targets were linked against these libraries. This included
1478 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1479 libraries. This caused unwanted dependencies to be created. This problem
1480 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1481 pcretest is linked with readline.
1482
1483 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1484 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1485 moved outside it again, because FALSE and TRUE are already defined in AIX,
1486 but BOOL is not.
1487
1488 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1489 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1490
1491 4. The pcregrep documentation said spaces were inserted as well as colons (or
1492 hyphens) following file names and line numbers when outputting matching
1493 lines. This is not true; no spaces are inserted. I have also clarified the
1494 wording for the --colour (or --color) option.
1495
1496 5. In pcregrep, when --colour was used with -o, the list of matching strings
1497 was not coloured; this is different to GNU grep, so I have changed it to be
1498 the same.
1499
1500 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1501 each matching line was coloured. Now it goes on to look for further matches
1502 of any of the test patterns, which is the same behaviour as GNU grep.
1503
1504 7. A pattern that could match an empty string could cause pcregrep to loop; it
1505 doesn't make sense to accept an empty string match in pcregrep, so I have
1506 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1507 seems to be how GNU grep behaves.
1508
1509 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1510 start or after a newline", because the conditional assertion was not being
1511 correctly handled. The rule now is that both the assertion and what follows
1512 in the first alternative must satisfy the test.
1513
1514 9. If auto-callout was enabled in a pattern with a conditional group whose
1515 condition was an assertion, PCRE could crash during matching, both with
1516 pcre_exec() and pcre_dfa_exec().
1517
1518 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1519 used for matching.
1520
1521 11. Unicode property support in character classes was not working for
1522 characters (bytes) greater than 127 when not in UTF-8 mode.
1523
1524 12. Added the -M command line option to pcretest.
1525
1526 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1527
1528 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1529
1530 16. Added comments and documentation about mis-use of no_arg in the C++
1531 wrapper.
1532
1533 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1534 from Martin Jerabek that uses macro names for all relevant character and
1535 string constants.
1536
1537 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1538 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1539 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1540 these, but not everybody uses configure.
1541
1542 19. A conditional group that had only one branch was not being correctly
1543 recognized as an item that could match an empty string. This meant that an
1544 enclosing group might also not be so recognized, causing infinite looping
1545 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1546 with the subject "ab", where knowledge that the repeated group can match
1547 nothing is needed in order to break the loop.
1548
1549 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1550 exec(), but without supplying a callout function, matching went wrong.
1551
1552 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1553 leak if the size of the offset vector was greater than 30. When the vector
1554 is smaller, the saved offsets during recursion go onto a local stack
1555 vector, but for larger vectors malloc() is used. It was failing to free
1556 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1557 error, in fact).
1558
1559 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1560 heapframe that is used only when UTF-8 support is enabled. This caused no
1561 problem, but was untidy.
1562
1563 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1564 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1565 included within another project.
1566
1567 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1568 slightly modified by me:
1569
1570 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1571 not building pcregrep.
1572
1573 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1574 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1575
1576 25. Forward references, both numeric and by name, in patterns that made use of
1577 duplicate group numbers, could behave incorrectly or give incorrect errors,
1578 because when scanning forward to find the reference group, PCRE was not
1579 taking into account the duplicate group numbers. A pattern such as
1580 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1581
1582 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1583 the feature of a custom pointer more persuasive (as requested by a user).
1584
1585 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1586 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1587 pre-defined.
1588
1589 28. Added support for (*UTF8) at the start of a pattern.
1590
1591 29. Arrange for flags added by the "release type" setting in CMake to be shown
1592 in the configuration summary.
1593
1594
1595 Version 7.8 05-Sep-08
1596 ---------------------
1597
1598 1. Replaced UCP searching code with optimized version as implemented for Ad
1599 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1600 stage table and inline lookup instead of a function, giving speed ups of 2
1601 to 5 times on some simple patterns that I tested. Permission was given to
1602 distribute the MultiStage2.py script that generates the tables (it's not in
1603 the tarball, but is in the Subversion repository).
1604
1605 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1606 scripts.
1607
1608 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1609 a group with a zero qualifier. The result of the study could be incorrect,
1610 or the function might crash, depending on the pattern.
1611
1612 4. Caseless matching was not working for non-ASCII characters in back
1613 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1614 It now works when Unicode Property Support is available.
1615
1616 5. In pcretest, an escape such as \x{de} in the data was always generating
1617 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1618 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1619 truncation.
1620
1621 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1622
1623 7. Added two (int) casts to pcregrep when printing the difference of two
1624 pointers, in case they are 64-bit values.
1625
1626 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1627 test 2 if it fails.
1628
1629 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1630 and a #define of that name to empty if it is not externally set. This is to
1631 allow users of MSVC to set it if necessary.
1632
1633 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1634 the convenience functions in the pcre_get.c source file.
1635
1636 11. An option change at the start of a pattern that had top-level alternatives
1637 could cause overwriting and/or a crash. This command provoked a crash in
1638 some environments:
1639
1640 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1641
1642 This potential security problem was recorded as CVE-2008-2371.
1643
1644 12. For a pattern where the match had to start at the beginning or immediately
1645 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1646 pcre_dfa_exec() could read past the end of the passed subject if there was
1647 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1648 pcretest so that it places the subject at the end of its malloc-ed buffer.
1649
1650 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1651 exec() might read past the end of the data buffer in UTF-8 mode.
1652
1653 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1654 the data contained the byte 0x85 as part of a UTF-8 character within its
1655 first line. This applied both to normal and DFA matching.
1656
1657 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1658 /^[^d]*?$/8 failed to match "abc".
1659
1660 16. Added a missing copyright notice to pcrecpp_internal.h.
1661
1662 17. Make it more clear in the documentation that values returned from
1663 pcre_exec() in ovector are byte offsets, not character counts.
1664
1665 18. Tidied a few places to stop certain compilers from issuing warnings.
1666
1667 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1668 supplied by Stefan Weber. I made a further small update for 7.8 because
1669 there is a change of source arrangements: the pcre_searchfuncs.c module is
1670 replaced by pcre_ucd.c.
1671
1672
1673 Version 7.7 07-May-08
1674 ---------------------
1675
1676 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1677 a string to a long long, pretend we don't even have a long long." This is
1678 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1679
1680 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1681 pre-7.6 versions, which defined a global no_arg variable instead of putting
1682 it in the RE class. (See also #8 below.)
1683
1684 3. Remove a line of dead code, identified by coverity and reported by Nuno
1685 Lopes.
1686
1687 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1688
1689 (1) The include/exclude patterns were being applied to the whole pathnames
1690 of files, instead of just to the final components.
1691
1692 (2) If there was more than one level of directory, the subdirectories were
1693 skipped unless they satisfied the include/exclude conditions. This is
1694 inconsistent with GNU grep (and could even be seen as contrary to the
1695 pcregrep specification - which I improved to make it absolutely clear).
1696 The action now is always to scan all levels of directory, and just
1697 apply the include/exclude patterns to regular files.
1698
1699 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1700 --exclude_dir in the tests to avoid scanning .svn directories.
1701
1702 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1703 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1704 doesn't support NULs in patterns.
1705
1706 7. Added some missing "const"s to declarations of static tables in
1707 pcre_compile.c and pcre_dfa_exec.c.
1708
1709 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1710 caused by fix #2 above. (Subsequently also a second patch to fix the
1711 first patch. And a third patch - this was a messy problem.)
1712
1713 9. Applied Craig's patch to remove the use of push_back().
1714
1715 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1716 matching function regexec().
1717
1718 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1719 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1720 references. PCRE supports relative numbers with this syntax (I don't think
1721 Oniguruma does).
1722
1723 12. Previously, a group with a zero repeat such as (...){0} was completely
1724 omitted from the compiled regex. However, this means that if the group
1725 was called as a subroutine from elsewhere in the pattern, things went wrong
1726 (an internal error was given). Such groups are now left in the compiled
1727 pattern, with a new opcode that causes them to be skipped at execution
1728 time.
1729
1730 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1731 to the way PCRE behaves:
1732
1733 (a) A lone ] character is dis-allowed (Perl treats it as data).
1734
1735 (b) A back reference to an unmatched subpattern matches an empty string
1736 (Perl fails the current match path).
1737
1738 (c) A data ] in a character class must be notated as \] because if the
1739 first data character in a class is ], it defines an empty class. (In
1740 Perl it is not possible to have an empty class.) The empty class []
1741 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1742 The negative empty class [^] matches any one character, independently
1743 of the DOTALL setting.
1744
1745 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1746 non-existent subpattern following a character class starting with ']' and
1747 containing () gave an internal compiling error instead of "reference to
1748 non-existent subpattern". Fortunately, when the pattern did exist, the
1749 compiled code was correct. (When scanning forwards to check for the
1750 existencd of the subpattern, it was treating the data ']' as terminating
1751 the class, so got the count wrong. When actually compiling, the reference
1752 was subsequently set up correctly.)
1753
1754 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1755 it was being rejected as not supported by pcre_dfa_exec(), even though
1756 other assertions are supported. I have made pcre_dfa_exec() support
1757 (*FAIL).
1758
1759 16. The implementation of 13c above involved the invention of a new opcode,
1760 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1761 cannot be changed at match time, I realized I could make a small
1762 improvement to matching performance by compiling OP_ALLANY instead of
1763 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1764 on the OP_ANY path.
1765
1766 17. Compiling pcretest on Windows with readline support failed without the
1767 following two fixes: (1) Make the unistd.h include conditional on
1768 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1769
1770 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1771 ncurses library to be included for pcretest when ReadLine support is
1772 requested, but also to allow for it to be overridden. This patch came from
1773 Daniel Bergström.
1774
1775 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1776 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1777 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1778 spotting this.
1779
1780
1781 Version 7.6 28-Jan-08
1782 ---------------------
1783
1784 1. A character class containing a very large number of characters with
1785 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1786 overflow.
1787
1788 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1789 HAVE_LONG_LONG is not defined.
1790
1791 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1792 bring them up to date and include new features. This patch includes:
1793
1794 - Fixed PH's badly added libz and libbz2 support.
1795 - Fixed a problem with static linking.
1796 - Added pcredemo. [But later removed - see 7 below.]
1797 - Fixed dftables problem and added an option.
1798 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1799 HAVE_LONG_LONG.
1800 - Added readline support for pcretest.
1801 - Added an listing of the option settings after cmake has run.
1802
1803 4. A user submitted a patch to Makefile that makes it easy to create
1804 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1805 Makefile.am that cause it to include this special target, without
1806 affecting anything else. Note that the same mingw target plus all
1807 the other distribution libraries and programs are now supported
1808 when configuring with CMake (see 6 below) instead of with
1809 Configure/Make.
1810
1811 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1812 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1813 exported in the Windows port". It has not yet been confirmed that the patch
1814 solves the problem, but it does no harm.
1815
1816 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1817 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1818 with CMake, and also correct the comment about stack recursion.
1819
1820 7. Remove the automatic building of pcredemo from the ./configure system and
1821 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1822 of a program that users should build themselves after PCRE is installed, so
1823 building it automatically is not really right. What is more, it gave
1824 trouble in some build environments.
1825
1826 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1827
1828
1829 Version 7.5 10-Jan-08
1830 ---------------------
1831
1832 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1833 values in parens when parsing an RE using the C++ wrapper."
1834
1835 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1836 Characters greater than 255 were excluded from the class instead of being
1837 included.
1838
1839 3. The same bug as (2) above applied to negated POSIX classes such as
1840 [:^space:].
1841
1842 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1843 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1844 I have changed it.
1845
1846 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1847 first named subpattern) and a construct such as (?&a) would reference the
1848 first named subpattern whose name started with "a" (in other words, the
1849 length check was missing). Both these problems are fixed. "Subpattern name
1850 expected" is now given for (?&) (a zero-length name), and this patch also
1851 makes it give the same error for \k'' (previously it complained that that
1852 was a reference to a non-existent subpattern).
1853
1854 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1855 this is right because (?- can be followed by option settings as well as by
1856 digits. I have, however, made the messages clearer.
1857
1858 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1859 than the number used in the conditional) now cause a compile-time error.
1860 This is actually not compatible with Perl, which accepts such patterns, but
1861 treats the conditional as always being FALSE (as PCRE used to), but it
1862 seems to me that giving a diagnostic is better.
1863
1864 8. Change "alphameric" to the more common word "alphanumeric" in comments
1865 and messages.
1866
1867 9. Fix two occurrences of "backslash" in comments that should have been
1868 "backspace".
1869
1870 10. Remove two redundant lines of code that can never be obeyed (their function
1871 was moved elsewhere).
1872
1873 11. The program that makes PCRE's Unicode character property table had a bug
1874 which caused it to generate incorrect table entries for sequences of
1875 characters that have the same character type, but are in different scripts.
1876 It amalgamated them into a single range, with the script of the first of
1877 them. In other words, some characters were in the wrong script. There were
1878 thirteen such cases, affecting characters in the following ranges:
1879
1880 U+002b0 - U+002c1
1881 U+0060c - U+0060d
1882 U+0061e - U+00612
1883 U+0064b - U+0065e
1884 U+0074d - U+0076d
1885 U+01800 - U+01805
1886 U+01d00 - U+01d77
1887 U+01d9b - U+01dbf
1888 U+0200b - U+0200f
1889 U+030fc - U+030fe
1890 U+03260 - U+0327f
1891 U+0fb46 - U+0fbb1
1892 U+10450 - U+1049d
1893
1894 12. The -o option (show only the matching part of a line) for pcregrep was not
1895 compatible with GNU grep in that, if there was more than one match in a
1896 line, it showed only the first of them. It now behaves in the same way as
1897 GNU grep.
1898
1899 13. If the -o and -v options were combined for pcregrep, it printed a blank
1900 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1901 does the same. The return code can be used to tell if there were any
1902 non-matching lines.
1903
1904 14. Added --file-offsets and --line-offsets to pcregrep.
1905
1906 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1907 infinitely looping recursion. The bug was that positive lookaheads were not
1908 being skipped when checking for a possible empty match (negative lookaheads
1909 and both kinds of lookbehind were skipped).
1910
1911 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1912 inclusion of <windows.h> to before rather than after the definition of
1913 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1914
1915 17. Specifying a possessive quantifier with a specific limit for a Unicode
1916 character property caused pcre_compile() to compile bad code, which led at
1917 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1918 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1919 caused the error; without that there was no problem.
1920
1921 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1922
1923 19. Added --enable-pcretest-libreadline.
1924
1925 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1926 RE::GlobalReplace(). As a result, the number of replacements returned was
1927 double what it should be. I removed one of the increments, but Craig sent a
1928 later patch that removed the other one (the right fix) and added unit tests
1929 that check the return values (which was not done before).
1930
1931 21. Several CMake things:
1932
1933 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1934 the names libpcre and libpcreposix, not just pcre and pcreposix.
