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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1235 - (show annotations)
Mon Dec 10 13:36:29 2012 UTC (6 years, 11 months ago) by ph10
File size: 240460 byte(s)
Ensure boolean macros are #undeffed in pcre.h.generic.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.33 xx-xxxx-201x
5 -------------------------
6
7 1. Added 'U' to some constants that are compared to unsigned integers, to
8 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9 variables that are added to signed variables, to ensure the result is
10 signed and can be negated.
11
12 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
13 CMake config files.
14
15 3. Revise the creation of config.h.generic so that all boolean macros are
16 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17 overriding via -D on the command line possible.
18
19
20 Version 8.32 30-November-2012
21 -----------------------------
22
23 1. Improved JIT compiler optimizations for first character search and single
24 character iterators.
25
26 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
27 Patch by Daniel Richard G.
28
29 3. Single character iterator optimizations in the JIT compiler.
30
31 4. Improved JIT compiler optimizations for character ranges.
32
33 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
34 Reported by Giuseppe D'Angelo.
35
36 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
37 of a line, was being set incorrectly in cases where .* appeared inside
38 atomic brackets at the start of a pattern, or where there was a subsequent
39 *PRUNE or *SKIP.
40
41 7. Improved instruction cache flush for POWER/PowerPC.
42 Patch by Daniel Richard G.
43
44 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
45 grep:
46
47 (a) There is now no limit to the number of patterns to be matched.
48
49 (b) An error is given if a pattern is too long.
50
51 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
52 are now supported.
53
54 (d) --exclude-from and --include-from (multiple use) have been added.
55
56 (e) Exclusions and inclusions now apply to all files and directories, not
57 just to those obtained from scanning a directory recursively.
58
59 (f) Multiple uses of -f and --file-list are now supported.
60
61 (g) In a Windows environment, the default for -d has been changed from
62 "read" (the GNU grep default) to "skip", because otherwise the presence
63 of a directory in the file list provokes an error.
64
65 (h) The documentation has been revised and clarified in places.
66
67 9. Improve the matching speed of capturing brackets.
68
69 10. Changed the meaning of \X so that it now matches a Unicode extended
70 grapheme cluster.
71
72 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
73 out POSIX threads when JIT support is configured.
74
75 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
76
77 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
78 structure could go wrong in environments where size_t is not the same size
79 as int.
80
81 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
82 set.
83
84 15. The EBCDIC support had decayed; later updates to the code had included
85 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
86 general tidy up of EBCDIC-related issues, and the documentation was also
87 not quite right. There is now a test that can be run on ASCII systems to
88 check some of the EBCDIC-related things (but is it not a full test).
89
90 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
91 in a small tidy to the code.
92
93 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
94
95 18. If the --only-matching (-o) option in pcregrep is specified multiple
96 times, each one causes appropriate output. For example, -o1 -o2 outputs the
97 substrings matched by the 1st and 2nd capturing parentheses. A separating
98 string can be specified by --om-separator (default empty).
99
100 19. Improving the first n character searches.
101
102 20. Turn case lists for horizontal and vertical white space into macros so that
103 they are defined only once.
104
105 21. This set of changes together give more compatible Unicode case-folding
106 behaviour for characters that have more than one other case when UCP
107 support is available.
108
109 (a) The Unicode property table now has offsets into a new table of sets of
110 three or more characters that are case-equivalent. The MultiStage2.py
111 script that generates these tables (the pcre_ucd.c file) now scans
112 CaseFolding.txt instead of UnicodeData.txt for character case
113 information.
114
115 (b) The code for adding characters or ranges of characters to a character
116 class has been abstracted into a generalized function that also handles
117 case-independence. In UTF-mode with UCP support, this uses the new data
118 to handle characters with more than one other case.
119
120 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
121 whose other case is greater than 256 are now correctly matched
122 caselessly. Previously, the high codepoint matched the low one, but not
123 vice versa.
124
125 (d) The processing of \h, \H, \v, and \ in character classes now makes use
126 of the new class addition function, using character lists defined as
127 macros alongside the case definitions of 20 above.
128
129 (e) Caseless back references now work with characters that have more than
130 one other case.
131
132 (f) General caseless matching of characters with more than one other case
133 is supported.
134
135 22. Unicode character properties were updated from Unicode 6.2.0
136
137 23. Improved CMake support under Windows. Patch by Daniel Richard G.
138
139 24. Add support for 32-bit character strings, and UTF-32
140
141 25. Major JIT compiler update (code refactoring and bugfixing).
142 Experimental Sparc 32 support is added.
143
144 26. Applied a modified version of Daniel Richard G's patch to create
145 pcre.h.generic and config.h.generic by "make" instead of in the
146 PrepareRelease script.
147
148 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
149 pcre_compile.c when checking for a zero character.
150
151 28. Introducing a native interface for JIT. Through this interface, the compiled
152 machine code can be directly executed. The purpose of this interface is to
153 provide fast pattern matching, so several sanity checks are not performed.
154 However, feature tests are still performed. The new interface provides
155 1.4x speedup compared to the old one.
156
157 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
158 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
159 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
160
161 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
162 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
163 the "old" RFC 2279). Instead, it ended up passing a negative length to
164 pcre_exec().
165
166 31. Add support for GCC's visibility feature to hide internal functions.
167
168 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
169 "unknown -C option" after outputting 0 or 1.
170
171 33. There is now support for generating a code coverage report for the test
172 suite in environments where gcc is the compiler and lcov is installed. This
173 is mainly for the benefit of the developers.
174
175 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
176 unaddressable using valgrind annotations, allowing valgrind to detect
177 invalid memory accesses. This is mainly for the benefit of the developers.
178
179 25. (*UTF) can now be used to start a pattern in any of the three libraries.
180
181 26. Give configure error if --enable-cpp but no C++ compiler found.
182
183
184 Version 8.31 06-July-2012
185 -------------------------
186
187 1. Fixing a wrong JIT test case and some compiler warnings.
188
189 2. Removed a bashism from the RunTest script.
190
191 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
192 to unsigned type, result still unsigned" that was given by an MS compiler
193 on encountering the code "-sizeof(xxx)".
194
195 4. Partial matching support is added to the JIT compiler.
196
197 5. Fixed several bugs concerned with partial matching of items that consist
198 of more than one character:
199
200 (a) /^(..)\1/ did not partially match "aba" because checking references was
201 done on an "all or nothing" basis. This also applied to repeated
202 references.
203
204 (b) \R did not give a hard partial match if \r was found at the end of the
205 subject.
206
207 (c) \X did not give a hard partial match after matching one or more
208 characters at the end of the subject.
209
210 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
211 a partial match for the string "\r".
212
213 (e) When newline was set to CRLF, the metacharacter "." did not recognize
214 a partial match for a CR character at the end of the subject string.
215
216 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
217 running pcretest, the text "(JIT)" added to the output whenever JIT is
218 actually used to run the match.
219
220 7. Individual JIT compile options can be set in pcretest by following -s+[+]
221 or /S+[+] with a digit between 1 and 7.
222
223 8. OP_NOT now supports any UTF character not just single-byte ones.
224
225 9. (*MARK) control verb is now supported by the JIT compiler.
226
227 10. The command "./RunTest list" lists the available tests without actually
228 running any of them. (Because I keep forgetting what they all are.)
229
230 11. Add PCRE_INFO_MAXLOOKBEHIND.
231
232 12. Applied a (slightly modified) user-supplied patch that improves performance
233 when the heap is used for recursion (compiled with --disable-stack-for-
234 recursion). Instead of malloc and free for each heap frame each time a
235 logical recursion happens, frames are retained on a chain and re-used where
236 possible. This sometimes gives as much as 30% improvement.
237
238 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
239 call.
240
241 14. As documented, (*COMMIT) is now confined to within a positive assertion.
242
243 15. It is now possible to link pcretest with libedit as an alternative to
244 libreadline.
245
246 16. (*COMMIT) control verb is now supported by the JIT compiler.
247
248 17. The Unicode data tables have been updated to Unicode 6.1.0.
249
250 18. Added --file-list option to pcregrep.
251
252 19. Added binary file support to pcregrep, including the -a, --binary-files,
253 -I, and --text options.
254
255 20. The madvise function is renamed for posix_madvise for QNX compatibility
256 reasons. Fixed by Giuseppe D'Angelo.
257
258 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
259
260 22. Changed the option for creating symbolic links for 16-bit man pages from
261 -s to -sf so that re-installing does not cause issues.
262
263 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
264
265 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
266 matches in certain environments (the workspace was not being correctly
267 retained). Also added to pcre_dfa_exec() a simple plausibility check on
268 some of the workspace data at the beginning of a restart.
269
270 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
271 was not doing so when it should - probably a typo introduced by SVN 528
272 (change 8.10/14).
273
274 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
275 \w+ when the character tables indicated that \x{c4} was a word character.
276 There were several related cases, all because the tests for doing a table
277 lookup were testing for characters less than 127 instead of 255.
278
279 27. If a pattern contains capturing parentheses that are not used in a match,
280 their slots in the ovector are set to -1. For those that are higher than
281 any matched groups, this happens at the end of processing. In the case when
282 there were back references that the ovector was too small to contain
283 (causing temporary malloc'd memory to be used during matching), and the
284 highest capturing number was not used, memory off the end of the ovector
285 was incorrectly being set to -1. (It was using the size of the temporary
286 memory instead of the true size.)
287
288 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
289 ovector size, it uses memory at the end of the block that it has got.
290
291 29. Check for an overlong MARK name and give an error at compile time. The
292 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
293
294 30. JIT compiler update.
295
296 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
297 Rill for the patch.
298
299 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
300
301 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
302
303 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
304 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
305 was enabled.
306
307 35. Improve JIT code generation for greedy plus quantifier.
308
309 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
310 1 to "aa" instead of to an empty string. The bug affected repeated groups
311 that could potentially match an empty string.
312
313 37. Optimizing single character iterators in JIT.
314
315 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
316 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
317 codepoints that are too big for the mode are faulted, and in a UTF mode,
318 disallowed codepoints are also faulted.
319
320 39. If PCRE was compiled with UTF support, in three places in the DFA
321 matcher there was code that should only have been obeyed in UTF mode, but
322 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
323 processing when bytes with values greater than 127 were present. In 16-bit
324 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
325 both cases the values are those that cannot be the first data item in a UTF
326 character. The three items that might have provoked this were recursions,
327 possessively repeated groups, and atomic groups.
328
329 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
330 and pcregrep, because some OS require shared objects to be explicitly
331 passed to ld, causing the link step to fail if they are not.
332
333 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
334 mode, patterns that started with \h* or \R* might be incorrectly matched.
335
336
337 Version 8.30 04-February-2012
338 -----------------------------
339
340 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
341 name is defined in ctype.h.
342
343 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
344 only in quite long subpatterns.
345
346 3. Removed the function pcre_info(), which has been obsolete and deprecated
347 since it was replaced by pcre_fullinfo() in February 2000.
348
349 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
350 match a (*MARK), and the match failed at the start of the subject, a
351 reference to memory before the start of the subject could occur. This bug
352 was introduced by fix 17 of release 8.21.
353
354 5. A reference to an unset group with zero minimum repetition was giving
355 totally wrong answers (in non-JavaScript-compatibility mode). For example,
356 /(another)?(\1?)test/ matched against "hello world test". This bug was
357 introduced in release 8.13.
358
359 6. Add support for 16-bit character strings (a large amount of work involving
360 many changes and refactorings).
361
362 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
363 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
364 from a file.
365
366 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
367 rounding is not applied in this particular case).
368
369 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
370 if they appear, or are escaped, in patterns.
371
372 10. Get rid of a number of -Wunused-but-set-variable warnings.
373
374 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
375 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
376 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
377 also returns the mark "x". This bug applied to capturing parentheses,
378 non-capturing parentheses, and atomic parentheses. It also applied to some
379 assertions.
380
381 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
382 information out of configure.ac instead of relying on pcre.h.generic, which
383 is not stored in the repository.
384
385 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
386 -lreadline.
387
388 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
389
390 15. Applied Graycode's patch to put the top-level frame on the stack rather
391 than the heap when not using the stack for recursion. This gives a
392 performance improvement in many cases when recursion is not deep.