1935
1936 (2) The above change means that pcretest and pcregrep are now correctly
1937 linked with the newly-built libraries, not previously installed ones.
1938
1939 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1940
1941 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1942 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1943 UTF-8 newline character). The key issue is that the pattern starts .*;
1944 this means that the match must be either at the beginning, or after a
1945 newline. The bug was in the code for advancing after a failed match and
1946 checking that the new position followed a newline. It was not taking
1947 account of UTF-8 characters correctly.
1948
1949 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1950 character classes. PCRE was not treating the sequence [:...:] as a
1951 character class unless the ... were all letters. Perl, however, seems to
1952 allow any characters between [: and :], though of course it rejects as
1953 unknown any "names" that contain non-letters, because all the known class
1954 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1955 for example, whereas PCRE did not - it did not recognize a POSIX character
1956 class. This seemed a bit dangerous, so the code has been changed to be
1957 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1958 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1959 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1960 Perl does, and where it didn't before.
1961
1962 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1963 Windows environments %n is disabled by default.
1964
1965
1966 Version 7.4 21-Sep-07
1967 ---------------------
1968
1969 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1970 means that a class such as [\s] counted as "explicit reference to CR or
1971 LF". That isn't really right - the whole point of the change was to try to
1972 help when there was an actual mention of one of the two characters. So now
1973 the change happens only if \r or \n (or a literal CR or LF) character is
1974 encountered.
1975
1976 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1977 of both had grown to the point where there were only 3 bits left.
1978 Fortunately, there was spare space in the data structure, and so I have
1979 moved the internal flags into a new 16-bit field to free up more option
1980 bits.
1981
1982 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1983 but did not set the internal JCHANGED flag - either of these is enough to
1984 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1985 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1986 start sets both bits.
1987
1988 4. Added options (at build time, compile time, exec time) to change \R from
1989 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1990
1991 5. doc/pcresyntax.html was missing from the distribution.
1992
1993 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1994 compatibility, even though it is no longer used.
1995
1996 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1997 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1998 windows.h file is present (where different names are used). [This was
1999 reversed later after testing - see 16 below.]
2000
2001 8. Changed all #include <config.h> to #include "config.h". There were also
2002 some further <pcre.h> cases that I changed to "pcre.h".
2003
2004 9. When pcregrep was used with the --colour option, it missed the line ending
2005 sequence off the lines that it output.
2006
2007 10. It was pointed out to me that arrays of string pointers cause lots of
2008 relocations when a shared library is dynamically loaded. A technique of
2009 using a single long string with a table of offsets can drastically reduce
2010 these. I have refactored PCRE in four places to do this. The result is
2011 dramatic:
2012
2013 Originally: 290
2014 After changing UCP table: 187
2015 After changing error message table: 43
2016 After changing table of "verbs" 36
2017 After changing table of Posix names 22
2018
2019 Thanks to the folks working on Gregex for glib for this insight.
2020
2021 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2022 unicode-properties was also set.
2023
2024 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2025
2026 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2027 checked only for CRLF.
2028
2029 14. Added casts to pcretest.c to avoid compiler warnings.
2030
2031 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2032
2033 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2034 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2035 entirely. This removes changes made in 7 above.
2036
2037 17. The CMake files have been updated, and there is now more information about
2038 building with CMake in the NON-UNIX-USE document.
2039
2040
2041 Version 7.3 28-Aug-07
2042 ---------------------
2043
2044 1. In the rejigging of the build system that eventually resulted in 7.1, the
2045 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2046 brackets there is not right, since it causes compilers to look for an
2047 installed pcre.h, not the version that is in the source that is being
2048 compiled (which of course may be different). I have changed it back to:
2049
2050 #include "pcre.h"
2051
2052 I have a vague recollection that the change was concerned with compiling in
2053 different directories, but in the new build system, that is taken care of
2054 by the VPATH setting the Makefile.
2055
2056 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2057 when the subject happened to end in the byte 0x85 (e.g. if the last
2058 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2059 characters but of course it shouldn't be taken as a newline when it is part
2060 of another character. The bug was that, for an unlimited repeat of . in
2061 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2062 characters when looking for a newline.
2063
2064 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2065
2066 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2067 in debug output.
2068
2069 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2070 long printing in the pcrecpp unittest when running under MinGW.
2071
2072 6. ESC_K was left out of the EBCDIC table.
2073
2074 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2075 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2076 limit also applies to "virtual nesting" when a pattern is recursive, and in
2077 this case 1000 isn't so big. I have been able to remove this limit at the
2078 expense of backing off one optimization in certain circumstances. Normally,
2079 when pcre_exec() would call its internal match() function recursively and
2080 immediately return the result unconditionally, it uses a "tail recursion"
2081 feature to save stack. However, when a subpattern that can match an empty
2082 string has an unlimited repetition quantifier, it no longer makes this
2083 optimization. That gives it a stack frame in which to save the data for
2084 checking that an empty string has been matched. Previously this was taken
2085 from the 1000-entry workspace that had been reserved. So now there is no
2086 explicit limit, but more stack is used.
2087
2088 8. Applied Daniel's patches to solve problems with the import/export magic
2089 syntax that is required for Windows, and which was going wrong for the
2090 pcreposix and pcrecpp parts of the library. These were overlooked when this
2091 problem was solved for the main library.
2092
2093 9. There were some crude static tests to avoid integer overflow when computing
2094 the size of patterns that contain repeated groups with explicit upper
2095 limits. As the maximum quantifier is 65535, the maximum group length was
2096 set at 30,000 so that the product of these two numbers did not overflow a
2097 32-bit integer. However, it turns out that people want to use groups that
2098 are longer than 30,000 bytes (though not repeat them that many times).
2099 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2100 made it possible to implement the integer overflow checks in a much more
2101 dynamic way, which I have now done. The artificial limitation on group
2102 length has been removed - we now have only the limit on the total length of
2103 the compiled pattern, which depends on the LINK_SIZE setting.
2104
2105 10. Fixed a bug in the documentation for get/copy named substring when
2106 duplicate names are permitted. If none of the named substrings are set, the
2107 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2108 empty string.
2109
2110 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2111 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2112 because the ] is interpreted as the first data character and the
2113 terminating ] is not found. PCRE has been made compatible with Perl in this
2114 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2115 cause memory overwriting.
2116
2117 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2118 string has been matched (to stop an infinite loop). It was not recognizing
2119 a conditional subpattern that could match an empty string if that
2120 subpattern was within another subpattern. For example, it looped when
2121 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2122 condition was not nested. This bug has been fixed.
2123
2124 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2125 past the start of the subject in the presence of bytes with the top bit
2126 set, for example "\x8aBCD".
2127
2128 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2129 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2130
2131 14. Optimized (?!) to (*FAIL).
2132
2133 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2134 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2135 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2136 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2137 does: it's just the validity check that is more restrictive.
2138
2139 16. Inserted checks for integer overflows during escape sequence (backslash)
2140 processing, and also fixed erroneous offset values for syntax errors during
2141 backslash processing.
2142
2143 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2144 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2145
2146 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2147 caused an overrun.
2148
2149 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2150 something other than just ASCII characters) inside a group that had an
2151 unlimited repeat caused a loop at compile time (while checking to see
2152 whether the group could match an empty string).
2153
2154 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2155 [\P{Any}] did so. (Error in the code for printing property names.)
2156
2157 21. An orphan \E inside a character class could cause a crash.
2158
2159 22. A repeated capturing bracket such as (A)? could cause a wild memory
2160 reference during compilation.
2161
2162 23. There are several functions in pcre_compile() that scan along a compiled
2163 expression for various reasons (e.g. to see if it's fixed length for look
2164 behind). There were bugs in these functions when a repeated \p or \P was
2165 present in the pattern. These operators have additional parameters compared
2166 with \d, etc, and these were not being taken into account when moving along
2167 the compiled data. Specifically:
2168
2169 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2170 length.
2171
2172 (b) An item such as \pL+ within a repeated group could cause crashes or
2173 loops.
2174
2175 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2176 "reference to non-existent subpattern" error.
2177
2178 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2179
2180 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2181 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2182
2183 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2184
2185 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2186 character were causing crashes (broken optimization).
2187
2188 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2189 \p or \P) caused a compile-time loop.
2190
2191 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2192 break. For example, the unstudied pattern [\r\n]A does not match the string
2193 "\r\nA" because change 7.0/46 below moves the current point on by two
2194 characters after failing to match at the start. However, the pattern \nA
2195 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2196 the same is true. There doesn't seem any very clean way out of this, but
2197 what I have chosen to do makes the common cases work: PCRE now takes note
2198 of whether there can be an explicit match for \r or \n anywhere in the
2199 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2200 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2201 pattern has explicit CR or LF references.
2202
2203 29. Added (*CR) etc for changing newline setting at start of pattern.
2204
2205
2206 Version 7.2 19-Jun-07
2207 ---------------------
2208
2209 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2210 which is apparently normally available under Windows.
2211
2212 2. Re-jig the pcregrep tests with different newline settings in an attempt
2213 to make them independent of the local environment's newline setting.
2214
2215 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2216
2217 4. Some of the "internals" tests were previously cut out when the link size
2218 was not 2, because the output contained actual offsets. The recent new
2219 "Z" feature of pcretest means that these can be cut out, making the tests
2220 usable with all link sizes.
2221
2222 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2223 stack recursion. This gives a massive performance boost under BSD, but just
2224 a small improvement under Linux. However, it saves one field in the frame
2225 in all cases.
2226
2227 6. Added more features from the forthcoming Perl 5.10:
2228
2229 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2230 recursion call. It refers to the nth most recently opened parentheses.
2231
2232 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2233 to be opened parentheses.
2234
2235 (c) Conditions that refer to capturing parentheses can be specified
2236 relatively, for example, (?(-2)... or (?(+3)...
2237
2238 (d) \K resets the start of the current match so that everything before
2239 is not part of it.
2240
2241 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2242
2243 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2244 reference syntax.
2245
2246 (g) (?| introduces a group in which the numbering of parentheses in each
2247 alternative starts with the same number.
2248
2249 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2250
2251 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2252 PCRE_INFO_JCHANGED.
2253
2254 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2255 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2256 for detecting groups that can match an empty string.
2257
2258 9. A pattern with a very large number of alternatives (more than several
2259 hundred) was running out of internal workspace during the pre-compile
2260 phase, where pcre_compile() figures out how much memory will be needed. A
2261 bit of new cunning has reduced the workspace needed for groups with
2262 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2263 workspace instead of running out of the 4096 that are available.
2264
2265 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2266
2267 11. Applied patch from Google to remove an optimization that didn't quite work.
2268 The report of the bug said:
2269
2270 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2271 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2272 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2273
2274 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2275 it matched the wrong number of bytes.
2276
2277
2278 Version 7.1 24-Apr-07
2279 ---------------------
2280
2281 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2282 that is more "standard", making use of automake and other Autotools. There
2283 is some re-arrangement of the files and adjustment of comments consequent
2284 on this.
2285
2286 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2287 for recursive directory scanning broke on some systems because the files
2288 are not scanned in any specific order and on different systems the order
2289 was different. A call to "sort" has been inserted into RunGrepTest for the
2290 approprate test as a short-term fix. In the longer term there may be an
2291 alternative.
2292
2293 3. I had an email from Eric Raymond about problems translating some of PCRE's
2294 man pages to HTML (despite the fact that I distribute HTML pages, some
2295 people do their own conversions for various reasons). The problems
2296 concerned the use of low-level troff macros .br and .in. I have therefore
2297 removed all such uses from the man pages (some were redundant, some could
2298 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2299 HTML has been updated to handle .nf/.fi and to complain if it encounters
2300 .br or .in.
2301
2302 4. Updated comments in configure.ac that get placed in config.h.in and also
2303 arranged for config.h to be included in the distribution, with the name
2304 config.h.generic, for the benefit of those who have to compile without
2305 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2306
2307 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2308 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2309 makevp.bat for the current PCRE, using the additional files
2310 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2311
2312 6. A Windows user reported a minor discrepancy with test 2, which turned out
2313 to be caused by a trailing space on an input line that had got lost in his
2314 copy. The trailing space was an accident, so I've just removed it.
2315
2316 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2317 that is needed.
2318
2319 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2320 as "const" (a) because they are and (b) because it helps the PHP
2321 maintainers who have recently made a script to detect big data structures
2322 in the php code that should be moved to the .rodata section. I remembered
2323 to update Builducptable as well, so it won't revert if ucptable.h is ever
2324 re-created.
2325
2326 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2327 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2328 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2329 support is not required. This saves 1.5-2K of code, which is important in
2330 some applications.
2331
2332 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2333 so as not to refer to the tables, even though these functions will never be
2334 called when UTF-8 support is disabled. Otherwise there are problems with a
2335 shared library.
2336
2337 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2338
2339 (a) It was defining its arguments as char * instead of void *.
2340
2341 (b) It was assuming that all moves were upwards in memory; this was true
2342 a long time ago when I wrote it, but is no longer the case.
2343
2344 The emulated memove() is provided for those environments that have neither
2345 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2346 is clearly not the case, as these two bugs were recently reported.
2347
2348 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2349 and Detrail to create the HTML documentation, the .txt form of the man
2350 pages, and it removes trailing spaces from listed files. It also creates
2351 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2352 case, it wraps all the #defines with #ifndefs. This script should be run
2353 before "make dist".
2354
2355 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2356 with Unicode property support.
2357
2358 (a) For a maximizing quantifier, if the two different cases of the
2359 character were of different lengths in their UTF-8 codings (there are
2360 some cases like this - I found 11), and the matching function had to
2361 back up over a mixture of the two cases, it incorrectly assumed they
2362 were both the same length.
2363
2364 (b) When PCRE was configured to use the heap rather than the stack for
2365 recursion during matching, it was not correctly preserving the data for
2366 the other case of a UTF-8 character when checking ahead for a match
2367 while processing a minimizing repeat. If the check also involved
2368 matching a wide character, but failed, corruption could cause an
2369 erroneous result when trying to check for a repeat of the original
2370 character.
2371
2372 13. Some tidying changes to the testing mechanism:
2373
2374 (a) The RunTest script now detects the internal link size and whether there
2375 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2376 values substituted by "configure". (The RunGrepTest script already did
2377 this for UTF-8.) The configure.ac script no longer substitutes the
2378 relevant variables.
2379
2380 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2381 with length and offset values. This means that the output is different
2382 for different internal link sizes. Test 2 is skipped for link sizes
2383 other than 2 because of this, bypassing the problem. Unfortunately,
2384 there was also a test in test 3 (the locale tests) that used /B and
2385 failed for link sizes other than 2. Rather than cut the whole test out,
2386 I have added a new /Z option to pcretest that replaces the length and
2387 offset values with spaces. This is now used to make test 3 independent
2388 of link size. (Test 2 will be tidied up later.)
2389
2390 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2391 segmentation fault instead of returning the appropriate error message.
2392
2393 15. In multiline mode when the newline sequence was set to "any", the pattern
2394 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2395 This doesn't seem right; it now treats the CRLF combination as the line
2396 ending, and so does not match in that case. It's only a pattern such as ^$
2397 that would hit this one: something like ^ABC$ would have failed after \r
2398 and then tried again after \r\n.
2399
2400 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2401 in an attempt to make files that differ only in their line terminators
2402 compare equal. This works on Linux.
2403
2404 17. Under certain error circumstances pcregrep might try to free random memory
2405 as it exited. This is now fixed, thanks to valgrind.
2406
2407 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2408 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2409 was because its rules for how to advance for /g after matching an empty
2410 string at the end of a line did not allow for this case. They now check for
2411 it specially.
2412
2413 20. pcretest is supposed to handle patterns and data of any length, by
2414 extending its buffers when necessary. It was getting this wrong when the
2415 buffer for a data line had to be extended.
2416
2417 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2418 CRLF as a newline sequence.
2419
2420 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2421 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2422 I have nevertheless tidied it up.
2423
2424 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2425
2426 24. Added a man page for pcre-config.
2427
2428
2429 Version 7.0 19-Dec-06
2430 ---------------------
2431
2432 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2433 moving to gcc 4.1.1.
2434
2435 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2436 sys/time.h, which is documented as needed for this function. It doesn't
2437 seem to matter on Linux, but it showed up on some releases of OS X.
2438
2439 3. It seems that there are systems where bytes whose values are greater than
2440 127 match isprint() in the "C" locale. The "C" locale should be the
2441 default when a C program starts up. In most systems, only ASCII printing
2442 characters match isprint(). This difference caused the output from pcretest
2443 to vary, making some of the tests fail. I have changed pcretest so that:
2444
2445 (a) When it is outputting text in the compiled version of a pattern, bytes
2446 other than 32-126 are always shown as hex escapes.
2447
2448 (b) When it is outputting text that is a matched part of a subject string,
2449 it does the same, unless a different locale has been set for the match
2450 (using the /L modifier). In this case, it uses isprint() to decide.
2451
2452 4. Fixed a major bug that caused incorrect computation of the amount of memory
2453 required for a compiled pattern when options that changed within the
2454 pattern affected the logic of the preliminary scan that determines the
2455 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2456 that the computed length was too small. The symptoms of this bug were
2457 either the PCRE error "internal error: code overflow" from pcre_compile(),
2458 or a glibc crash with a message such as "pcretest: free(): invalid next
2459 size (fast)". Examples of patterns that provoked this bug (shown in
2460 pcretest format) are:
2461
2462 /(?-x: )/x
2463 /(?x)(?-x: \s*#\s*)/
2464 /((?i)[\x{c0}])/8
2465 /(?i:[\x{c0}])/8
2466
2467 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2468 is now done differently.
2469
2470 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2471 wrapper classes; (b) implement a new function in the C++ scanner that is
2472 more efficient than the old way of doing things because it avoids levels of
2473 recursion in the regex matching; (c) add a paragraph to the documentation
2474 for the FullMatch() function.
2475
2476 6. The escape sequence \n was being treated as whatever was defined as
2477 "newline". Not only was this contrary to the documentation, which states
2478 that \n is character 10 (hex 0A), but it also went horribly wrong when
2479 "newline" was defined as CRLF. This has been fixed.
2480
2481 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2482 was being set to -1 for the "end of line" case (supposedly a value that no
2483 character can have). Though this value is never used (the check for end of
2484 line is "zero bytes in current character"), it caused compiler complaints.
2485 I've changed it to 0xffffffff.
2486
2487 8. In pcre_version.c, the version string was being built by a sequence of
2488 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2489 string (as it is for production releases) called a macro with an empty
2490 argument. The C standard says the result of this is undefined. The gcc
2491 compiler treats it as an empty string (which was what was wanted) but it is
2492 reported that Visual C gives an error. The source has been hacked around to
2493 avoid this problem.
2494
2495 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2496 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2497 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2498 of them did).
2499
2500 10. Originally, pcretest opened its input and output without "b"; then I was
2501 told that "b" was needed in some environments, so it was added for release
2502 5.0 to both the input and output. (It makes no difference on Unix-like
2503 systems.) Later I was told that it is wrong for the input on Windows. I've
2504 now abstracted the modes into two macros, to make it easier to fiddle with
2505 them, and removed "b" from the input mode under Windows.
2506
2507 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2508
2509 12. Added -help and --help to pcretest as an official way of being reminded
2510 of the options.
2511
2512 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2513 and pcrecpp.cc because they annoy compilers at high warning levels.
2514
2515 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2516
2517 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2518 scripts are not C programs :-) and which was not noticed because it works
2519 on Linux.
2520
2521 16. pcretest is supposed to handle any length of pattern and data line (as one
2522 line or as a continued sequence of lines) by extending its input buffer if
2523 necessary. This feature was broken for very long pattern lines, leading to
2524 a string of junk being passed to pcre_compile() if the pattern was longer
2525 than about 50K.
2526
2527 17. I have done a major re-factoring of the way pcre_compile() computes the
2528 amount of memory needed for a compiled pattern. Previously, there was code
2529 that made a preliminary scan of the pattern in order to do this. That was
2530 OK when PCRE was new, but as the facilities have expanded, it has become
2531 harder and harder to keep it in step with the real compile phase, and there
2532 have been a number of bugs (see for example, 4 above). I have now found a
2533 cunning way of running the real compile function in a "fake" mode that
2534 enables it to compute how much memory it would need, while actually only
2535 ever using a few hundred bytes of working memory and without too many
2536 tests of the mode. This should make future maintenance and development
2537 easier. A side effect of this work is that the limit of 200 on the nesting
2538 depth of parentheses has been removed (though this was never a serious
2539 limitation, I suspect). However, there is a downside: pcre_compile() now
2540 runs more slowly than before (30% or more, depending on the pattern). I
2541 hope this isn't a big issue. There is no effect on runtime performance.
2542
2543 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2544 newline (only possible for the last line of a file) and it was a
2545 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2546
2547 19. Added additional timing features to pcretest. (1) The -tm option now times
2548 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2549 separate command line item, by a number that specifies the number of
2550 repeats to use when timing. The default is 50000; this gives better
2551 precision, but takes uncomfortably long for very large patterns.
2552
2553 20. Extended pcre_study() to be more clever in cases where a branch of a
2554 subpattern has no definite first character. For example, (a*|b*)[cd] would
2555 previously give no result from pcre_study(). Now it recognizes that the
2556 first character must be a, b, c, or d.
2557
2558 21. There was an incorrect error "recursive call could loop indefinitely" if
2559 a subpattern (or the entire pattern) that was being tested for matching an
2560 empty string contained only one non-empty item after a nested subpattern.
2561 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2562 incorrectly, because the \d was being skipped in the check.
2563
2564 22. The pcretest program now has a new pattern option /B and a command line
2565 option -b, which is equivalent to adding /B to every pattern. This causes
2566 it to show the compiled bytecode, without the additional information that
2567 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2568 is the same as /B/I).
2569
2570 23. A new optimization is now able automatically to treat some sequences such
2571 as a*b as a*+b. More specifically, if something simple (such as a character
2572 or a simple class like \d) has an unlimited quantifier, and is followed by
2573 something that cannot possibly match the quantified thing, the quantifier
2574 is automatically "possessified".
2575
2576 24. A recursive reference to a subpattern whose number was greater than 39
2577 went wrong under certain circumstances in UTF-8 mode. This bug could also
2578 have affected the operation of pcre_study().
2579
2580 25. Realized that a little bit of performance could be had by replacing
2581 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2582
2583 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2584
2585 27. Possessive quantifiers such as a++ were previously implemented by turning
2586 them into atomic groups such as ($>a+). Now they have their own opcodes,
2587 which improves performance. This includes the automatically created ones
2588 from 23 above.
2589
2590 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2591 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2592 the first matched character to be a colon. This applied both to named and
2593 numbered groups.
2594
2595 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2596
2597 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2598 building PCRE on Windows easier, so I have included it in the distribution.
2599
2600 31. There is now a check in pcretest against a ridiculously large number being
2601 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2602 loop, the loop is abandoned.
2603
2604 32. Forward references to subpatterns in conditions such as (?(2)...) where
2605 subpattern 2 is defined later cause pcre_compile() to search forwards in
2606 the pattern for the relevant set of parentheses. This search went wrong
2607 when there were unescaped parentheses in a character class, parentheses
2608 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2609
2610 33. "Subroutine" calls and backreferences were previously restricted to
2611 referencing subpatterns earlier in the regex. This restriction has now
2612 been removed.
2613
2614 34. Added a number of extra features that are going to be in Perl 5.10. On the
2615 whole, these are just syntactic alternatives for features that PCRE had
2616 previously implemented using the Python syntax or my own invention. The
2617 other formats are all retained for compatibility.
2618
2619 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2620 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2621 also .NET compatible.
2622
2623 (b) A recursion or subroutine call to a named group can now be defined as
2624 (?&name) as well as (?P>name).
2625
2626 (c) A backreference to a named group can now be defined as \k<name> or
2627 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2628 5.10, are also .NET compatible.
2629
2630 (d) A conditional reference to a named group can now use the syntax
2631 (?(<name>) or (?('name') as well as (?(name).
2632
2633 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2634 groups (named and numbered) that are never evaluated inline, but can be
2635 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2636 is always false. There may be only one alternative in such a group.
2637
2638 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2639 as the simple (?(R). The condition is true only if the most recent
2640 recursion is that of the given number or name. It does not search out
2641 through the entire recursion stack.
2642
2643 (g) The escape \gN or \g{N} has been added, where N is a positive or
2644 negative number, specifying an absolute or relative reference.
2645
2646 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2647 some "unreachable code" warnings.
2648
2649 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2650 things, this adds five new scripts.
2651
2652 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2653 There were also incompatibilities regarding the handling of \Q..\E inside
2654 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2655 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2656
2657 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2658 matches an empty string, and forcibly breaks the loop. There were bugs in
2659 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2660 against aaaa the result was just "a" rather than "aaaa", for example. Two
2661 separate and independent bugs (that affected different cases) have been
2662 fixed.
2663
2664 39. Refactored the code to abolish the use of different opcodes for small
2665 capturing bracket numbers. This is a tidy that I avoided doing when I
2666 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2667 The new approach is not only tidier, it makes it possible to reduce the
2668 memory needed to fix the previous bug (38).
2669
2670 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2671 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2672 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2673 mode.
2674
2675 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2676 report.
2677
2678 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2679 copy construction and assignment in the C++ wrapper.
2680
2681 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2682 couple of bugs that could have given wrong results in the "--newline=crlf"
2683 case.
2684
2685 44. Added a number of casts and did some reorganization of signed/unsigned int
2686 variables following suggestions from Dair Grant. Also renamed the variable
2687 "this" as "item" because it is a C++ keyword.
2688
2689 45. Arranged for dftables to add
2690
2691 #include "pcre_internal.h"
2692
2693 to pcre_chartables.c because without it, gcc 4.x may remove the array
2694 definition from the final binary if PCRE is built into a static library and
2695 dead code stripping is activated.
2696
2697 46. For an unanchored pattern, if a match attempt fails at the start of a
2698 newline sequence, and the newline setting is CRLF or ANY, and the next two
2699 characters are CRLF, advance by two characters instead of one.
2700
2701
2702 Version 6.7 04-Jul-06
2703 ---------------------
2704
2705 1. In order to handle tests when input lines are enormously long, pcretest has
2706 been re-factored so that it automatically extends its buffers when
2707 necessary. The code is crude, but this _is_ just a test program. The
2708 default size has been increased from 32K to 50K.
2709
2710 2. The code in pcre_study() was using the value of the re argument before
2711 testing it for NULL. (Of course, in any sensible call of the function, it
2712 won't be NULL.)
2713
2714 3. The memmove() emulation function in pcre_internal.h, which is used on
2715 systems that lack both memmove() and bcopy() - that is, hardly ever -
2716 was missing a "static" storage class specifier.
2717
2718 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2719 containing an extended class (one that cannot be represented by a bitmap
2720 because it contains high-valued characters or Unicode property items, e.g.
2721 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2722 pattern, but PCRE should not loop if you do not (it no longer does).
2723 [Detail: two cases were found: (a) a repeated subpattern containing an
2724 extended class; (b) a recursive reference to a subpattern that followed a
2725 previous extended class. It wasn't skipping over the extended class
2726 correctly when UTF-8 mode was not set.]
2727
2728 5. A negated single-character class was not being recognized as fixed-length
2729 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2730 compile error "lookbehind assertion is not fixed length".
2731
2732 6. The RunPerlTest auxiliary script was showing an unexpected difference
2733 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2734 write a Perl script that can interpret lines of an input file either as
2735 byte characters or as UTF-8, which is what "perltest" was being required to
2736 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2737 can't do is switch easily at run time between having the "use utf8;" pragma
2738 or not. In the end, I fudged it by using the RunPerlTest script to insert
2739 "use utf8;" explicitly for the UTF-8 tests.
2740
2741 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2742 the end of the subject string, contrary to the documentation and to what
2743 Perl does. This was true of both matching functions. Now it matches only at
2744 the start of the subject and immediately after *internal* newlines.
2745
2746 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2747 a pointer to an int instead of a pointer to an unsigned long int. This
2748 caused problems on 64-bit systems.
2749
2750 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2751 instance of the 'standard' template library not being so standard".
2752
2753 10. There was no check on the number of named subpatterns nor the maximum
2754 length of a subpattern name. The product of these values is used to compute
2755 the size of the memory block for a compiled pattern. By supplying a very
2756 long subpattern name and a large number of named subpatterns, the size
2757 computation could be caused to overflow. This is now prevented by limiting
2758 the length of names to 32 characters, and the number of named subpatterns
2759 to 10,000.
2760
2761 11. Subpatterns that are repeated with specific counts have to be replicated in
2762 the compiled pattern. The size of memory for this was computed from the
2763 length of the subpattern and the repeat count. The latter is limited to
2764 65535, but there was no limit on the former, meaning that integer overflow
2765 could in principle occur. The compiled length of a repeated subpattern is
2766 now limited to 30,000 bytes in order to prevent this.
2767
2768 12. Added the optional facility to have named substrings with the same name.
2769
2770 13. Added the ability to use a named substring as a condition, using the
2771 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2772 are numbers (not recommended). Forward references are permitted.
2773
2774 14. Added forward references in named backreferences (if you see what I mean).
2775
2776 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2777 pattern could run off the end of the subject. For example, the pattern
2778 "(?s)(.{1,5})"8 did this with the subject "ab".