393
394 16. Experimental code added to "pcretest -C" to output the stack frame size.
395
396
397 Version 8.21 12-Dec-2011
398 ------------------------
399
400 1. Updating the JIT compiler.
401
402 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
403 are added as well.
404
405 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
406 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
407 calling _pcre_jit_exec. Some extra comments are added.
408
409 4. (*MARK) settings inside atomic groups that do not contain any capturing
410 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
411 was introduced by change 18 for 8.20.
412
413 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
414 ECMA-262 standard.
415
416 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
417 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
418 This bug was probably introduced by change 9 of 8.13.
419
420 7. While fixing 6 above, I noticed that a number of other items were being
421 incorrectly rejected as "not fixed length". This arose partly because newer
422 opcodes had not been added to the fixed-length checking code. I have (a)
423 corrected the bug and added tests for these items, and (b) arranged for an
424 error to occur if an unknown opcode is encountered while checking for fixed
425 length instead of just assuming "not fixed length". The items that were
426 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
427 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
428 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
429
430 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
431 being incorrectly compiled and would have given unpredicatble results.
432
433 9. A possessively repeated subpattern with minimum repeat count greater than
434 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
435 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
436 the first (A) could occur when it should not.
437
438 10. Add a cast and remove a redundant test from the code.
439
440 11. JIT should use pcre_malloc/pcre_free for allocation.
441
442 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
443 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
444 is anything other than /usr, -L is still shown).
445
446 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
447
448 14. Perl does not support \N without a following name in a [] class; PCRE now
449 also gives an error.
450
451 15. If a forward reference was repeated with an upper limit of around 2000,
452 it caused the error "internal error: overran compiling workspace". The
453 maximum number of forward references (including repeats) was limited by the
454 internal workspace, and dependent on the LINK_SIZE. The code has been
455 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
456 the default depends on LINK_SIZE. There is a new upper limit (for safety)
457 of around 200,000 forward references. While doing this, I also speeded up
458 the filling in of repeated forward references.
459
460 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
461 incorrectly expecting the subject to contain another "a" after the start.
462
463 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
464 in the match, the SKIP should be ignored. This was not happening; instead
465 the SKIP was being treated as NOMATCH. For patterns such as
466 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
467 tested.
468
469 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
470 now much more compatible with Perl, in particular in cases where the result
471 is a non-match for a non-anchored pattern. For example, if
472 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
473 "m", where previously it did not return a name. A side effect of this
474 change is that for partial matches, the last encountered mark name is
475 returned, as for non matches. A number of tests that were previously not
476 Perl-compatible have been moved into the Perl-compatible test files. The
477 refactoring has had the pleasing side effect of removing one argument from
478 the match() function, thus reducing its stack requirements.
479
480 19. If the /S+ option was used in pcretest to study a pattern using JIT,
481 subsequent uses of /S (without +) incorrectly behaved like /S+.
482
483 21. Retrieve executable code size support for the JIT compiler and fixing
484 some warnings.
485
486 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
487 not work when the shorter character appeared right at the end of the
488 subject string.
489
490 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
491 systems.
492
493 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
494 output it when the /M option is used in pcretest.
495
496 25. The CheckMan script was not being included in the distribution. Also, added
497 an explicit "perl" to run Perl scripts from the PrepareRelease script
498 because this is reportedly needed in Windows.
499
500 26. If study data was being save in a file and studying had not found a set of
501 "starts with" bytes for the pattern, the data written to the file (though
502 never used) was taken from uninitialized memory and so caused valgrind to
503 complain.
504
505 27. Updated RunTest.bat as provided by Sheri Pierce.
506
507 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
508
509 29. Computation of memory usage for the table of capturing group names was
510 giving an unnecessarily large value.
511
512
513 Version 8.20 21-Oct-2011
514 ------------------------
515
516 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
517 a POSIX class. After further experiments with Perl, which convinced me that
518 Perl has bugs and confusions, a closing square bracket is no longer allowed
519 in a POSIX name. This bug also affected patterns with classes that started
520 with full stops.
521
522 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
523 captured substring, but while checking the failing first alternative,
524 substring 1 is temporarily captured. If the output vector supplied to
525 pcre_exec() was not big enough for this capture, the yield of the function
526 was still zero ("insufficient space for captured substrings"). This cannot
527 be totally fixed without adding another stack variable, which seems a lot
528 of expense for a edge case. However, I have improved the situation in cases
529 such as /(a)(b)x|abc/ matched against "abc", where the return code
530 indicates that fewer than the maximum number of slots in the ovector have
531 been set.
532
533 3. Related to (2) above: when there are more back references in a pattern than
534 slots in the output vector, pcre_exec() uses temporary memory during
535 matching, and copies in the captures as far as possible afterwards. It was
536 using the entire output vector, but this conflicts with the specification
537 that only 2/3 is used for passing back captured substrings. Now it uses
538 only the first 2/3, for compatibility. This is, of course, another edge
539 case.
540
541 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
542 main code base, and can be used by building with --enable-jit. When this is
543 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
544 runtime --no-jit option is given.
545
546 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
547 ovector, the return from the function was zero, implying that there were
548 other matches that did not fit. The correct "exactly full" value is now
549 returned.
550
551 6. If a subpattern that was called recursively or as a subroutine contained
552 (*PRUNE) or any other control that caused it to give a non-standard return,
553 invalid errors such as "Error -26 (nested recursion at the same subject
554 position)" or even infinite loops could occur.
555
556 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
557 computing the minimum length on reaching *ACCEPT, and so ended up with the
558 wrong value of 1 rather than 0. Further investigation indicates that
559 computing a minimum subject length in the presence of *ACCEPT is difficult
560 (think back references, subroutine calls), and so I have changed the code
561 so that no minimum is registered for a pattern that contains *ACCEPT.
562
563 8. If (*THEN) was present in the first (true) branch of a conditional group,
564 it was not handled as intended. [But see 16 below.]
565
566 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
567 Sheri Pierce.
568
569 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
570 the first byte in a match must be "a".
571
572 11. Change 17 for 8.13 increased the recursion depth for patterns like
573 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
574 pattern contains any instances of (*THEN). If it does not, the old
575 optimizations are restored. It would be nice to do this on a per-group
576 basis, but at the moment that is not feasible.
577
578 12. In some environments, the output of pcretest -C is CRLF terminated. This
579 broke RunTest's code that checks for the link size. A single white space
580 character after the value is now allowed for.
581
582 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
583 For "fr", it uses the Windows-specific input and output files.
584
585 14. If (*THEN) appeared in a group that was called recursively or as a
586 subroutine, it did not work as intended. [But see next item.]
587
588 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
589 pattern fragments (but not containing any | characters). If A and B are
590 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
591 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
592 D. In other words, Perl considers parentheses that do not contain any |
593 characters to be part of a surrounding alternative, whereas PCRE was
594 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
595 differently. PCRE now behaves in the same way as Perl, except in the case
596 of subroutine/recursion calls such as (?1) which have in any case always
597 been different (but PCRE had them first :-).
598
599 16. Related to 15 above: Perl does not treat the | in a conditional group as
600 creating alternatives. Such a group is treated in the same way as an
601 ordinary group without any | characters when processing (*THEN). PCRE has
602 been changed to match Perl's behaviour.
603
604 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
605 RunGrepTest script failed.
606
607 18. Change 22 for version 13 caused atomic groups to use more stack. This is
608 inevitable for groups that contain captures, but it can lead to a lot of
609 stack use in large patterns. The old behaviour has been restored for atomic
610 groups that do not contain any capturing parentheses.
611
612 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
613 suppress the check for a minimum subject length at run time. (If it was
614 given to pcre_exec() or pcre_dfa_exec() it did work.)
615
616 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
617 fail to work when decoding hex characters in data strings in EBCDIC
618 environments.
619
620 21. It appears that in at least one Mac OS environment, the isxdigit() function
621 is implemented as a macro that evaluates to its argument more than once,
622 contravening the C 90 Standard (I haven't checked a later standard). There
623 was an instance in pcretest which caused it to go wrong when processing
624 \x{...} escapes in subject strings. The has been rewritten to avoid using
625 things like p++ in the argument of isxdigit().
626
627
628 Version 8.13 16-Aug-2011
629 ------------------------
630
631 1. The Unicode data tables have been updated to Unicode 6.0.0.
632
633 2. Two minor typos in pcre_internal.h have been fixed.
634
635 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
636 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
637 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
638
639 4. There were a number of related bugs in the code for matching backrefences
640 caselessly in UTF-8 mode when codes for the characters concerned were
641 different numbers of bytes. For example, U+023A and U+2C65 are an upper
642 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
643 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
644 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
645 2-byte code at the end of the subject (it thought there wasn't enough data
646 left).
647
648 5. Comprehensive information about what went wrong is now returned by
649 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
650 as the output vector has at least 2 elements. The offset of the start of
651 the failing character and a reason code are placed in the vector.
652
653 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
654 now returned is for the first byte of the failing character, instead of the
655 last byte inspected. This is an incompatible change, but I hope it is small
656 enough not to be a problem. It makes the returned offset consistent with
657 pcre_exec() and pcre_dfa_exec().
658
659 7. pcretest now gives a text phrase as well as the error number when
660 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
661 failure, the offset and reason code are output.
662
663 8. When \R was used with a maximizing quantifier it failed to skip backwards
664 over a \r\n pair if the subsequent match failed. Instead, it just skipped
665 back over a single character (\n). This seems wrong (because it treated the
666 two characters as a single entity when going forwards), conflicts with the
667 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
668 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
669 has been changed.
670
671 9. Some internal refactoring has changed the processing so that the handling
672 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
673 time (the PCRE_DOTALL option was changed this way some time ago: version
674 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
675 which was always a bit of a fudge. It also means that there is one less
676 argument for the match() function, which reduces its stack requirements
677 slightly. This change also fixes an incompatibility with Perl: the pattern
678 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
679
680 10. More internal refactoring has drastically reduced the number of recursive
681 calls to match() for possessively repeated groups such as (abc)++ when
682 using pcre_exec().
683
684 11. While implementing 10, a number of bugs in the handling of groups were
685 discovered and fixed:
686
687 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
688 (a|)*(?1) gave a compile-time internal error.
689 ((a|)+)+ did not notice that the outer group could match an empty string.
690 (^a|^)+ was not marked as anchored.
691 (.*a|.*)+ was not marked as matching at start or after a newline.
692
693 12. Yet more internal refactoring has removed another argument from the match()
694 function. Special calls to this function are now indicated by setting a
695 value in a variable in the "match data" data block.
696
697 13. Be more explicit in pcre_study() instead of relying on "default" for
698 opcodes that mean there is no starting character; this means that when new
699 ones are added and accidentally left out of pcre_study(), testing should
700 pick them up.
701
702 14. The -s option of pcretest has been documented for ages as being an old
703 synonym of -m (show memory usage). I have changed it to mean "force study
704 for every regex", that is, assume /S for every regex. This is similar to -i
705 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
706 using it. It makes it easier to run collections of tests with and without
707 study enabled, and thereby test pcre_study() more easily. All the standard
708 tests are now run with and without -s (but some patterns can be marked as
709 "never study" - see 20 below).
710
711 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
712 restoration of the capturing data to the outer values was not happening
713 correctly.
714
715 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
716 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
717 pattern had matched an empty string, and so incorrectly returned a no
718 match.
719
720 17. There was optimizing code for the last branch of non-capturing parentheses,
721 and also for the obeyed branch of a conditional subexpression, which used
722 tail recursion to cut down on stack usage. Unfortunately, now that there is
723 the possibility of (*THEN) occurring in these branches, tail recursion is
724 no longer possible because the return has to be checked for (*THEN). These
725 two optimizations have therefore been removed. [But see 8.20/11 above.]
726
727 18. If a pattern containing \R was studied, it was assumed that \R always
728 matched two bytes, thus causing the minimum subject length to be
729 incorrectly computed because \R can also match just one byte.
730
731 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
732 was incorrectly computed.