2779
2780 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2781 PCRE_CASELESS was set when matching characters that were quantified with ?
2782 or *.
2783
2784 17. A character class other than a single negated character that had a minimum
2785 but no maximum quantifier - for example [ab]{6,} - was not handled
2786 correctly by pce_dfa_exec(). It would match only one character.
2787
2788 18. A valid (though odd) pattern that looked like a POSIX character
2789 class but used an invalid character after [ (for example [[,abc,]]) caused
2790 pcre_compile() to give the error "Failed: internal error: code overflow" or
2791 in some cases to crash with a glibc free() error. This could even happen if
2792 the pattern terminated after [[ but there just happened to be a sequence of
2793 letters, a binary zero, and a closing ] in the memory that followed.
2794
2795 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2796 over the years. Originally (before any Unicode support), just the bottom 8
2797 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2798 output from "man perlunicode" includes this:
2799
2800 The regular expression compiler produces polymorphic opcodes. That
2801 is, the pattern adapts to the data and automatically switches to
2802 the Unicode character scheme when presented with Unicode data--or
2803 instead uses a traditional byte scheme when presented with byte
2804 data.
2805
2806 Sadly, a wide octal escape does not cause a switch, and in a string with
2807 no other multibyte characters, these octal escapes are treated as before.
2808 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2809 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2810 Unicode string.
2811
2812 I have not perpetrated such confusion in PCRE. Up till now, it took just
2813 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2814 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2815 translate to the appropriate multibyte character.
2816
2817 29. Applied some refactoring to reduce the number of warnings from Microsoft
2818 and Borland compilers. This has included removing the fudge introduced
2819 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2820 a warning about an unused variable.
2821
2822 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2823 characters since release 4.0, because Perl (from release 5.004) does not.
2824 [Or at least, is documented not to: some releases seem to be in conflict
2825 with the documentation.] However, when a pattern was studied with
2826 pcre_study() and all its branches started with \s, PCRE still included VT
2827 as a possible starting character. Of course, this did no harm; it just
2828 caused an unnecessary match attempt.
2829
2830 22. Removed a now-redundant internal flag bit that recorded the fact that case
2831 dependency changed within the pattern. This was once needed for "required
2832 byte" processing, but is no longer used. This recovers a now-scarce options
2833 bit. Also moved the least significant internal flag bit to the most-
2834 significant bit of the word, which was not previously used (hangover from
2835 the days when it was an int rather than a uint) to free up another bit for
2836 the future.
2837
2838 23. Added support for CRLF line endings as well as CR and LF. As well as the
2839 default being selectable at build time, it can now be changed at runtime
2840 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2841 specify that it is scanning data with non-default line endings.
2842
2843 24. Changed the definition of CXXLINK to make it agree with the definition of
2844 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2845
2846 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2847 recursions. This makes a big different to stack usage for some patterns.
2848
2849 26. If a subpattern containing a named recursion or subroutine reference such
2850 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2851 the space required for the compiled pattern went wrong and gave too small a
2852 value. Depending on the environment, this could lead to "Failed: internal
2853 error: code overflow at offset 49" or "glibc detected double free or
2854 corruption" errors.
2855
2856 27. Applied patches from Google (a) to support the new newline modes and (b) to
2857 advance over multibyte UTF-8 characters in GlobalReplace.
2858
2859 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2860 difference for some implementation of PCRE in some Windows version.
2861
2862 29. Added some extra testing facilities to pcretest:
2863
2864 \q<number> in a data line sets the "match limit" value
2865 \Q<number> in a data line sets the "match recursion limt" value
2866 -S <number> sets the stack size, where <number> is in megabytes
2867
2868 The -S option isn't available for Windows.
2869
2870
2871 Version 6.6 06-Feb-06
2872 ---------------------
2873
2874 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2875 in pcreposix.h. I have copied the definition from pcre.h.
2876
2877 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2878 because pcre.h is no longer a built file.
2879
2880 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2881 not normally included in the compiled code.
2882
2883
2884 Version 6.5 01-Feb-06
2885 ---------------------
2886
2887 1. When using the partial match feature with pcre_dfa_exec(), it was not
2888 anchoring the second and subsequent partial matches at the new starting
2889 point. This could lead to incorrect results. For example, with the pattern
2890 /1234/, partially matching against "123" and then "a4" gave a match.
2891
2892 2. Changes to pcregrep:
2893
2894 (a) All non-match returns from pcre_exec() were being treated as failures
2895 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2896 error message is output. Some extra information is given for the
2897 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2898 probably the only errors that are likely to be caused by users (by
2899 specifying a regex that has nested indefinite repeats, for instance).
2900 If there are more than 20 of these errors, pcregrep is abandoned.
2901
2902 (b) A binary zero was treated as data while matching, but terminated the
2903 output line if it was written out. This has been fixed: binary zeroes
2904 are now no different to any other data bytes.
2905
2906 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2907 used to set a locale for matching. The --locale=xxxx long option has
2908 been added (no short equivalent) to specify a locale explicitly on the
2909 pcregrep command, overriding the environment variables.
2910
2911 (d) When -B was used with -n, some line numbers in the output were one less
2912 than they should have been.
2913
2914 (e) Added the -o (--only-matching) option.
2915
2916 (f) If -A or -C was used with -c (count only), some lines of context were
2917 accidentally printed for the final match.
2918
2919 (g) Added the -H (--with-filename) option.
2920
2921 (h) The combination of options -rh failed to suppress file names for files
2922 that were found from directory arguments.
2923
2924 (i) Added the -D (--devices) and -d (--directories) options.
2925
2926 (j) Added the -F (--fixed-strings) option.
2927
2928 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2929
2930 (l) Added the --colo(u)r option.
2931
2932 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2933 is not present by default.
2934
2935 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2936 items such as (?R) or (?1), when the recursion could match a number of
2937 alternatives. If it matched one of the alternatives, but subsequently,
2938 outside the recursion, there was a failure, the code tried to back up into
2939 the recursion. However, because of the way PCRE is implemented, this is not
2940 possible, and the result was an incorrect result from the match.
2941
2942 In order to prevent this happening, the specification of recursion has
2943 been changed so that all such subpatterns are automatically treated as
2944 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2945
2946 4. I had overlooked the fact that, in some locales, there are characters for
2947 which isalpha() is true but neither isupper() nor islower() are true. In
2948 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2949 and ordfeminine) are like this. This affected the treatment of \w and \W
2950 when they appeared in character classes, but not when they appeared outside
2951 a character class. The bit map for "word" characters is now created
2952 separately from the results of isalnum() instead of just taking it from the
2953 upper, lower, and digit maps. (Plus the underscore character, of course.)
2954
2955 5. The above bug also affected the handling of POSIX character classes such as
2956 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2957 permanent tables. Instead, the bit maps for such a class were previously
2958 created as the appropriate unions of the upper, lower, and digit bitmaps.
2959 Now they are created by subtraction from the [[:word:]] class, which has
2960 its own bitmap.
2961
2962 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2963 It is created by subtracting the vertical space characters (\x09, \x0a,
2964 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2965 subtraction was done in the overall bitmap for a character class, meaning
2966 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2967 be recognized. This bug has been fixed.
2968
2969 7. Patches from the folks at Google:
2970
2971 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2972 real life, but is still worth protecting against".
2973
2974 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2975 regular expressions".
2976
2977 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2978 have it.
2979
2980 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2981 "configure" and the latter not, in order to fix a problem somebody had
2982 with compiling the Arg class on HP-UX.
2983
2984 (e) Improve the error-handling of the C++ wrapper a little bit.
2985
2986 (f) New tests for checking recursion limiting.
2987
2988 8. The pcre_memmove() function, which is used only if the environment does not
2989 have a standard memmove() function (and is therefore rarely compiled),
2990 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2991 returning a result (though PCRE never actually uses the result).
2992
2993 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2994 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2995 returned instead of calling malloc() with an overflowing number that would
2996 most likely cause subsequent chaos.
2997
2998 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2999
3000 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3001 with this option is matched, the nmatch and pmatch options of regexec() are
3002 ignored.
3003
3004 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3005 provided in case anyone wants to the the POSIX interface with UTF-8
3006 strings.
3007
3008 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3009 C++ linking (needed for some HP-UX environments).
3010
3011 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3012 (unused parameter) and in the pcre_printint() function (omitted "default"
3013 switch label when the default is to do nothing).
3014
3015 15. Added some code to make it possible, when PCRE is compiled as a C++
3016 library, to replace subject pointers for pcre_exec() with a smart pointer
3017 class, thus making it possible to process discontinuous strings.
3018
3019 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3020 much the same function. They were added by different people who were trying
3021 to make PCRE easy to compile on non-Unix systems. It has been suggested
3022 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3023 for compiling on Windows systems. I have therefore replaced it with
3024 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3025 defaults to "extern" for C or "extern C" for C++, which works fine on
3026 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3027 SCOPE with something explicit in config.h. In addition:
3028
3029 (a) pcreposix.h still had just "extern" instead of either of these macros;
3030 I have replaced it with PCRE_DATA_SCOPE.
3031
3032 (b) Functions such as _pcre_xclass(), which are internal to the library,
3033 but external in the C sense, all had PCRE_EXPORT in their definitions.
3034 This is apparently wrong for the Windows case, so I have removed it.
3035 (It makes no difference on Unix-like systems.)
3036
3037 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3038 of recursive calls to match(). This is different to MATCH_LIMIT because
3039 that limits the total number of calls to match(), not all of which increase
3040 the depth of recursion. Limiting the recursion depth limits the amount of
3041 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3042 when PCRE is compiled, and changed at run time. A patch from Google adds
3043 this functionality to the C++ interface.
3044
3045 18. Changes to the handling of Unicode character properties:
3046
3047 (a) Updated the table to Unicode 4.1.0.
3048
3049 (b) Recognize characters that are not in the table as "Cn" (undefined).
3050
3051 (c) I revised the way the table is implemented to a much improved format
3052 which includes recognition of ranges. It now supports the ranges that
3053 are defined in UnicodeData.txt, and it also amalgamates other
3054 characters into ranges. This has reduced the number of entries in the
3055 table from around 16,000 to around 3,000, thus reducing its size
3056 considerably. I realized I did not need to use a tree structure after
3057 all - a binary chop search is just as efficient. Having reduced the
3058 number of entries, I extended their size from 6 bytes to 8 bytes to
3059 allow for more data.
3060
3061 (d) Added support for Unicode script names via properties such as \p{Han}.
3062
3063 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3064 matching that character.
3065
3066 20. When matching a repeated Unicode property with a minimum greater than zero,
3067 (for example \pL{2,}), PCRE could look past the end of the subject if it
3068 reached it while seeking the minimum number of characters. This could
3069 happen only if some of the characters were more than one byte long, because
3070 there is a check for at least the minimum number of bytes.
3071
3072 21. Refactored the implementation of \p and \P so as to be more general, to
3073 allow for more different types of property in future. This has changed the
3074 compiled form incompatibly. Anybody with saved compiled patterns that use
3075 \p or \P will have to recompile them.
3076
3077 22. Added "Any" and "L&" to the supported property types.
3078
3079 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3080 but give a compile time error if the value is greater than 0xff.
3081
3082 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3083 accidentally not being installed or uninstalled.
3084
3085 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3086 made were to insert the current release number. This seemed silly, because
3087 it made things harder for people building PCRE on systems that don't run
3088 "configure". I have turned pcre.h into a distributed file, no longer built
3089 by "configure", with the version identification directly included. There is
3090 no longer a pcre.h.in file.
3091
3092 However, this change necessitated a change to the pcre-config script as
3093 well. It is built from pcre-config.in, and one of the substitutions was the
3094 release number. I have updated configure.ac so that ./configure now finds
3095 the release number by grepping pcre.h.
3096
3097 26. Added the ability to run the tests under valgrind.
3098
3099
3100 Version 6.4 05-Sep-05
3101 ---------------------
3102
3103 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3104 "--" to be printed when multiple files were scanned, even when none of the
3105 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3106 consider it to be a bug, and have restored the previous behaviour.
3107
3108 2. A couple of code tidies to get rid of compiler warnings.
3109
3110 3. The pcretest program used to cheat by referring to symbols in the library
3111 whose names begin with _pcre_. These are internal symbols that are not
3112 really supposed to be visible externally, and in some environments it is
3113 possible to suppress them. The cheating is now confined to including
3114 certain files from the library's source, which is a bit cleaner.
3115
3116 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3117 file's purpose clearer.
3118
3119 5. Reorganized pcre_ucp_findchar().
3120
3121
3122 Version 6.3 15-Aug-05
3123 ---------------------
3124
3125 1. The file libpcre.pc.in did not have general read permission in the tarball.
3126
3127 2. There were some problems when building without C++ support:
3128
3129 (a) If C++ support was not built, "make install" and "make test" still
3130 tried to test it.
3131
3132 (b) There were problems when the value of CXX was explicitly set. Some
3133 changes have been made to try to fix these, and ...
3134
3135 (c) --disable-cpp can now be used to explicitly disable C++ support.
3136
3137 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3138 backslash in a target when C++ was disabled. This confuses some
3139 versions of "make", apparently. Using an intermediate variable solves
3140 this. (Same for CPP_LOBJ.)
3141
3142 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3143 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3144 necessary on certain architectures.
3145
3146 4. Added a setting of -export-symbols-regex to the link command to remove
3147 those symbols that are exported in the C sense, but actually are local
3148 within the library, and not documented. Their names all begin with
3149 "_pcre_". This is not a perfect job, because (a) we have to except some
3150 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3151 available (and never for static libraries). I have made a note to try to
3152 find a way round (a) in the future.
3153
3154
3155 Version 6.2 01-Aug-05
3156 ---------------------
3157
3158 1. There was no test for integer overflow of quantifier values. A construction
3159 such as {1111111111111111} would give undefined results. What is worse, if
3160 a minimum quantifier for a parenthesized subpattern overflowed and became
3161 negative, the calculation of the memory size went wrong. This could have
3162 led to memory overwriting.
3163
3164 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3165
3166 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3167 operating environments where this matters.
3168
3169 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3170 PCRE options from within the C++ wrapper.
3171
3172 5. Named capturing subpatterns were not being correctly counted when a pattern
3173 was compiled. This caused two problems: (a) If there were more than 100
3174 such subpatterns, the calculation of the memory needed for the whole
3175 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3176 back references of the form \12, where the number was greater than 9, were
3177 not recognized as back references, even though there were sufficient
3178 previous subpatterns.
3179
3180 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3181 versions of gcc, e.g. 2.95.4.
3182
3183
3184 Version 6.1 21-Jun-05
3185 ---------------------
3186
3187 1. There was one reference to the variable "posix" in pcretest.c that was not
3188 surrounded by "#if !defined NOPOSIX".