733
734 20. If /S is present twice on a test pattern in pcretest input, it now
735 *disables* studying, thereby overriding the use of -s on the command line
736 (see 14 above). This is necessary for one or two tests to keep the output
737 identical in both cases.
738
739 21. When (*ACCEPT) was used in an assertion that matched an empty string and
740 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
741
742 22. When an atomic group that contained a capturing parenthesis was
743 successfully matched, but the branch in which it appeared failed, the
744 capturing was not being forgotten if a higher numbered group was later
745 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
746 group 1 to "a", when in fact it should be unset. This applied to multi-
747 branched capturing and non-capturing groups, repeated or not, and also to
748 positive assertions (capturing in negative assertions does not happen
749 in PCRE) and also to nested atomic groups.
750
751 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
752 subject after a captured substring, to make it easier to tell which of a
753 number of identical substrings has been captured.
754
755 24. The way atomic groups are processed by pcre_exec() has been changed so that
756 if they are repeated, backtracking one repetition now resets captured
757 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
758 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
759 "aaa". Previously, it would have been "a". As part of this code
760 refactoring, the way recursive calls are handled has also been changed.
761
762 25. If an assertion condition captured any substrings, they were not passed
763 back unless some other capturing happened later. For example, if
764 (?(?=(a))a) was matched against "a", no capturing was returned.
765
766 26. When studying a pattern that contained subroutine calls or assertions,
767 the code for finding the minimum length of a possible match was handling
768 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
769 group 1 called group 2 while simultaneously a separate group 2 called group
770 1). A stack overflow occurred in this case. I have fixed this by limiting
771 the recursion depth to 10.
772
773 27. Updated RunTest.bat in the distribution to the version supplied by Tom
774 Fortmann. This supports explicit test numbers on the command line, and has
775 argument validation and error reporting.
776
777 28. An instance of \X with an unlimited repeat could fail if at any point the
778 first character it looked at was a mark character.
779
780 29. Some minor code refactoring concerning Unicode properties and scripts
781 should reduce the stack requirement of match() slightly.
782
783 30. Added the '=' option to pcretest to check the setting of unused capturing
784 slots at the end of the pattern, which are documented as being -1, but are
785 not included in the return count.
786
787 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
788 compiled something random. Now it gives a compile-time error (as does
789 Perl).
790
791 32. A *MARK encountered during the processing of a positive assertion is now
792 recorded and passed back (compatible with Perl).
793
794 33. If --only-matching or --colour was set on a pcregrep call whose pattern
795 had alternative anchored branches, the search for a second match in a line
796 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
797 matched the line "0102" twice. The same bug affected patterns that started
798 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
799 twice.
800
801 34. Previously, PCRE did not allow quantification of assertions. However, Perl
802 does, and because of capturing effects, quantifying parenthesized
803 assertions may at times be useful. Quantifiers are now allowed for
804 parenthesized assertions.
805
806 35. A minor code tidy in pcre_compile() when checking options for \R usage.
807
808 36. \g was being checked for fancy things in a character class, when it should
809 just be a literal "g".
810
811 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
812 appearance of a nested POSIX class supersedes an apparent external class.
813 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
814 unescaped square brackets may also appear as part of class names. For
815 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
816 more like Perl. (But see 8.20/1 above.)
817
818 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
819 was because it thought it was \N{name}, which is not supported).
820
821 39. Add minix to OS list not supporting the -S option in pcretest.
822
823 40. PCRE tries to detect cases of infinite recursion at compile time, but it
824 cannot analyze patterns in sufficient detail to catch mutual recursions
825 such as ((?1))((?2)). There is now a runtime test that gives an error if a
826 subgroup is called recursively as a subpattern for a second time at the
827 same position in the subject string. In previous releases this might have
828 been caught by the recursion limit, or it might have run out of stack.
829
830 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
831 happen only once. PCRE was, however incorrectly giving a compile time error
832 "recursive call could loop indefinitely" because it cannot analyze the
833 pattern in sufficient detail. The compile time test no longer happens when
834 PCRE is compiling a conditional subpattern, but actual runaway loops are
835 now caught at runtime (see 40 above).
836
837 42. It seems that Perl allows any characters other than a closing parenthesis
838 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
839 has been changed to be the same.
840
841 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
842 as not to get warnings when autogen.sh is called. Also changed
843 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
844
845 44. To help people who use pcregrep to scan files containing exceedingly long
846 lines, the following changes have been made:
847
848 (a) The default value of the buffer size parameter has been increased from
849 8K to 20K. (The actual buffer used is three times this size.)
850
851 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
852 PCRE is built.
853
854 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
855 to be set at run time.
856
857 (d) Numerical values in pcregrep options can be followed by K or M, for
858 example --buffer-size=50K.
859
860 (e) If a line being scanned overflows pcregrep's buffer, an error is now
861 given and the return code is set to 2.
862
863 45. Add a pointer to the latest mark to the callout data block.
864
865 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
866 partial match of an empty string instead of no match. This was specific to
867 the use of ".".
868
869 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
870 complete match instead of a partial match. This bug was dependent on both
871 the PCRE_UTF8 and PCRE_DOTALL options being set.
872
873 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
874 starting byte set, because \b was not being ignored.
875
876
877 Version 8.12 15-Jan-2011
878 ------------------------
879
880 1. Fixed some typos in the markup of the man pages, and wrote a script that
881 checks for such things as part of the documentation building process.
882
883 2. On a big-endian 64-bit system, pcregrep did not correctly process the
884 --match-limit and --recursion-limit options (added for 8.11). In
885 particular, this made one of the standard tests fail. (The integer value
886 went into the wrong half of a long int.)
887
888 3. If the --colour option was given to pcregrep with -v (invert match), it
889 did strange things, either producing crazy output, or crashing. It should,
890 of course, ignore a request for colour when reporting lines that do not
891 match.
892
893 4. Another pcregrep bug caused similar problems if --colour was specified with
894 -M (multiline) and the pattern match finished with a line ending.
895
896 5. In pcregrep, when a pattern that ended with a literal newline sequence was
897 matched in multiline mode, the following line was shown as part of the
898 match. This seems wrong, so I have changed it.
899
900 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
901 the check for further matches in the same line (so they could be coloured)
902 to overrun the end of the current line. If another match was found, it was
903 incorrectly shown (and then shown again when found in the next line).
904
905 7. If pcregrep was compiled under Windows, there was a reference to the
906 function pcregrep_exit() before it was defined. I am assuming this was
907 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
908 reported by a user. I've moved the definition above the reference.
909
910
911 Version 8.11 10-Dec-2010
912 ------------------------
913
914 1. (*THEN) was not working properly if there were untried alternatives prior
915 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
916 backtracked to try for "b" instead of moving to the next alternative branch
917 at the same level (in this case, to look for "c"). The Perl documentation
918 is clear that when (*THEN) is backtracked onto, it goes to the "next
919 alternative in the innermost enclosing group".
920
921 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
922 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
923 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
924 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
925 (*THEN).
926
927 3. If \s appeared in a character class, it removed the VT character from
928 the class, even if it had been included by some previous item, for example
929 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
930 of \s, but is part of the POSIX "space" class.)
931
932 4. A partial match never returns an empty string (because you can always
933 match an empty string at the end of the subject); however the checking for
934 an empty string was starting at the "start of match" point. This has been
935 changed to the "earliest inspected character" point, because the returned
936 data for a partial match starts at this character. This means that, for
937 example, /(?<=abc)def/ gives a partial match for the subject "abc"
938 (previously it gave "no match").
939
940 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
941 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
942 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
943 has an implication that the given string is incomplete (because a partial
944 match is preferred over a full match). For this reason, these items now
945 give a partial match in this situation. [Aside: previously, the one case
946 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
947 match rather than a full match, which was wrong by the old rules, but is
948 now correct.]
949
950 6. There was a bug in the handling of #-introduced comments, recognized when
951 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
952 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
953 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
954 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
955 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
956 places in pcre_compile().
957
958 7. Related to (6) above, when pcre_compile() was skipping #-introduced
959 comments when looking ahead for named forward references to subpatterns,
960 the only newline sequence it recognized was NL. It now handles newlines
961 according to the set newline convention.
962
963 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
964 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
965 cater for a lack of strerror(). These oversights have been fixed.
966
967 9. Added --match-limit and --recursion-limit to pcregrep.
968
969 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
970
971 11. When the -o option was used, pcregrep was setting a return code of 1, even
972 when matches were found, and --line-buffered was not being honoured.
973
974 12. Added an optional parentheses number to the -o and --only-matching options
975 of pcregrep.
976
977 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
978 can match an empty string. The code to do it in pcretest and pcredemo
979 needed fixing:
980
981 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
982 only one byte after an empty string match just before CRLF (this case
983 just got forgotten; "any" and "anycrlf" were OK).
984
985 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
986 mode when an empty string match preceded an ASCII character followed by
987 a non-ASCII character. (The code for advancing by one character rather
988 than one byte was nonsense.)
989
990 (c) The pcredemo.c sample program did not have any code at all to handle
991 the cases when CRLF is a valid newline sequence.
992
993 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
994 as a starting offset was within the subject string. There is now a new
995 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
996 negative or greater than the length of the string. In order to test this,
997 pcretest is extended to allow the setting of negative starting offsets.
998
999 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1000 starting offset points to the beginning of a UTF-8 character was
1001 unnecessarily clumsy. I tidied it up.
1002
1003 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1004 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1005
1006 17. Nobody had reported that the --include_dir option, which was added in
1007 release 7.7 should have been called --include-dir (hyphen, not underscore)
1008 for compatibility with GNU grep. I have changed it to --include-dir, but
1009 left --include_dir as an undocumented synonym, and the same for
1010 --exclude-dir, though that is not available in GNU grep, at least as of
1011 release 2.5.4.
1012
1013 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1014 characters from a string of bytes) have been redefined so as not to use
1015 loops, in order to improve performance in some environments. At the same
1016 time, I abstracted some of the common code into auxiliary macros to save
1017 repetition (this should not affect the compiled code).
1018
1019 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1020 compile-time error is now given if \c is not followed by an ASCII
1021 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1022 different, and any byte value is allowed.)
1023
1024 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1025 START_OPTIMIZE option, which is now allowed at compile time - but just
1026 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1027 to pcregrep and other applications that have no direct access to PCRE
1028 options. The new /Y option in pcretest sets this option when calling
1029 pcre_compile().
1030
1031 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1032 back references. Groups containing recursive back references were forced to
1033 be atomic by that change, but in the case of named groups, the amount of
1034 memory required was incorrectly computed, leading to "Failed: internal
1035 error: code overflow". This has been fixed.
1036
1037 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1038 pcretest.c, to avoid build problems in some Borland environments.
1039
1040
1041 Version 8.10 25-Jun-2010
1042 ------------------------
1043
1044 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1045 THEN.
1046
1047 2. (*ACCEPT) was not working when inside an atomic group.
1048
1049 3. Inside a character class, \B is treated as a literal by default, but
1050 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1051 causes the error). The code is unchanged, but I tidied the documentation.
1052
1053 4. Inside a character class, PCRE always treated \R and \X as literals,
1054 whereas Perl faults them if its -w option is set. I have changed PCRE so
1055 that it faults them when PCRE_EXTRA is set.
1056
1057 5. Added support for \N, which always matches any character other than
1058 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1059
1060 6. When compiling pcregrep with newer versions of gcc which may have
1061 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1062 declared with attribute warn_unused_result" were given. Just casting the
1063 result to (void) does not stop the warnings; a more elaborate fudge is
1064 needed. I've used a macro to implement this.
1065
1066 7. Minor change to pcretest.c to avoid a compiler warning.
1067
1068 8. Added four artifical Unicode properties to help with an option to make
1069 \s etc use properties (see next item). The new properties are: Xan
1070 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1071
1072 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1073 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1074 this option. Modified pcretest to add /W to test this facility. Added
1075 REG_UCP to make it available via the POSIX interface.
1076
1077 10. Added --line-buffered to pcregrep.
1078
1079 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1080 studied, and the match started with a letter with a code point greater than
1081 127 whose first byte was different to the first byte of the other case of
1082 the letter, the other case of this starting letter was not recognized
1083 (#976).