3189
3190 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3191 the cross-check on the old pcre_info() function, for the benefit of the
3192 cut-down version of PCRE that is currently imported into Exim.
3193
3194 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3195 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3196 patterns that start (?i) but I don't think that matters. The use of (?i) is
3197 just an example; this all applies to the other options as well.
3198
3199 4. Since libtool seems to echo the compile commands it is issuing, the output
3200 from "make" can be reduced a bit by putting "@" in front of each libtool
3201 compile command.
3202
3203 5. Patch from the folks at Google for configure.in to be a bit more thorough
3204 in checking for a suitable C++ installation before trying to compile the
3205 C++ stuff. This should fix a reported problem when a compiler was present,
3206 but no suitable headers.
3207
3208 6. The man pages all had just "PCRE" as their title. I have changed them to
3209 be the relevant file name. I have also arranged that these names are
3210 retained in the file doc/pcre.txt, which is a concatenation in text format
3211 of all the man pages except the little individual ones for each function.
3212
3213 7. The NON-UNIX-USE file had not been updated for the different set of source
3214 files that come with release 6. I also added a few comments about the C++
3215 wrapper.
3216
3217
3218 Version 6.0 07-Jun-05
3219 ---------------------
3220
3221 1. Some minor internal re-organization to help with my DFA experiments.
3222
3223 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3224 didn't matter for the library itself when fully configured, but did matter
3225 when compiling without UCP support, or within Exim, where the ucp files are
3226 not imported.
3227
3228 3. Refactoring of the library code to split up the various functions into
3229 different source modules. The addition of the new DFA matching code (see
3230 below) to a single monolithic source would have made it really too
3231 unwieldy, quite apart from causing all the code to be include in a
3232 statically linked application, when only some functions are used. This is
3233 relevant even without the DFA addition now that patterns can be compiled in
3234 one application and matched in another.
3235
3236 The downside of splitting up is that there have to be some external
3237 functions and data tables that are used internally in different modules of
3238 the library but which are not part of the API. These have all had their
3239 names changed to start with "_pcre_" so that they are unlikely to clash
3240 with other external names.
3241
3242 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3243 a different (DFA) algorithm. Although it is slower than the original
3244 function, it does have some advantages for certain types of matching
3245 problem.
3246
3247 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3248 including restarting after a partial match.
3249
3250 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3251 defined when compiling for Windows was sent to me. I have put it into the
3252 code, though I have no means of testing or verifying it.
3253
3254 7. Added the pcre_refcount() auxiliary function.
3255
3256 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3257 match before or at the first newline in the subject string. In pcretest,
3258 the /f option on a pattern can be used to set this.
3259
3260 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3261 would behave wrongly. This has been present in PCRE since release 4.0.
3262
3263 10. A number of changes to the pcregrep command:
3264
3265 (a) Refactored how -x works; insert ^(...)$ instead of setting
3266 PCRE_ANCHORED and checking the length, in preparation for adding
3267 something similar for -w.
3268
3269 (b) Added the -w (match as a word) option.
3270
3271 (c) Refactored the way lines are read and buffered so as to have more
3272 than one at a time available.
3273
3274 (d) Implemented a pcregrep test script.
3275
3276 (e) Added the -M (multiline match) option. This allows patterns to match
3277 over several lines of the subject. The buffering ensures that at least
3278 8K, or the rest of the document (whichever is the shorter) is available
3279 for matching (and similarly the previous 8K for lookbehind assertions).
3280
3281 (f) Changed the --help output so that it now says
3282
3283 -w, --word-regex(p)
3284
3285 instead of two lines, one with "regex" and the other with "regexp"
3286 because that confused at least one person since the short forms are the
3287 same. (This required a bit of code, as the output is generated
3288 automatically from a table. It wasn't just a text change.)
3289
3290 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3291 option but starts with a hyphen. Could be a pattern or a path name
3292 starting with a hyphen, for instance.
3293
3294 (h) "-" can be given as a file name to represent stdin.
3295
3296 (i) When file names are being printed, "(standard input)" is used for
3297 the standard input, for compatibility with GNU grep. Previously
3298 "<stdin>" was used.
3299
3300 (j) The option --label=xxx can be used to supply a name to be used for
3301 stdin when file names are being printed. There is no short form.
3302
3303 (k) Re-factored the options decoding logic because we are going to add
3304 two more options that take data. Such options can now be given in four
3305 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3306
3307 (l) Added the -A, -B, and -C options for requesting that lines of context
3308 around matches be printed.
3309
3310 (m) Added the -L option to print the names of files that do not contain
3311 any matching lines, that is, the complement of -l.
3312
3313 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3314 continue to scan other files.
3315
3316 (o) The -s option was incorrectly implemented. For compatibility with other
3317 greps, it now suppresses the error message for a non-existent or non-
3318 accessible file (but not the return code). There is a new option called
3319 -q that suppresses the output of matching lines, which was what -s was
3320 previously doing.
3321
3322 (p) Added --include and --exclude options to specify files for inclusion
3323 and exclusion when recursing.
3324
3325 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3326 Hopefully, it now does.
3327
3328 12. Missing cast in pcre_study().
3329
3330 13. Added an "uninstall" target to the makefile.
3331
3332 14. Replaced "extern" in the function prototypes in Makefile.in with
3333 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3334 world, but is set differently for Windows.
3335
3336 15. Added a second compiling function called pcre_compile2(). The only
3337 difference is that it has an extra argument, which is a pointer to an
3338 integer error code. When there is a compile-time failure, this is set
3339 non-zero, in addition to the error test pointer being set to point to an
3340 error message. The new argument may be NULL if no error number is required
3341 (but then you may as well call pcre_compile(), which is now just a
3342 wrapper). This facility is provided because some applications need a
3343 numeric error indication, but it has also enabled me to tidy up the way
3344 compile-time errors are handled in the POSIX wrapper.
3345
3346 16. Added VPATH=.libs to the makefile; this should help when building with one
3347 prefix path and installing with another. (Or so I'm told by someone who
3348 knows more about this stuff than I do.)
3349
3350 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3351 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3352 match everything, including newlines. This is not POSIX-compatible, but
3353 somebody wanted the feature. From pcretest it can be activated by using
3354 both the P and the s flags.
3355
3356 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3357
3358 19. libpcre.pc was being incorrectly installed as executable.
3359
3360 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3361 it now also looks for '\r' so that it will work unmodified on Windows.
3362
3363 21. Added Google's contributed C++ wrapper to the distribution.
3364
3365 22. Added some untidy missing memory free() calls in pcretest, to keep
3366 Electric Fence happy when testing.
3367
3368
3369
3370 Version 5.0 13-Sep-04
3371 ---------------------
3372
3373 1. Internal change: literal characters are no longer packed up into items
3374 containing multiple characters in a single byte-string. Each character
3375 is now matched using a separate opcode. However, there may be more than one
3376 byte in the character in UTF-8 mode.
3377
3378 2. The pcre_callout_block structure has two new fields: pattern_position and
3379 next_item_length. These contain the offset in the pattern to the next match
3380 item, and its length, respectively.
3381
3382 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3383 insertion of callouts before each pattern item. Added the /C option to
3384 pcretest to make use of this.
3385
3386 4. On the advice of a Windows user, the lines
3387
3388 #if defined(_WIN32) || defined(WIN32)
3389 _setmode( _fileno( stdout ), 0x8000 );
3390 #endif /* defined(_WIN32) || defined(WIN32) */
3391
3392 have been added to the source of pcretest. This apparently does useful
3393 magic in relation to line terminators.
3394
3395 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3396 for the benefit of those environments where the "b" makes a difference.
3397
3398 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3399 to know about it. I have put a hack into configure.in that adds in code
3400 to set GCC=yes if CC=icc. This seems to end up at a point in the
3401 generated configure script that is early enough to affect the setting of
3402 compiler options, which is what is needed, but I have no means of testing
3403 whether it really works. (The user who reported this had patched the
3404 generated configure script, which of course I cannot do.)
3405
3406 LATER: After change 22 below (new libtool files), the configure script
3407 seems to know about icc (and also ecc). Therefore, I have commented out
3408 this hack in configure.in.
3409
3410 7. Added support for pkg-config (2 patches were sent in).
3411
3412 8. Negated POSIX character classes that used a combination of internal tables
3413 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3414 [[:^ascii]]. Typically, they would match almost any characters. The other
3415 POSIX classes were not broken in this way.
3416
3417 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3418 to find the match, as PCRE was deluded into thinking that the match had to
3419 start at the start point or following a newline. The same bug applied to
3420 patterns with negative forward assertions or any backward assertions
3421 preceding ".*" at the start, unless the pattern required a fixed first
3422 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3423
3424 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3425 starting at the last subject character, bytes beyond the end of the subject
3426 string were read.
3427
3428 11. Renamed the variable "class" as "classbits" to make life easier for C++
3429 users. (Previously there was a macro definition, but it apparently wasn't
3430 enough.)
3431
3432 12. Added the new field "tables" to the extra data so that tables can be passed
3433 in at exec time, or the internal tables can be re-selected. This allows
3434 a compiled regex to be saved and re-used at a later time by a different
3435 program that might have everything at different addresses.
3436
3437 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3438 -R library as well as a -L library.
3439
3440 14. The debugging options of pcretest (-d on the command line or D on a
3441 pattern) showed incorrect output for anything following an extended class
3442 that contained multibyte characters and which was followed by a quantifier.
3443
3444 15. Added optional support for general category Unicode character properties
3445 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3446 support. It adds about 90K to the size of the library. The meanings of the
3447 inbuilt class escapes such as \d and \s have NOT been changed.
3448
3449 16. Updated pcredemo.c to include calls to free() to release the memory for the
3450 compiled pattern.
3451
3452 17. The generated file chartables.c was being created in the source directory
3453 instead of in the building directory. This caused the build to fail if the
3454 source directory was different from the building directory, and was
3455 read-only.
3456
3457 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3458 file. No doubt somebody will tell me if they don't make sense... Also added
3459 Dan Mooney's comments about building on OpenVMS.
3460
3461 19. Added support for partial matching via the PCRE_PARTIAL option for
3462 pcre_exec() and the \P data escape in pcretest.
3463
3464 20. Extended pcretest with 3 new pattern features:
3465
3466 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3467 write the compiled pattern to the file whose name is "rest-of-line".
3468 This is a straight binary dump of the data, with the saved pointer to
3469 the character tables forced to be NULL. The study data, if any, is
3470 written too. After writing, pcretest reads a new pattern.
3471
3472 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3473 compiled pattern from the given file. There must not be any
3474 occurrences of "<" in the file name (pretty unlikely); if there are,
3475 pcretest will instead treat the initial "<" as a pattern delimiter.
3476 After reading in the pattern, pcretest goes on to read data lines as
3477 usual.
3478
3479 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3480 and 16-bit fields in a compiled pattern, to simulate a pattern that
3481 was compiled on a host of opposite endianness.
3482
3483 21. The pcre-exec() function can now cope with patterns that were compiled on
3484 hosts of opposite endianness, with this restriction:
3485
3486 As for any compiled expression that is saved and used later, the tables
3487 pointer field cannot be preserved; the extra_data field in the arguments
3488 to pcre_exec() should be used to pass in a tables address if a value
3489 other than the default internal tables were used at compile time.
3490
3491 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3492 now diagnosed as an error. Previously, most of the time, a negative number
3493 would have been treated as zero, but if in addition "ovector" was passed as
3494 NULL, a crash could occur.
3495
3496 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3497 new versions from the libtool 1.5 distribution (the last one is a copy of
3498 a file called libtool.m4). This seems to have fixed the need to patch
3499 "configure" to support Darwin 1.3 (which I used to do). However, I still
3500 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3501 workstation).
3502
3503 24. Changed the PCRE licence to be the more standard "BSD" licence.
3504
3505
3506 Version 4.5 01-Dec-03
3507 ---------------------
3508
3509 1. There has been some re-arrangement of the code for the match() function so
3510 that it can be compiled in a version that does not call itself recursively.
3511 Instead, it keeps those local variables that need separate instances for
3512 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3513 needs to "recurse". Keeping track of where control must go is done by means
3514 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3515 hide most of the details from the main code, and operates only if
3516 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3517 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3518 operating.
3519
3520 To make it easier for callers to provide specially tailored get/free
3521 functions for this usage, two new functions, pcre_stack_malloc, and
3522 pcre_stack_free, are used. They are always called in strict stacking order,
3523 and the size of block requested is always the same.
3524
3525 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3526 PCRE has been compiled to use the stack or the heap for recursion. The
3527 -C option of pcretest uses this to show which version is compiled.
3528
3529 A new data escape \S, is added to pcretest; it causes the amounts of store
3530 obtained and freed by both kinds of malloc/free at match time to be added
3531 to the output.
3532
3533 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3534 what's available on my current Linux desktop machine.
3535
3536 3. When matching a UTF-8 string, the test for a valid string at the start has
3537 been extended. If start_offset is not zero, PCRE now checks that it points
3538 to a byte that is the start of a UTF-8 character. If not, it returns
3539 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3540 this is necessary because there may be backward assertions in the pattern.
3541 When matching the same subject several times, it may save resources to use
3542 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3543
3544 4. The code for checking the validity of UTF-8 strings has been tightened so
3545 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3546 containing "overlong sequences".
3547
3548 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3549 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3550 should have been "&", but it just so happened that all the cases this let
3551 through by mistake were picked up later in the function.
3552
3553 6. I had used a variable called "isblank" - this is a C99 function, causing
3554 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3555
3556 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3557 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3558 so the version is shown; (c) stop "make test" echoing "./RunTest".
3559
3560 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3561
3562 9. The prototype for memmove() for systems that don't have it was using
3563 size_t, but the inclusion of the header that defines size_t was later. I've
3564 moved the #includes for the C headers earlier to avoid this.
3565
3566 10. Added some adjustments to the code to make it easier to compiler on certain
3567 special systems:
3568
3569 (a) Some "const" qualifiers were missing.
3570 (b) Added the macro EXPORT before all exported functions; by default this
3571 is defined to be empty.
3572 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3573 that it reads its output file name as an argument instead of writing
3574 to the standard output and assuming this can be redirected.
3575
3576 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3577 class containing characters with values greater than 255, PCRE compilation
3578 went into a loop.
3579
3580 12. A recursive reference to a subpattern that was within another subpattern
3581 that had a minimum quantifier of zero caused PCRE to crash. For example,
3582 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3583 recursion. If the recursively-called subpattern itself had a zero repeat,
3584 that was OK.
3585
3586 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3587 buffer into which it was copied (for escape processing) was still set at
3588 1024, so long lines caused crashes.
3589
3590 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3591 "internal error: code overflow...". This applied to any character class
3592 that was followed by a possessive quantifier.
3593
3594 15. Modified the Makefile to add libpcre.la as a prerequisite for
3595 libpcreposix.la because I was told this is needed for a parallel build to
3596 work.