1084
1085 12. If a pattern that was studied started with a repeated Unicode property
1086 test, for example, \p{Nd}+, there was the theoretical possibility of
1087 setting up an incorrect bitmap of starting bytes, but fortunately it could
1088 not have actually happened in practice until change 8 above was made (it
1089 added property types that matched character-matching opcodes).
1090
1091 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1092 possible starting bytes for non-anchored patterns.
1093
1094 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1095 \R, and also a number of cases that involve Unicode properties, both
1096 explicit and implicit when PCRE_UCP is set.
1097
1098 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1099 input, it could crash or give wrong results if characters with values
1100 greater than 0xc0 were present in the subject string. (Detail: it assumed
1101 UTF-8 input when processing these items.)
1102
1103 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1104 size_t is 64-bit (#991).
1105
1106 17. Added a check for running out of memory when PCRE is compiled with
1107 --disable-stack-for-recursion (#990).
1108
1109 18. If the last data line in a file for pcretest does not have a newline on
1110 the end, a newline was missing in the output.
1111
1112 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1113 less than 128) in its various bitmaps. However, there is a facility for
1114 generating tables according to the current locale when PCRE is compiled. It
1115 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1116 characters, are recognized by isspace() and therefore were getting set in
1117 these tables, and indeed these tables seem to approximate to ISO 8859. This
1118 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1119 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1120 which of course cannot start UTF-8 characters. I have changed the code so
1121 that only real ASCII characters (less than 128) and the correct starting
1122 bytes for UTF-8 encodings are set for characters greater than 127 when in
1123 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1124 altogether.)
1125
1126 20. Added the /T option to pcretest so as to be able to run tests with non-
1127 standard character tables, thus making it possible to include the tests
1128 used for 19 above in the standard set of tests.
1129
1130 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1131 reference to a subpattern the other side of a comment that contains an
1132 opening parenthesis caused either an internal compiling error, or a
1133 reference to the wrong subpattern.
1134
1135
1136 Version 8.02 19-Mar-2010
1137 ------------------------
1138
1139 1. The Unicode data tables have been updated to Unicode 5.2.0.
1140
1141 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1142 configured.
1143
1144 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1145 original author of that file, following a query about its status.
1146
1147 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1148 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1149
1150 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1151 quantifier applied to a forward-referencing subroutine call, could compile
1152 incorrect code or give the error "internal error: previously-checked
1153 referenced subpattern not found".
1154
1155 6. Both MS Visual Studio and Symbian OS have problems with initializing
1156 variables to point to external functions. For these systems, therefore,
1157 pcre_malloc etc. are now initialized to local functions that call the
1158 relevant global functions.
1159
1160 7. There were two entries missing in the vectors called coptable and poptable
1161 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1162 I've fixed the data, and added a kludgy way of testing at compile time that
1163 the lengths are correct (equal to the number of opcodes).
1164
1165 8. Following on from 7, I added a similar kludge to check the length of the
1166 eint vector in pcreposix.c.
1167
1168 9. Error texts for pcre_compile() are held as one long string to avoid too
1169 much relocation at load time. To find a text, the string is searched,
1170 counting zeros. There was no check for running off the end of the string,
1171 which could happen if a new error number was added without updating the
1172 string.
1173
1174 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1175
1176 11. \K was not working if it appeared in an atomic group or in a group that
1177 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1178 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1179 the assertion is positive, but not if it is negative.
1180
1181 12. Change 11 fortuitously reduced the size of the stack frame used in the
1182 "match()" function of pcre_exec.c by one pointer. Forthcoming
1183 implementation of support for (*MARK) will need an extra pointer on the
1184 stack; I have reserved it now, so that the stack frame size does not
1185 decrease.
1186
1187 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1188 item in branch that calls a recursion is a subroutine call - as in the
1189 second branch in the above example - was incorrectly given the compile-
1190 time error "recursive call could loop indefinitely" because pcre_compile()
1191 was not correctly checking the subroutine for matching a non-empty string.
1192
1193 14. The checks for overrunning compiling workspace could trigger after an
1194 overrun had occurred. This is a "should never occur" error, but it can be
1195 triggered by pathological patterns such as hundreds of nested parentheses.
1196 The checks now trigger 100 bytes before the end of the workspace.
1197
1198 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1199
1200
1201 Version 8.01 19-Jan-2010
1202 ------------------------
1203
1204 1. If a pattern contained a conditional subpattern with only one branch (in
1205 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1206 computed the wrong minimum data length (which is of course zero for such
1207 subpatterns). This could cause incorrect "no match" results.
1208
1209 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1210 the pattern is reset in the first branch, pcre_compile() failed with
1211 "internal error: code overflow at offset...". This happened only when
1212 the reset was to the original external option setting. (An optimization
1213 abstracts leading options settings into an external setting, which was the
1214 cause of this.)
1215
1216 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1217 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1218 assertion pattern did not match (meaning that the assertion was true), it
1219 was incorrectly treated as false if the SKIP had been reached during the
1220 matching. This also applied to assertions used as conditions.
1221
1222 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1223 assertion subpattern, including such a pattern used as a condition,
1224 unpredictable results occurred, instead of the error return
1225 PCRE_ERROR_DFA_UITEM.
1226
1227 5. The C++ GlobalReplace function was not working like Perl for the special
1228 situation when an empty string is matched. It now does the fancy magic
1229 stuff that is necessary.
1230
1231 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1232 removed. (These were left over from very, very early versions of PCRE.)
1233
1234 7. Some cosmetic changes to the code to make life easier when compiling it
1235 as part of something else:
1236
1237 (a) Change DEBUG to PCRE_DEBUG.
1238
1239 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1240 called "current" as "current_branch", to prevent a collision with the
1241 Linux macro when compiled as a kernel module.
1242
1243 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1244 prevent a collision with the Linux macro when compiled as a kernel
1245 module.
1246
1247 8. In pcre_compile() there are some checks for integer overflows that used to
1248 cast potentially large values to (double). This has been changed to that
1249 when building, a check for int64_t is made, and if it is found, it is used
1250 instead, thus avoiding the use of floating point arithmetic. (There is no
1251 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1252 double.
1253
1254 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1255 2005 (difference between two addresses compared to an unsigned value).
1256
1257 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1258 custom one, because of the following reported problem in Windows:
1259
1260 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1261 under Win32.
1262 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1263 therefore missing the function definition.
1264 - The compiler thus generates a "C" signature for the test function.
1265 - The linker fails to find the "C" function.
1266 - PCRE fails to configure if asked to do so against libbz2.
1267
1268 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1269 messages were output:
1270
1271 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1272 rerunning libtoolize, to keep the correct libtool macros in-tree.
1273 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1274
1275 I have done both of these things.
1276
1277 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1278 most of the time, it *can* run out if it is given a pattern that contains a
1279 runaway infinite recursion. I updated the discussion in the pcrestack man
1280 page.
1281
1282 13. Now that we have gone to the x.xx style of version numbers, the minor
1283 version may start with zero. Using 08 or 09 is a bad idea because users
1284 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1285 interpreted as invalid octal numbers. I've updated the previous comment in
1286 configure.ac, and also added a check that gives an error if 08 or 09 are
1287 used.
1288
1289 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1290 causing partial matching to fail when the end of the subject matched \W
1291 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1292
1293 15. There were some discrepancies between the declarations in pcre_internal.h
1294 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1295 their definitions. The declarations used "const uschar *" and the
1296 definitions used USPTR. Even though USPTR is normally defined as "const
1297 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1298 reported that: "This difference in casting confuses some C++ compilers, for
1299 example, SunCC recognizes above declarations as different functions and
1300 generates broken code for hbpcre." I have changed the declarations to use
1301 USPTR.
1302
1303 16. GNU libtool is named differently on some systems. The autogen.sh script now
1304 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1305 (FreeBSD).
1306
1307 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1308 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1309 comment: "Figure out how to create a longlong from a string: strtoll and
1310 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1311 instance, but it only takes 2 args instead of 3!"
1312
1313 18. A subtle bug concerned with back references has been fixed by a change of
1314 specification, with a corresponding code fix. A pattern such as
1315 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1316 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1317 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1318 same bug. Such groups have to be quantified to be useful, or contained
1319 inside another quantified group. (If there's no repetition, the reference
1320 can never match.) The problem arises because, having left the group and
1321 moved on to the rest of the pattern, a later failure that backtracks into
1322 the group uses the captured value from the final iteration of the group
1323 rather than the correct earlier one. I have fixed this in PCRE by forcing
1324 any group that contains a reference to itself to be an atomic group; that
1325 is, there cannot be any backtracking into it once it has completed. This is
1326 similar to recursive and subroutine calls.
1327
1328
1329 Version 8.00 19-Oct-09
1330 ----------------------
1331
1332 1. The table for translating pcre_compile() error codes into POSIX error codes
1333 was out-of-date, and there was no check on the pcre_compile() error code
1334 being within the table. This could lead to an OK return being given in
1335 error.
1336
1337 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1338 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1339 in a Windows environment.
1340
1341 3. The pcregrep --count option prints the count for each file even when it is
1342 zero, as does GNU grep. However, pcregrep was also printing all files when
1343 --files-with-matches was added. Now, when both options are given, it prints
1344 counts only for those files that have at least one match. (GNU grep just
1345 prints the file name in this circumstance, but including the count seems
1346 more useful - otherwise, why use --count?) Also ensured that the
1347 combination -clh just lists non-zero counts, with no names.
1348
1349 4. The long form of the pcregrep -F option was incorrectly implemented as
1350 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1351 but it seems right to fix it, and I didn't think it was worth preserving
1352 the old behaviour.
1353
1354 5. The command line items --regex=pattern and --regexp=pattern were not
1355 recognized by pcregrep, which required --regex pattern or --regexp pattern
1356 (with a space rather than an '='). The man page documented the '=' forms,
1357 which are compatible with GNU grep; these now work.
1358
1359 6. No libpcreposix.pc file was created for pkg-config; there was just
1360 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1361
1362 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1363 when UCP support is not needed, by modifying the Python script that
1364 generates it from Unicode data files. This should not matter if the module
1365 is correctly used as a library, but I received one complaint about 50K of
1366 unwanted data. My guess is that the person linked everything into his
1367 program rather than using a library. Anyway, it does no harm.
1368
1369 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1370 was a minimum greater than 1 for a wide character in a possessive
1371 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1372 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1373 character. Chaos in the form of incorrect output or a compiling loop could
1374 result.
1375
1376 9. The restrictions on what a pattern can contain when partial matching is
1377 requested for pcre_exec() have been removed. All patterns can now be
1378 partially matched by this function. In addition, if there are at least two
1379 slots in the offset vector, the offset of the earliest inspected character
1380 for the match and the offset of the end of the subject are set in them when
1381 PCRE_ERROR_PARTIAL is returned.
1382
1383 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1384 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1385 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1386 and may be more useful for multi-segment matching.
1387
1388 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1389 used to be given if ever the end of the subject was reached; now it is
1390 given only if matching could not proceed because another character was
1391 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1392 string "Z", which now yields "no match" instead of "partial match". In the
1393 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1394 final character ended with (*FAIL).
1395
1396 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1397 if the pattern had a "must contain" character that was already found in the
1398 earlier partial match, unless partial matching was again requested. For
1399 example, with the pattern /dog.(body)?/, the "must contain" character is
1400 "g". If the first part-match was for the string "dog", restarting with
1401 "sbody" failed. This bug has been fixed.
1402
1403 13. The string returned by pcre_dfa_exec() after a partial match has been
1404 changed so that it starts at the first inspected character rather than the
1405 first character of the match. This makes a difference only if the pattern
1406 starts with a lookbehind assertion or \b or \B (\K is not supported by
1407 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1408 matching functions compatible, and I think it's the right thing to do.
1409
1410 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1411 so that the demonstration program is easily available in environments where
1412 PCRE has not been installed from source.
1413
1414 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1415 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1416 library.
1417
1418 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1419 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1420 is not the first non-POSIX option to be added. Clearly some people find
1421 these options useful.
1422
1423 17. If a caller to the POSIX matching function regexec() passes a non-zero
1424 value for nmatch with a NULL value for pmatch, the value of
1425 nmatch is forced to zero.