3597
3598 16. If a pattern that contained .* following optional items at the start was
3599 studied, the wrong optimizing data was generated, leading to matching
3600 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3601 matching string must start with a or b or c. The correct conclusion for
3602 this pattern is that a match can start with any character.
3603
3604
3605 Version 4.4 13-Aug-03
3606 ---------------------
3607
3608 1. In UTF-8 mode, a character class containing characters with values between
3609 127 and 255 was not handled correctly if the compiled pattern was studied.
3610 In fixing this, I have also improved the studying algorithm for such
3611 classes (slightly).
3612
3613 2. Three internal functions had redundant arguments passed to them. Removal
3614 might give a very teeny performance improvement.
3615
3616 3. Documentation bug: the value of the capture_top field in a callout is *one
3617 more than* the number of the hightest numbered captured substring.
3618
3619 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3620 in incorrectly linking with a previously installed version. They now link
3621 explicitly with libpcre.la.
3622
3623 5. configure.in no longer needs to recognize Cygwin specially.
3624
3625 6. A problem in pcre.in for Windows platforms is fixed.
3626
3627 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3628 pcretest, it used to include the size of the study block as part of its
3629 output. Unfortunately, the structure contains a field that has a different
3630 size on different hardware architectures. This meant that the tests that
3631 showed this size failed. As the block is currently always of a fixed size,
3632 this information isn't actually particularly useful in pcretest output, so
3633 I have just removed it.
3634
3635 8. Three pre-processor statements accidentally did not start in column 1.
3636 Sadly, there are *still* compilers around that complain, even though
3637 standard C has not required this for well over a decade. Sigh.
3638
3639 9. In pcretest, the code for checking callouts passed small integers in the
3640 callout_data field, which is a void * field. However, some picky compilers
3641 complained about the casts involved for this on 64-bit systems. Now
3642 pcretest passes the address of the small integer instead, which should get
3643 rid of the warnings.
3644
3645 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3646 both compile and run time, and gives an error if an invalid UTF-8 sequence
3647 is found. There is a option for disabling this check in cases where the
3648 string is known to be correct and/or the maximum performance is wanted.
3649
3650 11. In response to a bug report, I changed one line in Makefile.in from
3651
3652 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3653 to
3654 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3655
3656 to look similar to other lines, but I have no way of telling whether this
3657 is the right thing to do, as I do not use Windows. No doubt I'll get told
3658 if it's wrong...
3659
3660
3661 Version 4.3 21-May-03
3662 ---------------------
3663
3664 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3665 Makefile.
3666
3667 2. Some refactoring to improve the quality of the code:
3668
3669 (i) The utf8_table... variables are now declared "const".
3670
3671 (ii) The code for \cx, which used the "case flipping" table to upper case
3672 lower case letters, now just substracts 32. This is ASCII-specific,
3673 but the whole concept of \cx is ASCII-specific, so it seems
3674 reasonable.
3675
3676 (iii) PCRE was using its character types table to recognize decimal and
3677 hexadecimal digits in the pattern. This is silly, because it handles
3678 only 0-9, a-f, and A-F, but the character types table is locale-
3679 specific, which means strange things might happen. A private
3680 table is now used for this - though it costs 256 bytes, a table is
3681 much faster than multiple explicit tests. Of course, the standard
3682 character types table is still used for matching digits in subject
3683 strings against \d.
3684
3685 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3686 ending in _t are). So I've renamed it as ESC_tee.
3687
3688 3. The first argument for regexec() in the POSIX wrapper should have been
3689 defined as "const".
3690
3691 4. Changed pcretest to use malloc() for its buffers so that they can be
3692 Electric Fenced for debugging.
3693
3694 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3695 to read one or more bytes before the start of the subject string. Often this
3696 had no effect on PCRE's behaviour, but in some circumstances it could
3697 provoke a segmentation fault.
3698
3699 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3700 to try to read one or more bytes before the start of the subject string.
3701
3702 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3703 UTF-8 support could misbehave in various ways if the subject string
3704 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3705 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3706 back over UTF-8 characters.)
3707
3708
3709 Version 4.2 14-Apr-03
3710 ---------------------
3711
3712 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3713
3714 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3715 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3716 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3717 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3718 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3719 and BUILD_EXEEXT
3720 Note: automatic setting of the BUILD variables is not yet working
3721 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3722 compile-time but not at link-time
3723 [LINK]: use for linking executables only
3724 make different versions for Windows and non-Windows
3725 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3726 libraries
3727 [LINK_FOR_BUILD]: new variable
3728 [OBJEXT]: use throughout
3729 [EXEEXT]: use throughout
3730 <winshared>: new target
3731 <wininstall>: new target
3732 <dftables.o>: use native compiler
3733 <dftables>: use native linker
3734 <install>: handle Windows platform correctly
3735 <clean>: ditto
3736 <check>: ditto
3737 copy DLL to top builddir before testing
3738
3739 As part of these changes, -no-undefined was removed again. This was reported
3740 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3741 in any case.
3742
3743 3. Some tidies to get rid of compiler warnings:
3744
3745 . In the match_data structure, match_limit was an unsigned long int, whereas
3746 match_call_count was an int. I've made them both unsigned long ints.
3747
3748 . In pcretest the fact that a const uschar * doesn't automatically cast to
3749 a void * provoked a warning.
3750
3751 . Turning on some more compiler warnings threw up some "shadow" variables
3752 and a few more missing casts.
3753
3754 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3755 option, a class that contained a single character with a value between 128
3756 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3757
3758 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3759 option, a class that contained several characters, but with at least one
3760 whose value was between 128 and 255 caused PCRE to crash.
3761
3762
3763 Version 4.1 12-Mar-03
3764 ---------------------
3765
3766 1. Compiling with gcc -pedantic found a couple of places where casts were
3767 needed, and a string in dftables.c that was longer than standard compilers are
3768 required to support.
3769
3770 2. Compiling with Sun's compiler found a few more places where the code could
3771 be tidied up in order to avoid warnings.
3772
3773 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3774 first of these names is deprecated in the latest Autoconf in favour of the name
3775 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3776 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3777 analogy I have changed it to CFLAGS_FOR_BUILD.
3778
3779 4. Added -no-undefined to the linking command in the Makefile, because this is
3780 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3781 linking step for the pcreposix library.
3782
3783 5. PCRE was failing to diagnose the case of two named groups with the same
3784 name.
3785
3786 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3787 literal character that is needed in the subject for a match, and scans along to
3788 ensure that it is present before embarking on the full matching process. This
3789 saves time in cases of nested unlimited repeats that are never going to match.
3790 Problem: the scan can take a lot of time if the subject is very long (e.g.
3791 megabytes), thus penalizing straightforward matches. It is now done only if the
3792 amount of subject to be scanned is less than 1000 bytes.
3793
3794 7. A lesser problem with the same optimization is that it was recording the
3795 first character of an anchored pattern as "needed", thus provoking a search
3796 right along the subject, even when the first match of the pattern was going to
3797 fail. The "needed" character is now not set for anchored patterns, unless it
3798 follows something in the pattern that is of non-fixed length. Thus, it still
3799 fulfils its original purpose of finding quick non-matches in cases of nested
3800 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3801
3802
3803 Version 4.0 17-Feb-03
3804 ---------------------
3805
3806 1. If a comment in an extended regex that started immediately after a meta-item
3807 extended to the end of string, PCRE compiled incorrect data. This could lead to
3808 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3809
3810 2. Moved to autoconf 2.53 and libtool 1.4.2.
3811
3812 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3813 the special perltest8 script is no longer needed - all the tests can be run
3814 from a single perltest script.
3815
3816 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3817 by \s. It has now been removed in PCRE. This means it isn't recognized as
3818 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3819 class [:space:] *does* include VT, thereby creating a mess.
3820
3821 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3822 space and tab.
3823
3824 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3825 its new features into the main test script, reducing the number of scripts.
3826
3827 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3828 were backward compatible, and made the (?i) apply to the whole pattern, as if
3829 /i were given. Now it behaves more logically, and applies the option setting
3830 only to what follows. PCRE has been changed to follow suit. However, if it
3831 finds options settings right at the start of the pattern, it extracts them into
3832 the global options, as before. Thus, they show up in the info data.
3833
3834 8. Added support for the \Q...\E escape sequence. Characters in between are
3835 treated as literals. This is slightly different from Perl in that $ and @ are
3836 also handled as literals inside the quotes. In Perl, they will cause variable
3837 interpolation. Note the following examples:
3838
3839 Pattern PCRE matches Perl matches
3840
3841 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3842 \Qabc\$xyz\E abc\$xyz abc\$xyz
3843 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3844
3845 For compatibility with Perl, \Q...\E sequences are recognized inside character
3846 classes as well as outside them.
3847
3848 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3849 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3850 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3851 signed/unsigned warnings.
3852
3853 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3854 option for pcretest, so I've replaced it by a simple function that does just
3855 that job.
3856
3857 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3858 "pcregrep -".
3859
3860 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3861 Java package. This provides some syntactic sugar for simple cases of what my
3862 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3863 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3864 item, you can use this simplified notation. Note that only makes sense with
3865 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3866 greediness, whatever the setting of the PCRE_UNGREEDY option.
3867
3868 13. A change of greediness default within a pattern was not taking effect at
3869 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3870 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3871 was abstracted outside.
3872
3873 14. PCRE now supports the \G assertion. It is true when the current matching
3874 position is at the start point of the match. This differs from \A when the
3875 starting offset is non-zero. Used with the /g option of pcretest (or similar
3876 code), it works in the same way as it does for Perl's /g option. If all
3877 alternatives of a regex begin with \G, the expression is anchored to the start
3878 match position, and the "anchored" flag is set in the compiled expression.
3879
3880 15. Some bugs concerning the handling of certain option changes within patterns
3881 have been fixed. These applied to options other than (?ims). For example,
3882 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3883 been the other way round. Some of this was related to change 7 above.
3884
3885 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3886 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3887 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3888 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3889
3890 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3891 mode. Unlike ".", it always matches newline, whatever the setting of
3892 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3893 assertions. Perl allows it, but it doesn't (in general) work because it can't
3894 calculate the length of the lookbehind. At least, that's the case for Perl
3895 5.8.0 - I've been told they are going to document that it doesn't work in
3896 future.
3897
3898 18. Added an error diagnosis for escapes that PCRE does not support: these are
3899 \L, \l, \N, \P, \p, \U, \u, and \X.
3900
3901 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3902 reading past the end of the pattern in cases such as /[abcd/.
3903
3904 20. PCRE was getting more memory than necessary for patterns with classes that
3905 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3906
3907 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3908 compiling PCRE for use with Virtual Pascal.
3909
3910 22. Small fix to the Makefile to make it work properly if the build is done
3911 outside the source tree.
3912
3913 23. Added a new extension: a condition to go with recursion. If a conditional
3914 subpattern starts with (?(R) the "true" branch is used if recursion has
3915 happened, whereas the "false" branch is used only at the top level.
3916
3917 24. When there was a very long string of literal characters (over 255 bytes
3918 without UTF support, over 250 bytes with UTF support), the computation of how
3919 much memory was required could be incorrect, leading to segfaults or other
3920 strange effects.
3921
3922 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3923 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3924 there was a subsequent back reference to those brackets. This meant that, for
3925 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3926 possible to check for precisely this case. All we can do is abandon the
3927 optimization if .* occurs inside capturing brackets when there are any back
3928 references whatsoever. (See below for a better fix that came later.)
3929
3930 26. The handling of the optimization for finding the first character of a
3931 non-anchored pattern, and for finding a character that is required later in the
3932 match were failing in some cases. This didn't break the matching; it just
3933 failed to optimize when it could. The way this is done has been re-implemented.
3934
3935 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3936
3937 28. Added a new feature that provides some of the functionality that Perl
3938 provides with (?{...}). The facility is termed a "callout". The way it is done
3939 in PCRE is for the caller to provide an optional function, by setting
3940 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3941 global variable. By default it is unset, which disables all calling out. To get
3942 the function called, the regex must include (?C) at appropriate points. This
3943 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3944 This provides a means of identifying different callout points. When PCRE
3945 reaches such a point in the regex, if pcre_callout has been set, the external
3946 function is called. It is provided with data in a structure called
3947 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3948 matching continues; if it returns a non-zero value, the match at the current
3949 point fails. However, backtracking will occur if possible. [This was changed
3950 later and other features added - see item 49 below.]
3951
3952 29. pcretest is upgraded to test the callout functionality. It provides a
3953 callout function that displays information. By default, it shows the start of
3954 the match and the current position in the text. There are some new data escapes
3955 to vary what happens:
3956
3957 \C+ in addition, show current contents of captured substrings
3958 \C- do not supply a callout function
3959 \C!n return 1 when callout number n is reached
3960 \C!n!m return 1 when callout number n is reached for the mth time
3961
3962 30. If pcregrep was called with the -l option and just a single file name, it
3963 output "<stdin>" if a match was found, instead of the file name.
3964
3965 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3966 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3967 pcre_exec(). This saves a malloc/free per call. The default value of
3968 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3969 when configuring.
3970
3971 32. The default maximum size of a compiled pattern is 64K. There have been a
3972 few cases of people hitting this limit. The code now uses macros to handle the
3973 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3974 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3975 configuring. Tests 2 and 5 work only with 2-byte links because they output
3976 debugging information about compiled patterns.
3977
3978 33. Internal code re-arrangements:
3979
3980 (a) Moved the debugging function for printing out a compiled regex into
3981 its own source file (printint.c) and used #include to pull it into
3982 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3983 separate copies.
3984
3985 (b) Defined the list of op-code names for debugging as a macro in
3986 internal.h so that it is next to the definition of the opcodes.
3987
3988 (c) Defined a table of op-code lengths for simpler skipping along compiled
3989 code. This is again a macro in internal.h so that it is next to the
3990 definition of the opcodes.
3991
3992 34. Added support for recursive calls to individual subpatterns, along the
3993 lines of Robin Houston's patch (but implemented somewhat differently).
3994
3995 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3996 allow it to read and process whole directories in Win32. This code was
3997 contributed by Lionel Fourquaux; it has not been tested by me.
3998
3999 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4000 used to name a group. Names consist of alphanumerics and underscores, and must
4001 be unique. Back references use the syntax (?P=name) and recursive calls use
4002 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4003 numbers. The function pcre_fullinfo() can be used after compilation to extract
4004 a name/number map. There are three relevant calls:
4005
4006 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4007 PCRE_INFO_NAMECOUNT yields the number of entries
4008 PCRE_INFO_NAMETABLE yields a pointer to the map.
4009
4010 The map is a vector of fixed-size entries. The size of each entry depends on
4011 the length of the longest name used. The first two bytes of each entry are the
4012 group number, most significant byte first. There follows the corresponding
4013 name, zero terminated. The names are in alphabetical order.
4014
4015 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4016 case instead of 255. Making it the same both with and without UTF-8 support
4017 means that the same test output works with both.
4018
4019 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4020 calling malloc() with a zero argument.
4021
4022 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4023 optimization. I've improved things by keeping a bitmap of backreferences with
4024 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4025 fact referenced, the optimization can be applied. It is unlikely that a
4026 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4027 the match to follow \n) will appear inside brackets with a number greater than
4028 31, but if it does, any back reference > 31 suppresses the optimization.
4029
4030 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4031 of disabling numbered capturing parentheses. Any opening parenthesis that is
4032 not followed by ? behaves as if it were followed by ?: but named parentheses
4033 can still be used for capturing (and they will acquire numbers in the usual
4034 way).
4035
4036 41. Redesigned the return codes from the match() function into yes/no/error so
4037 that errors can be passed back from deep inside the nested calls. A malloc
4038 failure while inside a recursive subpattern call now causes the
4039 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4040
4041 42. It is now possible to set a limit on the number of times the match()
4042 function is called in a call to pcre_exec(). This facility makes it possible to
4043 limit the amount of recursion and backtracking, though not in a directly
4044 obvious way, because the match() function is used in a number of different
4045 circumstances. The count starts from zero for each position in the subject
4046 string (for non-anchored patterns). The default limit is, for compatibility, a
4047 large number, namely 10 000 000. You can change this in two ways:
4048
4049 (a) When configuring PCRE before making, you can use --with-match-limit=n
4050 to set a default value for the compiled library.
4051
4052 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4053 a different value is set. See 45 below.
4054
4055 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4056
4057 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4058 of things that can be changed at compile time. The first argument specifies
4059 what is wanted and the second points to where the information is to be placed.
4060 The current list of available information is:
4061
4062 PCRE_CONFIG_UTF8
4063
4064 The output is an integer that is set to one if UTF-8 support is available;
4065 otherwise it is set to zero.
4066
4067 PCRE_CONFIG_NEWLINE
4068
4069 The output is an integer that it set to the value of the code that is used for
4070 newline. It is either LF (10) or CR (13).
4071
4072 PCRE_CONFIG_LINK_SIZE
4073
4074 The output is an integer that contains the number of bytes used for internal
4075 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4076
4077 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4078
4079 The output is an integer that contains the threshold above which the POSIX
4080 interface uses malloc() for output vectors. See item 31 above.
4081
4082 PCRE_CONFIG_MATCH_LIMIT
4083
4084 The output is an unsigned integer that contains the default limit of the number
4085 of match() calls in a pcre_exec() execution. See 42 above.
4086
4087 44. pcretest has been upgraded by the addition of the -C option. This causes it
4088 to extract all the available output from the new pcre_config() function, and to
4089 output it. The program then exits immediately.
4090
4091 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4092 order to support additional features. One way would have been to define
4093 pcre_exec2() (for example) with extra arguments, but this would not have been
4094 extensible, and would also have required all calls to the original function to
4095 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4096 is used for passing in "extra" data from pcre_study().
4097
4098 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4099 contains the following fields:
4100
4101 flags a bitmap indicating which of the following fields are set
4102 study_data opaque data from pcre_study()
4103 match_limit a way of specifying a limit on match() calls for a specific
4104 call to pcre_exec()
4105 callout_data data for callouts (see 49 below)
4106
4107 The flag bits are also defined in pcre.h, and are
4108
4109 PCRE_EXTRA_STUDY_DATA
4110 PCRE_EXTRA_MATCH_LIMIT
4111 PCRE_EXTRA_CALLOUT_DATA
4112
4113 The pcre_study() function now returns one of these new pcre_extra blocks, with
4114 the actual study data pointed to by the study_data field, and the
4115 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4116 before. That is, this change is entirely upwards-compatible and requires no
4117 change to existing code.
4118
4119 If you want to pass in additional data to pcre_exec(), you can either place it
4120 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4121 block.
4122
4123 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4124 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4125 times with different match limits, until it finds the minimum value needed for
4126 pcre_exec() to complete. The value is then output. This can be instructive; for
4127 most simple matches the number is quite small, but for pathological cases it
4128 gets very large very quickly.
4129
4130 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4131 returns the size of the data block pointed to by the study_data field in a
4132 pcre_extra block, that is, the value that was passed as the argument to
4133 pcre_malloc() when PCRE was getting memory in which to place the information
4134 created by pcre_study(). The fourth argument should point to a size_t variable.
4135 pcretest has been extended so that this information is shown after a successful
4136 pcre_study() call when information about the compiled regex is being displayed.
4137
4138 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4139 because what follows is always an absolute path. (Later: it turns out that this
4140 is more than cosmetic for MinGW, because it doesn't like empty path
4141 components.)
4142
4143 49. Some changes have been made to the callout feature (see 28 above):
4144
4145 (i) A callout function now has three choices for what it returns:
4146
4147 0 => success, carry on matching
4148 > 0 => failure at this point, but backtrack if possible
4149 < 0 => serious error, return this value from pcre_exec()
4150
4151 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4152 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4153 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4154 use by callout functions. It will never be used by PCRE itself.
4155
4156 (ii) The pcre_extra structure (see 45 above) has a void * field called
4157 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4158 pcre_callout_block structure has a field of the same name. The contents of
4159 the field passed in the pcre_extra structure are passed to the callout
4160 function in the corresponding field in the callout block. This makes it
4161 easier to use the same callout-containing regex from multiple threads. For
4162 testing, the pcretest program has a new data escape
4163
4164 \C*n pass the number n (may be negative) as callout_data
4165
4166 If the callout function in pcretest receives a non-zero value as
4167 callout_data, it returns that value.
4168
4169 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4170 there were some redundant $(CFLAGS) in commands that are now specified as
4171 $(LINK), which already includes $(CFLAGS).
4172
4173 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4174 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4175 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4176 one-byte characters throughout. Note that case-insensitive matching applies
4177 only to characters whose values are less than 256. PCRE doesn't support the
4178 notion of cases for higher-valued characters.
4179
4180 (i) A character class whose characters are all within 0-255 is handled as
4181 a bit map, and the map is inverted for negative classes. Previously, a
4182 character > 255 always failed to match such a class; however it should
4183 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4184
4185 (ii) A negated character class with a single character < 255 is coded as
4186 "not this character" (OP_NOT). This wasn't working properly when the test
4187 character was multibyte, either singly or repeated.
4188
4189 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4190 mode, for example: \x{100}{2,3}.
4191
4192 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4193 singly or repeated) now correctly test multibyte characters. However,
4194 PCRE doesn't recognize any characters with values greater than 255 as
4195 digits, spaces, or word characters. Such characters always match \D, \S,
4196 and \W, and never match \d, \s, or \w.
4197
4198 (v) Classes may now contain characters and character ranges with values
4199 greater than 255. For example: [ab\x{100}-\x{400}].
4200
4201 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4202 PCRE in UTF-8 mode.
4203
4204 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4205 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4206 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4207 value.)
4208
4209 53. The single man page has become too large. I have therefore split it up into
4210 a number of separate man pages. These also give rise to individual HTML pages;
4211 these are now put in a separate directory, and there is an index.html page that
4212 lists them all. Some hyperlinking between the pages has been installed.
4213
4214 54. Added convenience functions for handling named capturing parentheses.
4215
4216 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4217 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4218 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4219 are faulted.
4220
4221 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4222 calling configure. These values are used when compiling the dftables.c program
4223 which is run to generate the source of the default character tables. They
4224 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4225 you will need to set these values.
4226
4227 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4228
4229
4230 Version 3.9 02-Jan-02
4231 ---------------------
4232
4233 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4234
4235 2. If --disable-static was given, the building process failed when trying to
4236 build pcretest and pcregrep. (For some reason it was using libtool to compile
4237 them, which is not right, as they aren't part of the library.)
4238
4239
4240 Version 3.8 18-Dec-01
4241 ---------------------
4242
4243 1. The experimental UTF-8 code was completely screwed up. It was packing the
4244 bytes in the wrong order. How dumb can you get?
4245
4246
4247 Version 3.7 29-Oct-01
4248 ---------------------
4249
4250 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4251 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4252 this didn't happen under Solaris 8, where I normally test things.
4253
4254 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4255 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4256 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4257 this fix an hour or so after the initial 3.7 release.)
4258
4259
4260 Version 3.6 23-Oct-01
4261 ---------------------
4262
4263 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4264 offsets passed as NULL with zero offset count.
4265
4266 2. The config.guess and config.sub files had not been updated when I moved to
4267 the latest autoconf.
4268
4269
4270 Version 3.5 15-Aug-01
4271 ---------------------
4272
4273 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4274 had been forgotten.
4275
4276 2. By using declared but undefined structures, we can avoid using "void"
4277 definitions in pcre.h while keeping the internal definitions of the structures
4278 private.
4279
4280 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4281 user point of view, this means that both static and shared libraries are built
4282 by default, but this can be individually controlled. More of the work of
4283 handling this static/shared cases is now inside libtool instead of PCRE's make
4284 file.
4285
4286 4. The pcretest utility is now installed along with pcregrep because it is
4287 useful for users (to test regexs) and by doing this, it automatically gets
4288 relinked by libtool. The documentation has been turned into a man page, so
4289 there are now .1, .txt, and .html versions in /doc.
4290
4291 5. Upgrades to pcregrep:
4292 (i) Added long-form option names like gnu grep.
4293 (ii) Added --help to list all options with an explanatory phrase.
4294 (iii) Added -r, --recursive to recurse into sub-directories.
4295 (iv) Added -f, --file to read patterns from a file.
4296
4297 6. pcre_exec() was referring to its "code" argument before testing that
4298 argument for NULL (and giving an error if it was NULL).
4299
4300 7. Upgraded Makefile.in to allow for compiling in a different directory from
4301 the source directory.
4302
4303 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4304 options bits, the pointer it was passed was to an int instead of to an unsigned
4305 long int. This mattered only on 64-bit systems.
4306
4307 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4308 generated) instead of pcre.in, which it its source. Also made the same change
4309 in several of the .c files.
4310
4311 10. A new release of gcc defines printf() as a macro, which broke pcretest
4312 because it had an ifdef in the middle of a string argument for printf(). Fixed
4313 by using separate calls to printf().
4314
4315 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4316 script, to force use of CR or LF instead of \n in the source. On non-Unix
4317 systems, the value can be set in config.h.
4318
4319 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4320 absolute limit. Changed the text of the error message to make this clear, and
4321 likewise updated the man page.
4322
4323 13. The limit of 99 on the number of capturing subpatterns has been removed.
4324 The new limit is 65535, which I hope will not be a "real" limit.
4325
4326
4327 Version 3.4 22-Aug-00
4328 ---------------------
4329
4330 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4331
4332 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4333
4334
4335 Version 3.3 01-Aug-00
4336 ---------------------
4337
4338 1. If an octal character was given, but the value was greater than \377, it
4339 was not getting masked to the least significant bits, as documented. This could
4340 lead to crashes in some systems.
4341
4342 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4343 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4344
4345 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4346 These just pass their arguments on to (pcre_free)(), but they are provided
4347 because some uses of PCRE bind it to non-C systems that can call its functions,
4348 but cannot call free() or pcre_free() directly.
4349
4350 4. Add "make test" as a synonym for "make check". Corrected some comments in
4351 the Makefile.
4352
4353 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4354 Makefile.
4355
4356 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4357 command called pgrep for grepping around the active processes.
4358
4359 7. Added the beginnings of support for UTF-8 character strings.
4360
4361 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4362 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4363 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4364 out for the ar command.)
4365
4366
4367 Version 3.2 12-May-00
4368 ---------------------
4369
4370 This is purely a bug fixing release.
4371
4372 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4373 of ZA. This was just one example of several cases that could provoke this bug,
4374 which was introduced by change 9 of version 2.00. The code for breaking
4375 infinite loops after an iteration that matches an empty string was't working
4376 correctly.
4377
4378 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4379 when matched against abbab (for example). After matching an empty string, it
4380 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4381 caused it to match further down the string than it should.
4382
4383 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4384 was there because it doesn't seem to be needed, and it causes trouble on some
4385 systems, as it is not a Standard C header. It has been removed.
4386
4387 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4388 were reported on the Macintosh. The changes were from
4389
4390 while ((c = *(++ptr)) != 0 && c != '\n');
4391 to
4392 while ((c = *(++ptr)) != 0 && c != '\n') ;
4393
4394 Totally extraordinary, but if that's what it takes...
4395
4396 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4397 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4398 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4399 assumes the way PCRE uses memmove() (always moving upwards).
4400
4401 6. PCRE is being used in one environment where strchr() is not available. There
4402 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4403 faster code anyway.
4404
4405
4406 Version 3.1 09-Feb-00
4407 ---------------------
4408
4409 The only change in this release is the fixing of some bugs in Makefile.in for
4410 the "install" target:
4411
4412 (1) It was failing to install pcreposix.h.
4413
4414 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4415
4416
4417 Version 3.0 01-Feb-00
4418 ---------------------
4419
4420 1. Add support for the /+ modifier to perltest (to output $` like it does in
4421 pcretest).
4422
4423 2. Add support for the /g modifier to perltest.
4424
4425 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4426 matches null strings.
4427
4428 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4429 pattern. Perl treats empty patterns specially - it reuses the most recent
4430 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4431 effect.
4432
4433 5. The POSIX interface was broken in that it was just handing over the POSIX
4434 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4435 required a bigger vector, with some working space on the end. This means that
4436 the POSIX wrapper now has to get and free some memory, and copy the results.
4437
4438 6. Added some simple autoconf support, placing the test data and the
4439 documentation in separate directories, re-organizing some of the
4440 information files, and making it build pcre-config (a GNU standard). Also added
4441 libtool support for building PCRE as a shared library, which is now the
4442 default.
4443
4444 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4445 09 are not valid octal constants. Single digits will be used for minor values
4446 less than 10.
4447
4448 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4449 existing programs that set these in the POSIX interface can use PCRE without
4450 modification.
4451
4452 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4453 return all that pcre_info() returns, plus additional data. The pcre_info()
4454 function is retained for compatibility, but is considered to be obsolete.
4455
4456 10. Added experimental recursion feature (?R) to handle one common case that
4457 Perl 5.6 will be able to do with (?p{...}).
4458
4459 11. Added support for POSIX character classes like [:alpha:], which Perl is
4460 adopting.
4461
4462
4463 Version 2.08 31-Aug-99
4464 ----------------------
4465
4466 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4467 trying to match at the startoffset position, but instead was moving forward to
4468 the next newline as if a previous match had failed.
4469
4470 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4471 and could get into a loop if a null string was matched other than at the start
4472 of the subject.
4473
4474 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4475 be distinguished at compile time, and for completeness also added PCRE_DATE.