1426
1427 18. RunGrepTest did not have a test for the availability of the -u option of
1428 the diff command, as RunTest does. It now checks in the same way as
1429 RunTest, and also checks for the -b option.
1430
1431 19. If an odd number of negated classes containing just a single character
1432 interposed, within parentheses, between a forward reference to a named
1433 subpattern and the definition of the subpattern, compilation crashed with
1434 an internal error, complaining that it could not find the referenced
1435 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1436 [The bug was that it was starting one character too far in when skipping
1437 over the character class, thus treating the ] as data rather than
1438 terminating the class. This meant it could skip too much.]
1439
1440 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1441 /g option in pcretest when the pattern contains \K, which makes it possible
1442 to have an empty string match not at the start, even when the pattern is
1443 anchored. Updated pcretest and pcredemo to use this option.
1444
1445 21. If the maximum number of capturing subpatterns in a recursion was greater
1446 than the maximum at the outer level, the higher number was returned, but
1447 with unset values at the outer level. The correct (outer level) value is
1448 now given.
1449
1450 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1451 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1452 make it do so. The string so far is captured, making this feature
1453 compatible with Perl.
1454
1455 23. The tests have been re-organized, adding tests 11 and 12, to make it
1456 possible to check the Perl 5.10 features against Perl 5.10.
1457
1458 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1459 pattern matches a fixed length string. PCRE did not allow this; now it
1460 does. Neither allows recursion.
1461
1462 25. I finally figured out how to implement a request to provide the minimum
1463 length of subject string that was needed in order to match a given pattern.
1464 (It was back references and recursion that I had previously got hung up
1465 on.) This code has now been added to pcre_study(); it finds a lower bound
1466 to the length of subject needed. It is not necessarily the greatest lower
1467 bound, but using it to avoid searching strings that are too short does give
1468 some useful speed-ups. The value is available to calling programs via
1469 pcre_fullinfo().
1470
1471 26. While implementing 25, I discovered to my embarrassment that pcretest had
1472 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1473 study optimizations had never been tested with that matching function.
1474 Oops. What is worse, even when it was passed study data, there was a bug in
1475 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1476 were also very few tests of studied patterns with pcre_dfa_exec().
1477
1478 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1479 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1480 on the other side of the coin, they are no longer allowed to have different
1481 names, because these cannot be distinguished in PCRE, and this has caused
1482 confusion. (This is a difference from Perl.)
1483
1484 28. When duplicate subpattern names are present (necessarily with different
1485 numbers, as required by 27 above), and a test is made by name in a
1486 conditional pattern, either for a subpattern having been matched, or for
1487 recursion in such a pattern, all the associated numbered subpatterns are
1488 tested, and the overall condition is true if the condition is true for any
1489 one of them. This is the way Perl works, and is also more like the way
1490 testing by number works.
1491
1492
1493 Version 7.9 11-Apr-09
1494 ---------------------
1495
1496 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1497 (pcretest), all targets were linked against these libraries. This included
1498 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1499 libraries. This caused unwanted dependencies to be created. This problem
1500 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1501 pcretest is linked with readline.
1502
1503 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1504 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1505 moved outside it again, because FALSE and TRUE are already defined in AIX,
1506 but BOOL is not.
1507
1508 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1509 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1510
1511 4. The pcregrep documentation said spaces were inserted as well as colons (or
1512 hyphens) following file names and line numbers when outputting matching
1513 lines. This is not true; no spaces are inserted. I have also clarified the
1514 wording for the --colour (or --color) option.
1515
1516 5. In pcregrep, when --colour was used with -o, the list of matching strings
1517 was not coloured; this is different to GNU grep, so I have changed it to be
1518 the same.
1519
1520 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1521 each matching line was coloured. Now it goes on to look for further matches
1522 of any of the test patterns, which is the same behaviour as GNU grep.
1523
1524 7. A pattern that could match an empty string could cause pcregrep to loop; it
1525 doesn't make sense to accept an empty string match in pcregrep, so I have
1526 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1527 seems to be how GNU grep behaves.
1528
1529 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1530 start or after a newline", because the conditional assertion was not being
1531 correctly handled. The rule now is that both the assertion and what follows
1532 in the first alternative must satisfy the test.
1533
1534 9. If auto-callout was enabled in a pattern with a conditional group whose
1535 condition was an assertion, PCRE could crash during matching, both with
1536 pcre_exec() and pcre_dfa_exec().
1537
1538 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1539 used for matching.
1540
1541 11. Unicode property support in character classes was not working for
1542 characters (bytes) greater than 127 when not in UTF-8 mode.
1543
1544 12. Added the -M command line option to pcretest.
1545
1546 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1547
1548 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1549
1550 16. Added comments and documentation about mis-use of no_arg in the C++
1551 wrapper.
1552
1553 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1554 from Martin Jerabek that uses macro names for all relevant character and
1555 string constants.
1556
1557 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1558 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1559 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1560 these, but not everybody uses configure.
1561
1562 19. A conditional group that had only one branch was not being correctly
1563 recognized as an item that could match an empty string. This meant that an
1564 enclosing group might also not be so recognized, causing infinite looping
1565 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1566 with the subject "ab", where knowledge that the repeated group can match
1567 nothing is needed in order to break the loop.
1568
1569 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1570 exec(), but without supplying a callout function, matching went wrong.
1571
1572 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1573 leak if the size of the offset vector was greater than 30. When the vector
1574 is smaller, the saved offsets during recursion go onto a local stack
1575 vector, but for larger vectors malloc() is used. It was failing to free
1576 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1577 error, in fact).
1578
1579 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1580 heapframe that is used only when UTF-8 support is enabled. This caused no
1581 problem, but was untidy.
1582
1583 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1584 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1585 included within another project.
1586
1587 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1588 slightly modified by me:
1589
1590 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1591 not building pcregrep.
1592
1593 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1594 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1595
1596 25. Forward references, both numeric and by name, in patterns that made use of
1597 duplicate group numbers, could behave incorrectly or give incorrect errors,
1598 because when scanning forward to find the reference group, PCRE was not
1599 taking into account the duplicate group numbers. A pattern such as
1600 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1601
1602 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1603 the feature of a custom pointer more persuasive (as requested by a user).
1604
1605 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1606 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1607 pre-defined.
1608
1609 28. Added support for (*UTF8) at the start of a pattern.
1610
1611 29. Arrange for flags added by the "release type" setting in CMake to be shown
1612 in the configuration summary.
1613
1614
1615 Version 7.8 05-Sep-08
1616 ---------------------
1617
1618 1. Replaced UCP searching code with optimized version as implemented for Ad
1619 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1620 stage table and inline lookup instead of a function, giving speed ups of 2
1621 to 5 times on some simple patterns that I tested. Permission was given to
1622 distribute the MultiStage2.py script that generates the tables (it's not in
1623 the tarball, but is in the Subversion repository).
1624
1625 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1626 scripts.
1627
1628 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1629 a group with a zero qualifier. The result of the study could be incorrect,
1630 or the function might crash, depending on the pattern.
1631
1632 4. Caseless matching was not working for non-ASCII characters in back
1633 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1634 It now works when Unicode Property Support is available.
1635
1636 5. In pcretest, an escape such as \x{de} in the data was always generating
1637 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1638 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1639 truncation.
1640
1641 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1642
1643 7. Added two (int) casts to pcregrep when printing the difference of two
1644 pointers, in case they are 64-bit values.
1645
1646 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1647 test 2 if it fails.
1648
1649 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1650 and a #define of that name to empty if it is not externally set. This is to
1651 allow users of MSVC to set it if necessary.
1652
1653 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1654 the convenience functions in the pcre_get.c source file.
1655
1656 11. An option change at the start of a pattern that had top-level alternatives
1657 could cause overwriting and/or a crash. This command provoked a crash in
1658 some environments:
1659
1660 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1661
1662 This potential security problem was recorded as CVE-2008-2371.
1663
1664 12. For a pattern where the match had to start at the beginning or immediately
1665 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1666 pcre_dfa_exec() could read past the end of the passed subject if there was
1667 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1668 pcretest so that it places the subject at the end of its malloc-ed buffer.
1669
1670 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1671 exec() might read past the end of the data buffer in UTF-8 mode.
1672
1673 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1674 the data contained the byte 0x85 as part of a UTF-8 character within its
1675 first line. This applied both to normal and DFA matching.
1676
1677 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1678 /^[^d]*?$/8 failed to match "abc".
1679
1680 16. Added a missing copyright notice to pcrecpp_internal.h.
1681
1682 17. Make it more clear in the documentation that values returned from
1683 pcre_exec() in ovector are byte offsets, not character counts.
1684
1685 18. Tidied a few places to stop certain compilers from issuing warnings.
1686
1687 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1688 supplied by Stefan Weber. I made a further small update for 7.8 because
1689 there is a change of source arrangements: the pcre_searchfuncs.c module is
1690 replaced by pcre_ucd.c.
1691
1692
1693 Version 7.7 07-May-08
1694 ---------------------
1695
1696 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1697 a string to a long long, pretend we don't even have a long long." This is
1698 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1699
1700 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1701 pre-7.6 versions, which defined a global no_arg variable instead of putting
1702 it in the RE class. (See also #8 below.)
1703
1704 3. Remove a line of dead code, identified by coverity and reported by Nuno
1705 Lopes.
1706
1707 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1708
1709 (1) The include/exclude patterns were being applied to the whole pathnames
1710 of files, instead of just to the final components.
1711
1712 (2) If there was more than one level of directory, the subdirectories were
1713 skipped unless they satisfied the include/exclude conditions. This is
1714 inconsistent with GNU grep (and could even be seen as contrary to the
1715 pcregrep specification - which I improved to make it absolutely clear).
1716 The action now is always to scan all levels of directory, and just
1717 apply the include/exclude patterns to regular files.
1718
1719 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1720 --exclude_dir in the tests to avoid scanning .svn directories.
1721
1722 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1723 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1724 doesn't support NULs in patterns.
1725
1726 7. Added some missing "const"s to declarations of static tables in
1727 pcre_compile.c and pcre_dfa_exec.c.
1728
1729 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1730 caused by fix #2 above. (Subsequently also a second patch to fix the
1731 first patch. And a third patch - this was a messy problem.)
1732
1733 9. Applied Craig's patch to remove the use of push_back().
1734
1735 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1736 matching function regexec().
1737
1738 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1739 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1740 references. PCRE supports relative numbers with this syntax (I don't think
1741 Oniguruma does).
1742
1743 12. Previously, a group with a zero repeat such as (...){0} was completely
1744 omitted from the compiled regex. However, this means that if the group
1745 was called as a subroutine from elsewhere in the pattern, things went wrong
1746 (an internal error was given). Such groups are now left in the compiled
1747 pattern, with a new opcode that causes them to be skipped at execution
1748 time.
1749
1750 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1751 to the way PCRE behaves:
1752
1753 (a) A lone ] character is dis-allowed (Perl treats it as data).
1754
1755 (b) A back reference to an unmatched subpattern matches an empty string
1756 (Perl fails the current match path).
1757
1758 (c) A data ] in a character class must be notated as \] because if the
1759 first data character in a class is ], it defines an empty class. (In
1760 Perl it is not possible to have an empty class.) The empty class []
1761 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1762 The negative empty class [^] matches any one character, independently
1763 of the DOTALL setting.
1764
1765 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1766 non-existent subpattern following a character class starting with ']' and
1767 containing () gave an internal compiling error instead of "reference to
1768 non-existent subpattern". Fortunately, when the pattern did exist, the
1769 compiled code was correct. (When scanning forwards to check for the
1770 existencd of the subpattern, it was treating the data ']' as terminating
1771 the class, so got the count wrong. When actually compiling, the reference
1772 was subsequently set up correctly.)
1773
1774 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1775 it was being rejected as not supported by pcre_dfa_exec(), even though
1776 other assertions are supported. I have made pcre_dfa_exec() support
1777 (*FAIL).