4476
4477 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4478 in GnuWin32 environments.
4479
4480
4481 Version 2.07 29-Jul-99
4482 ----------------------
4483
4484 1. The documentation is now supplied in plain text form and HTML as well as in
4485 the form of man page sources.
4486
4487 2. C++ compilers don't like assigning (void *) values to other pointer types.
4488 In particular this affects malloc(). Although there is no problem in Standard
4489 C, I've put in casts to keep C++ compilers happy.
4490
4491 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4492 should be (const char *).
4493
4494 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4495 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4496 However, I haven't made this a standard facility. The documentation doesn't
4497 mention it, and the Makefile doesn't support it.
4498
4499 5. The Makefile now contains an "install" target, with editable destinations at
4500 the top of the file. The pcretest program is not installed.
4501
4502 6. pgrep -V now gives the PCRE version number and date.
4503
4504 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4505 causing the entire string to be ignored, instead of just the last character.
4506
4507 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4508 non-matching string, it can take a very, very long time, even for strings of
4509 quite modest length, because of the nested recursion. PCRE now does better in
4510 some of these cases. It does this by remembering the last required literal
4511 character in the pattern, and pre-searching the subject to ensure it is present
4512 before running the real match. In other words, it applies a heuristic to detect
4513 some types of certain failure quickly, and in the above example, if presented
4514 with a string that has no trailing " it gives "no match" very quickly.
4515
4516 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4517 other alternatives are tried instead.
4518
4519
4520 Version 2.06 09-Jun-99
4521 ----------------------
4522
4523 1. Change pcretest's output for amount of store used to show just the code
4524 space, because the remainder (the data block) varies in size between 32-bit and
4525 64-bit systems.
4526
4527 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4528 start matching at. This allows lookbehinds to work when searching for multiple
4529 occurrences in a string.
4530
4531 3. Added additional options to pcretest for testing multiple occurrences:
4532
4533 /+ outputs the rest of the string that follows a match
4534 /g loops for multiple occurrences, using the new startoffset argument
4535 /G loops for multiple occurrences by passing an incremented pointer
4536
4537 4. PCRE wasn't doing the "first character" optimization for patterns starting
4538 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4539 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4540 the letter 'x'. On long subject strings, this gives a significant speed-up.
4541
4542
4543 Version 2.05 21-Apr-99
4544 ----------------------
4545
4546 1. Changed the type of magic_number from int to long int so that it works
4547 properly on 16-bit systems.
4548
4549 2. Fixed a bug which caused patterns starting with .* not to work correctly
4550 when the subject string contained newline characters. PCRE was assuming
4551 anchoring for such patterns in all cases, which is not correct because .* will
4552 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4553 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4554 must be retried after every newline in the subject.
4555
4556
4557 Version 2.04 18-Feb-99
4558 ----------------------
4559
4560 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4561 computation of the store needed to hold the pattern was incorrect (too large).
4562 If such patterns were nested a few deep, this could multiply and become a real
4563 problem.
4564
4565 2. Added /M option to pcretest to show the memory requirement of a specific
4566 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4567
4568 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4569 compiled in such a way that the backtracking after subsequent failure was
4570 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4571 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4572
4573
4574 Version 2.03 02-Feb-99
4575 ----------------------
4576
4577 1. Fixed typo and small mistake in man page.
4578
4579 2. Added 4th condition (GPL supersedes if conflict) and created separate
4580 LICENCE file containing the conditions.
4581
4582 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4583 Perl, that is the internal \ allows the delimiter to be included in the
4584 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4585 the final delimiter, add \ to the end of the pattern (to test the error).
4586
4587 4. Added the convenience functions for extracting substrings after a successful
4588 match. Updated pcretest to make it able to test these functions.
4589
4590
4591 Version 2.02 14-Jan-99
4592 ----------------------
4593
4594 1. Initialized the working variables associated with each extraction so that
4595 their saving and restoring doesn't refer to uninitialized store.
4596
4597 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4598 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4599 fix the problem.
4600
4601 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4602 calls, and wasn't printing the correct value for compiling calls. Increased the
4603 default value of LOOPREPEAT, and the number of significant figures in the
4604 times.
4605
4606 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4607
4608 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4609 a building problem on Windows NT with a FAT file system.
4610
4611
4612 Version 2.01 21-Oct-98
4613 ----------------------
4614
4615 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4616 to character tables built by pcre_maketables() in the current locale. If NULL
4617 is passed, the default tables are used.
4618
4619
4620 Version 2.00 24-Sep-98
4621 ----------------------
4622
4623 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4624 it any more.
4625
4626 2. Allow quantification of (?>) groups, and make it work correctly.
4627
4628 3. The first character computation wasn't working for (?>) groups.
4629
4630 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4631 end of the subject) and add 5.005's \z, which really does match only at the
4632 very end of the subject.
4633
4634 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4635
4636 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4637 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4638 localized options. All options to pcre_study() were also removed.
4639
4640 7. Add other new features from 5.005:
4641
4642 $(?<= positive lookbehind
4643 $(?<! negative lookbehind
4644 (?imsx-imsx) added the unsetting capability
4645 such a setting is global if at outer level; local otherwise
4646 (?imsx-imsx:) non-capturing groups with option setting
4647 (?(cond)re|re) conditional pattern matching
4648
4649 A backreference to itself in a repeated group matches the previous
4650 captured string.
4651
4652 8. General tidying up of studying (both automatic and via "study")
4653 consequential on the addition of new assertions.
4654
4655 9. As in 5.005, unlimited repeated groups that could match an empty substring
4656 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4657 runtime if any iteration does actually match an empty substring.
4658
4659 10. Include the RunTest script in the distribution.
4660
4661 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4662 discrepancies, some of which were old and were also with respect to 5.004. They
4663 have now been fixed.
4664
4665
4666 Version 1.09 28-Apr-98
4667 ----------------------
4668
4669 1. A negated single character class followed by a quantifier with a minimum
4670 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4671 program crashes, or just wrong answers. This did not apply to negated classes
4672 containing more than one character, or to minima other than one.
4673
4674
4675 Version 1.08 27-Mar-98
4676 ----------------------
4677
4678 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4679
4680 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4681 latter must appear before anything that relies on it in the pattern.
4682
4683
4684 Version 1.07 16-Feb-98
4685 ----------------------
4686
4687 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4688 repeat of a potentially empty string).
4689
4690
4691 Version 1.06 23-Jan-98
4692 ----------------------
4693
4694 1. Added Markus Oberhumer's little patches for C++.
4695
4696 2. Literal strings longer than 255 characters were broken.
4697
4698
4699 Version 1.05 23-Dec-97
4700 ----------------------
4701
4702 1. Negated character classes containing more than one character were failing if
4703 PCRE_CASELESS was set at run time.
4704
4705
4706 Version 1.04 19-Dec-97
4707 ----------------------
4708
4709 1. Corrected the man page, where some "const" qualifiers had been omitted.
4710
4711 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4712 input syntax.
4713
4714 3. Fixed memory leak which occurred when a regex with back references was
4715 matched with an offsets vector that wasn't big enough. The temporary memory
4716 that is used in this case wasn't being freed if the match failed.
4717
4718 4. Tidied pcretest to ensure it frees memory that it gets.
4719
4720 5. Temporary memory was being obtained in the case where the passed offsets
4721 vector was exactly big enough.
4722
4723 6. Corrected definition of offsetof() from change 5 below.
4724
4725 7. I had screwed up change 6 below and broken the rules for the use of
4726 setjmp(). Now fixed.
4727
4728
4729 Version 1.03 18-Dec-97
4730 ----------------------
4731
4732 1. A erroneous regex with a missing opening parenthesis was correctly
4733 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4734 on some systems.
4735
4736 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4737 it was reported that one broken compiler failed on the former because "code" is
4738 also an independent variable.
4739
4740 3. The erroneous regex a[]b caused an array overrun reference.
4741
4742 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4743 fail on data ending with that character. (It was going on too far, and checking
4744 the next character, typically a binary zero.) This was specific to the
4745 optimized code for single-character negative classes.
4746
4747 5. Added a contributed patch from the TIN world which does the following:
4748
4749 + Add an undef for memmove, in case the the system defines a macro for it.
4750
4751 + Add a definition of offsetof(), in case there isn't one. (I don't know
4752 the reason behind this - offsetof() is part of the ANSI standard - but
4753 it does no harm).
4754
4755 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4756 most of the places where whitespace preceded '#'. I have given up and
4757 allowed the remaining 2 cases to be at the margin.
4758
4759 + Rename some variables in pcre to eliminate shadowing. This seems very
4760 pedantic, but does no harm, of course.
4761
4762 6. Moved the call to setjmp() into its own function, to get rid of warnings
4763 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4764
4765 7. Constructs such as \d{8,} were compiling into the equivalent of
4766 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4767 outcome, but in this particular case used more store than had been allocated,
4768 which caused the bug to be discovered because it threw up an internal error.
4769
4770 8. The debugging code in both pcre and pcretest for outputting the compiled
4771 form of a regex was going wrong in the case of back references followed by
4772 curly-bracketed repeats.
4773
4774
4775 Version 1.02 12-Dec-97
4776 ----------------------
4777
4778 1. Typos in pcre.3 and comments in the source fixed.
4779
4780 2. Applied a contributed patch to get rid of places where it used to remove
4781 'const' from variables, and fixed some signed/unsigned and uninitialized
4782 variable warnings.
4783
4784 3. Added the "runtest" target to Makefile.
4785
4786 4. Set default compiler flag to -O2 rather than just -O.
4787
4788
4789 Version 1.01 19-Nov-97
4790 ----------------------
4791
4792 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4793 like /([ab]*)*/, that is, for classes with more than one character in them.
4794
4795 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4796 as /((?>a*))*/ (a PCRE_EXTRA facility).
4797
4798
4799 Version 1.00 18-Nov-97
4800 ----------------------
4801
4802 1. Added compile-time macros to support systems such as SunOS4 which don't have
4803 memmove() or strerror() but have other things that can be used instead.
4804
4805 2. Arranged that "make clean" removes the executables.
4806
4807
4808 Version 0.99 27-Oct-97
4809 ----------------------
4810
4811 1. Fixed bug in code for optimizing classes with only one character. It was
4812 initializing a 32-byte map regardless, which could cause it to run off the end
4813 of the memory it had got.
4814
4815 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4816
4817
4818 Version 0.98 22-Oct-97
4819 ----------------------
4820
4821 1. Fixed bug in code for handling temporary memory usage when there are more
4822 back references than supplied space in the ovector. This could cause segfaults.
4823
4824
4825 Version 0.97 21-Oct-97
4826 ----------------------
4827
4828 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4829
4830 2. Optimized negated single characters not to use a bit map.
4831
4832 3. Brought error texts together as macro definitions; clarified some of them;
4833 fixed one that was wrong - it said "range out of order" when it meant "invalid
4834 escape sequence".
4835
4836 4. Changed some char * arguments to const char *.
4837
4838 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4839
4840 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4841 pcretest.
4842
4843
4844 Version 0.96 16-Oct-97
4845 ----------------------
4846
4847 1. Added a simple "pgrep" utility to the distribution.
4848
4849 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4850 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4851 where "ddd" means "one or more decimal digits".
4852
4853 3. Fixed serious bug. If a pattern had a back reference, but the call to
4854 pcre_exec() didn't supply a large enough ovector to record the related
4855 identifying subpattern, the match always failed. PCRE now remembers the number
4856 of the largest back reference, and gets some temporary memory in which to save
4857 the offsets during matching if necessary, in order to ensure that
4858 backreferences always work.
4859
4860 4. Increased the compatibility with Perl in a number of ways:
4861
4862 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4863 to request this handling. The option can be set at compile or exec time.
4864
4865 (b) $ matches before a terminating newline by default; an option
4866 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4867 mode). The option can be set at compile or exec time.
4868
4869 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4870 the same as Perl's. If the decimal number it represents is less than 10
4871 or there aren't that many previous left capturing parentheses, an octal
4872 escape is read. Inside a character class, it's always an octal escape,
4873 even if it is a single digit.
4874
4875 (d) An escaped but undefined alphabetic character is taken as a literal,
4876 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4877 escapes.
4878
4879 (e) {0} is now permitted. (The previous item is removed from the compiled
4880 pattern).
4881
4882 5. Changed all the names of code files so that the basic parts are no longer
4883 than 10 characters, and abolished the teeny "globals.c" file.
4884
4885 6. Changed the handling of character classes; they are now done with a 32-byte
4886 bit map always.
4887
4888 7. Added the -d and /D options to pcretest to make it possible to look at the
4889 internals of compilation without having to recompile pcre.
4890
4891
4892 Version 0.95 23-Sep-97
4893 ----------------------
4894
4895 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4896 \x20 at the start of a run of normal characters. These were being treated as
4897 real characters, instead of the source characters being re-checked.
4898
4899
4900 Version 0.94 18-Sep-97
4901 ----------------------
4902
4903 1. The functions are now thread-safe, with the caveat that the global variables
4904 containing pointers to malloc() and free() or alternative functions are the
4905 same for all threads.
4906
4907 2. Get pcre_study() to generate a bitmap of initial characters for non-
4908 anchored patterns when this is possible, and use it if passed to pcre_exec().
4909
4910
4911 Version 0.93 15-Sep-97
4912 ----------------------
4913
4914 1. /(b)|(:+)/ was computing an incorrect first character.
4915
4916 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4917 but not actually doing anything yet.
4918
4919 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4920 as Perl does (e.g. [-az] or [az-]).
4921
4922 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4923 all possible positions.
4924
4925 5. Split up into different modules to avoid including unneeded functions in a
4926 compiled binary. However, compile and exec are still in one module. The "study"
4927 function is split off.
4928
4929 6. The character tables are now in a separate module whose source is generated
4930 by an auxiliary program - but can then be edited by hand if required. There are
4931 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4932 toupper() in the code.
4933
4934 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4935 make them global. Abolish the function for setting them, as the caller can now
4936 set them directly.
4937
4938
4939 Version 0.92 11-Sep-97
4940 ----------------------
4941
4942 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4943 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4944
4945 2. Caseless matching was not working in character classes if the characters in
4946 the pattern were in upper case.
4947
4948 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4949
4950 4. Make PCRE_ANCHORED public and accept as a compile option.
4951
4952 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4953 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4954 pass them.
4955
4956 6. Give an error if bad option bits passed at compile or run time.
4957
4958 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4959 pcretest to cause it to pass that flag.
4960
4961 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4962 options, and the first character, if set.
4963
4964 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4965
4966
4967 Version 0.91 10-Sep-97
4968 ----------------------
4969
4970 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4971 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4972
4973 2. PCRE was looping on encountering an indefinitely repeated back reference to
4974 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4975 Perl does - treats the match as successful.
4976
4977 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5