1778
1779 16. The implementation of 13c above involved the invention of a new opcode,
1780 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1781 cannot be changed at match time, I realized I could make a small
1782 improvement to matching performance by compiling OP_ALLANY instead of
1783 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1784 on the OP_ANY path.
1785
1786 17. Compiling pcretest on Windows with readline support failed without the
1787 following two fixes: (1) Make the unistd.h include conditional on
1788 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1789
1790 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1791 ncurses library to be included for pcretest when ReadLine support is
1792 requested, but also to allow for it to be overridden. This patch came from
1793 Daniel Bergström.
1794
1795 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1796 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1797 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1798 spotting this.
1799
1800
1801 Version 7.6 28-Jan-08
1802 ---------------------
1803
1804 1. A character class containing a very large number of characters with
1805 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1806 overflow.
1807
1808 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1809 HAVE_LONG_LONG is not defined.
1810
1811 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1812 bring them up to date and include new features. This patch includes:
1813
1814 - Fixed PH's badly added libz and libbz2 support.
1815 - Fixed a problem with static linking.
1816 - Added pcredemo. [But later removed - see 7 below.]
1817 - Fixed dftables problem and added an option.
1818 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1819 HAVE_LONG_LONG.
1820 - Added readline support for pcretest.
1821 - Added an listing of the option settings after cmake has run.
1822
1823 4. A user submitted a patch to Makefile that makes it easy to create
1824 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1825 Makefile.am that cause it to include this special target, without
1826 affecting anything else. Note that the same mingw target plus all
1827 the other distribution libraries and programs are now supported
1828 when configuring with CMake (see 6 below) instead of with
1829 Configure/Make.
1830
1831 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1832 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1833 exported in the Windows port". It has not yet been confirmed that the patch
1834 solves the problem, but it does no harm.
1835
1836 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1837 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1838 with CMake, and also correct the comment about stack recursion.
1839
1840 7. Remove the automatic building of pcredemo from the ./configure system and
1841 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1842 of a program that users should build themselves after PCRE is installed, so
1843 building it automatically is not really right. What is more, it gave
1844 trouble in some build environments.
1845
1846 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1847
1848
1849 Version 7.5 10-Jan-08
1850 ---------------------
1851
1852 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1853 values in parens when parsing an RE using the C++ wrapper."
1854
1855 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1856 Characters greater than 255 were excluded from the class instead of being
1857 included.
1858
1859 3. The same bug as (2) above applied to negated POSIX classes such as
1860 [:^space:].
1861
1862 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1863 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1864 I have changed it.
1865
1866 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1867 first named subpattern) and a construct such as (?&a) would reference the
1868 first named subpattern whose name started with "a" (in other words, the
1869 length check was missing). Both these problems are fixed. "Subpattern name
1870 expected" is now given for (?&) (a zero-length name), and this patch also
1871 makes it give the same error for \k'' (previously it complained that that
1872 was a reference to a non-existent subpattern).
1873
1874 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1875 this is right because (?- can be followed by option settings as well as by
1876 digits. I have, however, made the messages clearer.
1877
1878 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1879 than the number used in the conditional) now cause a compile-time error.
1880 This is actually not compatible with Perl, which accepts such patterns, but
1881 treats the conditional as always being FALSE (as PCRE used to), but it
1882 seems to me that giving a diagnostic is better.
1883
1884 8. Change "alphameric" to the more common word "alphanumeric" in comments
1885 and messages.
1886
1887 9. Fix two occurrences of "backslash" in comments that should have been
1888 "backspace".
1889
1890 10. Remove two redundant lines of code that can never be obeyed (their function
1891 was moved elsewhere).
1892
1893 11. The program that makes PCRE's Unicode character property table had a bug
1894 which caused it to generate incorrect table entries for sequences of
1895 characters that have the same character type, but are in different scripts.
1896 It amalgamated them into a single range, with the script of the first of
1897 them. In other words, some characters were in the wrong script. There were
1898 thirteen such cases, affecting characters in the following ranges:
1899
1900 U+002b0 - U+002c1
1901 U+0060c - U+0060d
1902 U+0061e - U+00612
1903 U+0064b - U+0065e
1904 U+0074d - U+0076d
1905 U+01800 - U+01805
1906 U+01d00 - U+01d77
1907 U+01d9b - U+01dbf
1908 U+0200b - U+0200f
1909 U+030fc - U+030fe
1910 U+03260 - U+0327f
1911 U+0fb46 - U+0fbb1
1912 U+10450 - U+1049d
1913
1914 12. The -o option (show only the matching part of a line) for pcregrep was not
1915 compatible with GNU grep in that, if there was more than one match in a
1916 line, it showed only the first of them. It now behaves in the same way as
1917 GNU grep.
1918
1919 13. If the -o and -v options were combined for pcregrep, it printed a blank
1920 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1921 does the same. The return code can be used to tell if there were any
1922 non-matching lines.
1923
1924 14. Added --file-offsets and --line-offsets to pcregrep.
1925
1926 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1927 infinitely looping recursion. The bug was that positive lookaheads were not
1928 being skipped when checking for a possible empty match (negative lookaheads
1929 and both kinds of lookbehind were skipped).
1930
1931 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1932 inclusion of <windows.h> to before rather than after the definition of
1933 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1934
1935 17. Specifying a possessive quantifier with a specific limit for a Unicode
1936 character property caused pcre_compile() to compile bad code, which led at
1937 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1938 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1939 caused the error; without that there was no problem.
1940
1941 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1942
1943 19. Added --enable-pcretest-libreadline.
1944
1945 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1946 RE::GlobalReplace(). As a result, the number of replacements returned was
1947 double what it should be. I removed one of the increments, but Craig sent a
1948 later patch that removed the other one (the right fix) and added unit tests
1949 that check the return values (which was not done before).
1950
1951 21. Several CMake things:
1952
1953 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1954 the names libpcre and libpcreposix, not just pcre and pcreposix.
1955
1956 (2) The above change means that pcretest and pcregrep are now correctly
1957 linked with the newly-built libraries, not previously installed ones.
1958
1959 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1960
1961 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1962 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1963 UTF-8 newline character). The key issue is that the pattern starts .*;
1964 this means that the match must be either at the beginning, or after a
1965 newline. The bug was in the code for advancing after a failed match and
1966 checking that the new position followed a newline. It was not taking
1967 account of UTF-8 characters correctly.
1968
1969 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1970 character classes. PCRE was not treating the sequence [:...:] as a
1971 character class unless the ... were all letters. Perl, however, seems to
1972 allow any characters between [: and :], though of course it rejects as
1973 unknown any "names" that contain non-letters, because all the known class
1974 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1975 for example, whereas PCRE did not - it did not recognize a POSIX character
1976 class. This seemed a bit dangerous, so the code has been changed to be
1977 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1978 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1979 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1980 Perl does, and where it didn't before.
1981
1982 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1983 Windows environments %n is disabled by default.
1984
1985
1986 Version 7.4 21-Sep-07
1987 ---------------------
1988
1989 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1990 means that a class such as [\s] counted as "explicit reference to CR or
1991 LF". That isn't really right - the whole point of the change was to try to
1992 help when there was an actual mention of one of the two characters. So now
1993 the change happens only if \r or \n (or a literal CR or LF) character is
1994 encountered.
1995
1996 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1997 of both had grown to the point where there were only 3 bits left.
1998 Fortunately, there was spare space in the data structure, and so I have
1999 moved the internal flags into a new 16-bit field to free up more option
2000 bits.
2001
2002 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2003 but did not set the internal JCHANGED flag - either of these is enough to
2004 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2005 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2006 start sets both bits.
2007
2008 4. Added options (at build time, compile time, exec time) to change \R from
2009 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2010
2011 5. doc/pcresyntax.html was missing from the distribution.
2012
2013 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2014 compatibility, even though it is no longer used.
2015
2016 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2017 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2018 windows.h file is present (where different names are used). [This was
2019 reversed later after testing - see 16 below.]
2020
2021 8. Changed all #include <config.h> to #include "config.h". There were also
2022 some further <pcre.h> cases that I changed to "pcre.h".
2023
2024 9. When pcregrep was used with the --colour option, it missed the line ending
2025 sequence off the lines that it output.
2026
2027 10. It was pointed out to me that arrays of string pointers cause lots of
2028 relocations when a shared library is dynamically loaded. A technique of
2029 using a single long string with a table of offsets can drastically reduce
2030 these. I have refactored PCRE in four places to do this. The result is
2031 dramatic:
2032
2033 Originally: 290
2034 After changing UCP table: 187
2035 After changing error message table: 43
2036 After changing table of "verbs" 36
2037 After changing table of Posix names 22
2038
2039 Thanks to the folks working on Gregex for glib for this insight.
2040
2041 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2042 unicode-properties was also set.
2043
2044 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2045
2046 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2047 checked only for CRLF.
2048
2049 14. Added casts to pcretest.c to avoid compiler warnings.
2050
2051 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2052
2053 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2054 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2055 entirely. This removes changes made in 7 above.
2056
2057 17. The CMake files have been updated, and there is now more information about
2058 building with CMake in the NON-UNIX-USE document.
2059
2060
2061 Version 7.3 28-Aug-07
2062 ---------------------
2063
2064 1. In the rejigging of the build system that eventually resulted in 7.1, the
2065 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2066 brackets there is not right, since it causes compilers to look for an
2067 installed pcre.h, not the version that is in the source that is being
2068 compiled (which of course may be different). I have changed it back to:
2069
2070 #include "pcre.h"
2071
2072 I have a vague recollection that the change was concerned with compiling in
2073 different directories, but in the new build system, that is taken care of
2074 by the VPATH setting the Makefile.
2075
2076 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2077 when the subject happened to end in the byte 0x85 (e.g. if the last
2078 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2079 characters but of course it shouldn't be taken as a newline when it is part
2080 of another character. The bug was that, for an unlimited repeat of . in
2081 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2082 characters when looking for a newline.
2083
2084 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2085
2086 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2087 in debug output.
2088
2089 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2090 long printing in the pcrecpp unittest when running under MinGW.
2091
2092 6. ESC_K was left out of the EBCDIC table.
2093
2094 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2095 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2096 limit also applies to "virtual nesting" when a pattern is recursive, and in
2097 this case 1000 isn't so big. I have been able to remove this limit at the
2098 expense of backing off one optimization in certain circumstances. Normally,
2099 when pcre_exec() would call its internal match() function recursively and
2100 immediately return the result unconditionally, it uses a "tail recursion"
2101 feature to save stack. However, when a subpattern that can match an empty
2102 string has an unlimited repetition quantifier, it no longer makes this
2103 optimization. That gives it a stack frame in which to save the data for
2104 checking that an empty string has been matched. Previously this was taken
2105 from the 1000-entry workspace that had been reserved. So now there is no
2106 explicit limit, but more stack is used.
2107
2108 8. Applied Daniel's patches to solve problems with the import/export magic
2109 syntax that is required for Windows, and which was going wrong for the
2110 pcreposix and pcrecpp parts of the library. These were overlooked when this
2111 problem was solved for the main library.
2112
2113 9. There were some crude static tests to avoid integer overflow when computing
2114 the size of patterns that contain repeated groups with explicit upper
2115 limits. As the maximum quantifier is 65535, the maximum group length was
2116 set at 30,000 so that the product of these two numbers did not overflow a
2117 32-bit integer. However, it turns out that people want to use groups that
2118 are longer than 30,000 bytes (though not repeat them that many times).
2119 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2120 made it possible to implement the integer overflow checks in a much more
2121 dynamic way, which I have now done. The artificial limitation on group
2122 length has been removed - we now have only the limit on the total length of
2123 the compiled pattern, which depends on the LINK_SIZE setting.
2124
2125 10. Fixed a bug in the documentation for get/copy named substring when
2126 duplicate names are permitted. If none of the named substrings are set, the
2127 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2128 empty string.
2129
2130 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2131 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2132 because the ] is interpreted as the first data character and the
2133 terminating ] is not found. PCRE has been made compatible with Perl in this
2134 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2135 cause memory overwriting.
2136
2137 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2138 string has been matched (to stop an infinite loop). It was not recognizing
2139 a conditional subpattern that could match an empty string if that
2140 subpattern was within another subpattern. For example, it looped when
2141 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2142 condition was not nested. This bug has been fixed.
2143
2144 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2145 past the start of the subject in the presence of bytes with the top bit
2146 set, for example "\x8aBCD".
2147
2148 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2149 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2150
2151 14. Optimized (?!) to (*FAIL).
2152
2153 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2154 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2155 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2156 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2157 does: it's just the validity check that is more restrictive.
2158
2159 16. Inserted checks for integer overflows during escape sequence (backslash)
2160 processing, and also fixed erroneous offset values for syntax errors during
2161 backslash processing.
2162
2163 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2164 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2165
2166 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2167 caused an overrun.
2168
2169 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2170 something other than just ASCII characters) inside a group that had an
2171 unlimited repeat caused a loop at compile time (while checking to see
2172 whether the group could match an empty string).
2173
2174 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2175 [\P{Any}] did so. (Error in the code for printing property names.)
2176
2177 21. An orphan \E inside a character class could cause a crash.
2178
2179 22. A repeated capturing bracket such as (A)? could cause a wild memory
2180 reference during compilation.
2181
2182 23. There are several functions in pcre_compile() that scan along a compiled
2183 expression for various reasons (e.g. to see if it's fixed length for look
2184 behind). There were bugs in these functions when a repeated \p or \P was
2185 present in the pattern. These operators have additional parameters compared
2186 with \d, etc, and these were not being taken into account when moving along
2187 the compiled data. Specifically:
2188
2189 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2190 length.
2191
2192 (b) An item such as \pL+ within a repeated group could cause crashes or
2193 loops.
2194
2195 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2196 "reference to non-existent subpattern" error.
2197
2198 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2199
2200 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2201 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2202
2203 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2204
2205 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2206 character were causing crashes (broken optimization).
2207
2208 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2209 \p or \P) caused a compile-time loop.
2210
2211 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2212 break. For example, the unstudied pattern [\r\n]A does not match the string
2213 "\r\nA" because change 7.0/46 below moves the current point on by two
2214 characters after failing to match at the start. However, the pattern \nA
2215 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2216 the same is true. There doesn't seem any very clean way out of this, but
2217 what I have chosen to do makes the common cases work: PCRE now takes note
2218 of whether there can be an explicit match for \r or \n anywhere in the
2219 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2220 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2221 pattern has explicit CR or LF references.
2222
2223 29. Added (*CR) etc for changing newline setting at start of pattern.
2224
2225
2226 Version 7.2 19-Jun-07
2227 ---------------------
2228
2229 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2230 which is apparently normally available under Windows.
2231
2232 2. Re-jig the pcregrep tests with different newline settings in an attempt
2233 to make them independent of the local environment's newline setting.
2234
2235 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2236
2237 4. Some of the "internals" tests were previously cut out when the link size
2238 was not 2, because the output contained actual offsets. The recent new
2239 "Z" feature of pcretest means that these can be cut out, making the tests
2240 usable with all link sizes.
2241
2242 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2243 stack recursion. This gives a massive performance boost under BSD, but just
2244 a small improvement under Linux. However, it saves one field in the frame
2245 in all cases.
2246
2247 6. Added more features from the forthcoming Perl 5.10:
2248
2249 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2250 recursion call. It refers to the nth most recently opened parentheses.
2251
2252 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2253 to be opened parentheses.
2254
2255 (c) Conditions that refer to capturing parentheses can be specified
2256 relatively, for example, (?(-2)... or (?(+3)...
2257
2258 (d) \K resets the start of the current match so that everything before
2259 is not part of it.
2260
2261 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2262
2263 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2264 reference syntax.
2265
2266 (g) (?| introduces a group in which the numbering of parentheses in each
2267 alternative starts with the same number.
2268
2269 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2270
2271 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2272 PCRE_INFO_JCHANGED.
2273
2274 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2275 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2276 for detecting groups that can match an empty string.
2277
2278 9. A pattern with a very large number of alternatives (more than several
2279 hundred) was running out of internal workspace during the pre-compile
2280 phase, where pcre_compile() figures out how much memory will be needed. A
2281 bit of new cunning has reduced the workspace needed for groups with
2282 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2283 workspace instead of running out of the 4096 that are available.
2284
2285 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2286
2287 11. Applied patch from Google to remove an optimization that didn't quite work.
2288 The report of the bug said:
2289
2290 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2291 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2292 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2293
2294 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2295 it matched the wrong number of bytes.
2296
2297
2298 Version 7.1 24-Apr-07
2299 ---------------------
2300
2301 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2302 that is more "standard", making use of automake and other Autotools. There
2303 is some re-arrangement of the files and adjustment of comments consequent
2304 on this.
2305
2306 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2307 for recursive directory scanning broke on some systems because the files
2308 are not scanned in any specific order and on different systems the order
2309 was different. A call to "sort" has been inserted into RunGrepTest for the
2310 approprate test as a short-term fix. In the longer term there may be an
2311 alternative.
2312
2313 3. I had an email from Eric Raymond about problems translating some of PCRE's
2314 man pages to HTML (despite the fact that I distribute HTML pages, some
2315 people do their own conversions for various reasons). The problems
2316 concerned the use of low-level troff macros .br and .in. I have therefore
2317 removed all such uses from the man pages (some were redundant, some could
2318 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2319 HTML has been updated to handle .nf/.fi and to complain if it encounters
2320 .br or .in.
2321
2322 4. Updated comments in configure.ac that get placed in config.h.in and also
2323 arranged for config.h to be included in the distribution, with the name
2324 config.h.generic, for the benefit of those who have to compile without
2325 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2326
2327 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2328 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2329 makevp.bat for the current PCRE, using the additional files
2330 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2331
2332 6. A Windows user reported a minor discrepancy with test 2, which turned out
2333 to be caused by a trailing space on an input line that had got lost in his
2334 copy. The trailing space was an accident, so I've just removed it.
2335
2336 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2337 that is needed.
2338
2339 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2340 as "const" (a) because they are and (b) because it helps the PHP
2341 maintainers who have recently made a script to detect big data structures
2342 in the php code that should be moved to the .rodata section. I remembered
2343 to update Builducptable as well, so it won't revert if ucptable.h is ever
2344 re-created.
2345
2346 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2347 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2348 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2349 support is not required. This saves 1.5-2K of code, which is important in
2350 some applications.
2351
2352 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2353 so as not to refer to the tables, even though these functions will never be
2354 called when UTF-8 support is disabled. Otherwise there are problems with a
2355 shared library.
2356
2357 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2358
2359 (a) It was defining its arguments as char * instead of void *.
2360
2361 (b) It was assuming that all moves were upwards in memory; this was true
2362 a long time ago when I wrote it, but is no longer the case.
2363
2364 The emulated memove() is provided for those environments that have neither
2365 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2366 is clearly not the case, as these two bugs were recently reported.
2367
2368 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2369 and Detrail to create the HTML documentation, the .txt form of the man
2370 pages, and it removes trailing spaces from listed files. It also creates
2371 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2372 case, it wraps all the #defines with #ifndefs. This script should be run
2373 before "make dist".
2374
2375 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2376 with Unicode property support.
2377
2378 (a) For a maximizing quantifier, if the two different cases of the
2379 character were of different lengths in their UTF-8 codings (there are
2380 some cases like this - I found 11), and the matching function had to
2381 back up over a mixture of the two cases, it incorrectly assumed they
2382 were both the same length.
2383
2384 (b) When PCRE was configured to use the heap rather than the stack for
2385 recursion during matching, it was not correctly preserving the data for
2386 the other case of a UTF-8 character when checking ahead for a match
2387 while processing a minimizing repeat. If the check also involved
2388 matching a wide character, but failed, corruption could cause an
2389 erroneous result when trying to check for a repeat of the original
2390 character.
2391
2392 13. Some tidying changes to the testing mechanism:
2393
2394 (a) The RunTest script now detects the internal link size and whether there
2395 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2396 values substituted by "configure". (The RunGrepTest script already did
2397 this for UTF-8.) The configure.ac script no longer substitutes the
2398 relevant variables.
2399
2400 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2401 with length and offset values. This means that the output is different
2402 for different internal link sizes. Test 2 is skipped for link sizes
2403 other than 2 because of this, bypassing the problem. Unfortunately,
2404 there was also a test in test 3 (the locale tests) that used /B and
2405 failed for link sizes other than 2. Rather than cut the whole test out,
2406 I have added a new /Z option to pcretest that replaces the length and
2407 offset values with spaces. This is now used to make test 3 independent
2408 of link size. (Test 2 will be tidied up later.)
2409
2410 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2411 segmentation fault instead of returning the appropriate error message.
2412
2413 15. In multiline mode when the newline sequence was set to "any", the pattern
2414 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2415 This doesn't seem right; it now treats the CRLF combination as the line
2416 ending, and so does not match in that case. It's only a pattern such as ^$
2417 that would hit this one: something like ^ABC$ would have failed after \r
2418 and then tried again after \r\n.
2419
2420 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2421 in an attempt to make files that differ only in their line terminators
2422 compare equal. This works on Linux.
2423
2424 17. Under certain error circumstances pcregrep might try to free random memory
2425 as it exited. This is now fixed, thanks to valgrind.
2426
2427 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2428 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2429 was because its rules for how to advance for /g after matching an empty
2430 string at the end of a line did not allow for this case. They now check for
2431 it specially.
2432
2433 20. pcretest is supposed to handle patterns and data of any length, by
2434 extending its buffers when necessary. It was getting this wrong when the
2435 buffer for a data line had to be extended.
2436
2437 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2438 CRLF as a newline sequence.
2439
2440 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2441 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2442 I have nevertheless tidied it up.
2443
2444 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2445
2446 24. Added a man page for pcre-config.
2447
2448
2449 Version 7.0 19-Dec-06
2450 ---------------------
2451
2452 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2453 moving to gcc 4.1.1.
2454
2455 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2456 sys/time.h, which is documented as needed for this function. It doesn't
2457 seem to matter on Linux, but it showed up on some releases of OS X.
2458
2459 3. It seems that there are systems where bytes whose values are greater than
2460 127 match isprint() in the "C" locale. The "C" locale should be the
2461 default when a C program starts up. In most systems, only ASCII printing
2462 characters match isprint(). This difference caused the output from pcretest
2463 to vary, making some of the tests fail. I have changed pcretest so that:
2464
2465 (a) When it is outputting text in the compiled version of a pattern, bytes
2466 other than 32-126 are always shown as hex escapes.
2467
2468 (b) When it is outputting text that is a matched part of a subject string,
2469 it does the same, unless a different locale has been set for the match
2470 (using the /L modifier). In this case, it uses isprint() to decide.
2471
2472 4. Fixed a major bug that caused incorrect computation of the amount of memory
2473 required for a compiled pattern when options that changed within the
2474 pattern affected the logic of the preliminary scan that determines the
2475 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2476 that the computed length was too small. The symptoms of this bug were
2477 either the PCRE error "internal error: code overflow" from pcre_compile(),
2478 or a glibc crash with a message such as "pcretest: free(): invalid next
2479 size (fast)". Examples of patterns that provoked this bug (shown in
2480 pcretest format) are:
2481
2482 /(?-x: )/x
2483 /(?x)(?-x: \s*#\s*)/
2484 /((?i)[\x{c0}])/8
2485 /(?i:[\x{c0}])/8
2486
2487 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2488 is now done differently.
2489
2490 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2491 wrapper classes; (b) implement a new function in the C++ scanner that is
2492 more efficient than the old way of doing things because it avoids levels of
2493 recursion in the regex matching; (c) add a paragraph to the documentation
2494 for the FullMatch() function.
2495
2496 6. The escape sequence \n was being treated as whatever was defined as
2497 "newline". Not only was this contrary to the documentation, which states
2498 that \n is character 10 (hex 0A), but it also went horribly wrong when
2499 "newline" was defined as CRLF. This has been fixed.
2500
2501 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2502 was being set to -1 for the "end of line" case (supposedly a value that no
2503 character can have). Though this value is never used (the check for end of
2504 line is "zero bytes in current character"), it caused compiler complaints.
2505 I've changed it to 0xffffffff.
2506
2507 8. In pcre_version.c, the version string was being built by a sequence of
2508 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2509 string (as it is for production releases) called a macro with an empty
2510 argument. The C standard says the result of this is undefined. The gcc
2511 compiler treats it as an empty string (which was what was wanted) but it is
2512 reported that Visual C gives an error. The source has been hacked around to
2513 avoid this problem.
2514
2515 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2516 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2517 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2518 of them did).
2519
2520 10. Originally, pcretest opened its input and output without "b"; then I was
2521 told that "b" was needed in some environments, so it was added for release
2522 5.0 to both the input and output. (It makes no difference on Unix-like
2523 systems.) Later I was told that it is wrong for the input on Windows. I've
2524 now abstracted the modes into two macros, to make it easier to fiddle with
2525 them, and removed "b" from the input mode under Windows.
2526
2527 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2528
2529 12. Added -help and --help to pcretest as an official way of being reminded
2530 of the options.
2531
2532 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2533 and pcrecpp.cc because they annoy compilers at high warning levels.
2534
2535 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2536
2537 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2538 scripts are not C programs :-) and which was not noticed because it works
2539 on Linux.
2540
2541 16. pcretest is supposed to handle any length of pattern and data line (as one
2542 line or as a continued sequence of lines) by extending its input buffer if
2543 necessary. This feature was broken for very long pattern lines, leading to
2544 a string of junk being passed to pcre_compile() if the pattern was longer
2545 than about 50K.
2546
2547 17. I have done a major re-factoring of the way pcre_compile() computes the
2548 amount of memory needed for a compiled pattern. Previously, there was code
2549 that made a preliminary scan of the pattern in order to do this. That was
2550 OK when PCRE was new, but as the facilities have expanded, it has become
2551 harder and harder to keep it in step with the real compile phase, and there
2552 have been a number of bugs (see for example, 4 above). I have now found a
2553 cunning way of running the real compile function in a "fake" mode that
2554 enables it to compute how much memory it would need, while actually only
2555 ever using a few hundred bytes of working memory and without too many
2556 tests of the mode. This should make future maintenance and development
2557 easier. A side effect of this work is that the limit of 200 on the nesting
2558 depth of parentheses has been removed (though this was never a serious
2559 limitation, I suspect). However, there is a downside: pcre_compile() now
2560 runs more slowly than before (30% or more, depending on the pattern). I
2561 hope this isn't a big issue. There is no effect on runtime performance.
2562
2563 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2564 newline (only possible for the last line of a file) and it was a
2565 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2566
2567 19. Added additional timing features to pcretest. (1) The -tm option now times
2568 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2569 separate command line item, by a number that specifies the number of
2570 repeats to use when timing. The default is 50000; this gives better
2571 precision, but takes uncomfortably long for very large patterns.
2572
2573 20. Extended pcre_study() to be more clever in cases where a branch of a
2574 subpattern has no definite first character. For example, (a*|b*)[cd] would
2575 previously give no result from pcre_study(). Now it recognizes that the
2576 first character must be a, b, c, or d.
2577
2578 21. There was an incorrect error "recursive call could loop indefinitely" if
2579 a subpattern (or the entire pattern) that was being tested for matching an
2580 empty string contained only one non-empty item after a nested subpattern.
2581 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2582 incorrectly, because the \d was being skipped in the check.
2583
2584 22. The pcretest program now has a new pattern option /B and a command line
2585 option -b, which is equivalent to adding /B to every pattern. This causes
2586 it to show the compiled bytecode, without the additional information that
2587 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2588 is the same as /B/I).
2589
2590 23. A new optimization is now able automatically to treat some sequences such
2591 as a*b as a*+b. More specifically, if something simple (such as a character
2592 or a simple class like \d) has an unlimited quantifier, and is followed by
2593 something that cannot possibly match the quantified thing, the quantifier
2594 is automatically "possessified".
2595
2596 24. A recursive reference to a subpattern whose number was greater than 39
2597 went wrong under certain circumstances in UTF-8 mode. This bug could also
2598 have affected the operation of pcre_study().
2599
2600 25. Realized that a little bit of performance could be had by replacing
2601 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2602
2603 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2604
2605 27. Possessive quantifiers such as a++ were previously implemented by turning
2606 them into atomic groups such as ($>a+). Now they have their own opcodes,
2607 which improves performance. This includes the automatically created ones
2608 from 23 above.
2609
2610 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2611 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2612 the first matched character to be a colon. This applied both to named and
2613 numbered groups.
2614
2615 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2616
2617 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2618 building PCRE on Windows easier, so I have included it in the distribution.
2619
2620 31. There is now a check in pcretest against a ridiculously large number being
2621 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2622 loop, the loop is abandoned.
2623
2624 32. Forward references to subpatterns in conditions such as (?(2)...) where
2625 subpattern 2 is defined later cause pcre_compile() to search forwards in
2626 the pattern for the relevant set of parentheses. This search went wrong
2627 when there were unescaped parentheses in a character class, parentheses
2628 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2629
2630 33. "Subroutine" calls and backreferences were previously restricted to
2631 referencing subpatterns earlier in the regex. This restriction has now
2632 been removed.
2633
2634 34. Added a number of extra features that are going to be in Perl 5.10. On the
2635 whole, these are just syntactic alternatives for features that PCRE had
2636 previously implemented using the Python syntax or my own invention. The
2637 other formats are all retained for compatibility.
2638
2639 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2640 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2641 also .NET compatible.
2642
2643 (b) A recursion or subroutine call to a named group can now be defined as
2644 (?&name) as well as (?P>name).
2645
2646 (c) A backreference to a named group can now be defined as \k<name> or
2647 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2648 5.10, are also .NET compatible.
2649
2650 (d) A conditional reference to a named group can now use the syntax
2651 (?(<name>) or (?('name') as well as (?(name).
2652
2653 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2654 groups (named and numbered) that are never evaluated inline, but can be
2655 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2656 is always false. There may be only one alternative in such a group.
2657
2658 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2659 as the simple (?(R). The condition is true only if the most recent
2660 recursion is that of the given number or name. It does not search out
2661 through the entire recursion stack.
2662
2663 (g) The escape \gN or \g{N} has been added, where N is a positive or
2664 negative number, specifying an absolute or relative reference.
2665
2666 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2667 some "unreachable code" warnings.
2668
2669 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2670 things, this adds five new scripts.
2671
2672 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2673 There were also incompatibilities regarding the handling of \Q..\E inside
2674 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2675 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2676
2677 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2678 matches an empty string, and forcibly breaks the loop. There were bugs in
2679 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2680 against aaaa the result was just "a" rather than "aaaa", for example. Two
2681 separate and independent bugs (that affected different cases) have been
2682 fixed.
2683
2684 39. Refactored the code to abolish the use of different opcodes for small
2685 capturing bracket numbers. This is a tidy that I avoided doing when I
2686 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2687 The new approach is not only tidier, it makes it possible to reduce the
2688 memory needed to fix the previous bug (38).
2689
2690 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2691 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2692 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2693 mode.
2694
2695 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2696 report.
2697
2698 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2699 copy construction and assignment in the C++ wrapper.
2700
2701 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2702 couple of bugs that could have given wrong results in the "--newline=crlf"
2703 case.
2704
2705 44. Added a number of casts and did some reorganization of signed/unsigned int
2706 variables following suggestions from Dair Grant. Also renamed the variable
2707 "this" as "item" because it is a C++ keyword.
2708
2709 45. Arranged for dftables to add
2710
2711 #include "pcre_internal.h"
2712
2713 to pcre_chartables.c because without it, gcc 4.x may remove the array
2714 definition from the final binary if PCRE is built into a static library and
2715 dead code stripping is activated.
2716
2717 46. For an unanchored pattern, if a match attempt fails at the start of a
2718 newline sequence, and the newline setting is CRLF or ANY, and the next two
2719 characters are CRLF, advance by two characters instead of one.
2720
2721
2722 Version 6.7 04-Jul-06
2723 ---------------------
2724
2725 1. In order to handle tests when input lines are enormously long, pcretest has
2726 been re-factored so that it automatically extends its buffers when
2727 necessary. The code is crude, but this _is_ just a test program. The
2728 default size has been increased from 32K to 50K.
2729
2730 2. The code in pcre_study() was using the value of the re argument before
2731 testing it for NULL. (Of course, in any sensible call of the function, it
2732 won't be NULL.)
2733
2734 3. The memmove() emulation function in pcre_internal.h, which is used on
2735 systems that lack both memmove() and bcopy() - that is, hardly ever -
2736 was missing a "static" storage class specifier.
2737
2738 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2739 containing an extended class (one that cannot be represented by a bitmap
2740 because it contains high-valued characters or Unicode property items, e.g.
2741 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2742 pattern, but PCRE should not loop if you do not (it no longer does).
2743 [Detail: two cases were found: (a) a repeated subpattern containing an
2744 extended class; (b) a recursive reference to a subpattern that followed a
2745 previous extended class. It wasn't skipping over the extended class
2746 correctly when UTF-8 mode was not set.]
2747
2748 5. A negated single-character class was not being recognized as fixed-length
2749 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2750 compile error "lookbehind assertion is not fixed length".
2751
2752 6. The RunPerlTest auxiliary script was showing an unexpected difference
2753 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2754 write a Perl script that can interpret lines of an input file either as
2755 byte characters or as UTF-8, which is what "perltest" was being required to
2756 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2757 can't do is switch easily at run time between having the "use utf8;" pragma
2758 or not. In the end, I fudged it by using the RunPerlTest script to insert
2759 "use utf8;" explicitly for the UTF-8 tests.
2760
2761 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2762 the end of the subject string, contrary to the documentation and to what
2763 Perl does. This was true of both matching functions. Now it matches only at
2764 the start of the subject and immediately after *internal* newlines.
2765
2766 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2767 a pointer to an int instead of a pointer to an unsigned long int. This
2768 caused problems on 64-bit systems.
2769
2770 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2771 instance of the 'standard' template library not being so standard".
2772
2773 10. There was no check on the number of named subpatterns nor the maximum
2774 length of a subpattern name. The product of these values is used to compute
2775 the size of the memory block for a compiled pattern. By supplying a very
2776 long subpattern name and a large number of named subpatterns, the size
2777 computation could be caused to overflow. This is now prevented by limiting
2778 the length of names to 32 characters, and the number of named subpatterns
2779 to 10,000.
2780
2781 11. Subpatterns that are repeated with specific counts have to be replicated in
2782 the compiled pattern. The size of memory for this was computed from the
2783 length of the subpattern and the repeat count. The latter is limited to
2784 65535, but there was no limit on the former, meaning that integer overflow
2785 could in principle occur. The compiled length of a repeated subpattern is
2786 now limited to 30,000 bytes in order to prevent this.
2787
2788 12. Added the optional facility to have named substrings with the same name.
2789
2790 13. Added the ability to use a named substring as a condition, using the
2791 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2792 are numbers (not recommended). Forward references are permitted.
2793
2794 14. Added forward references in named backreferences (if you see what I mean).
2795
2796 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2797 pattern could run off the end of the subject. For example, the pattern
2798 "(?s)(.{1,5})"8 did this with the subject "ab".
2799
2800 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2801 PCRE_CASELESS was set when matching characters that were quantified with ?
2802 or *.
2803
2804 17. A character class other than a single negated character that had a minimum
2805 but no maximum quantifier - for example [ab]{6,} - was not handled
2806 correctly by pce_dfa_exec(). It would match only one character.
2807
2808 18. A valid (though odd) pattern that looked like a POSIX character
2809 class but used an invalid character after [ (for example [[,abc,]]) caused
2810 pcre_compile() to give the error "Failed: internal error: code overflow" or
2811 in some cases to crash with a glibc free() error. This could even happen if
2812 the pattern terminated after [[ but there just happened to be a sequence of
2813 letters, a binary zero, and a closing ] in the memory that followed.
2814
2815 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2816 over the years. Originally (before any Unicode support), just the bottom 8
2817 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2818 output from "man perlunicode" includes this:
2819
2820 The regular expression compiler produces polymorphic opcodes. That
2821 is, the pattern adapts to the data and automatically switches to
2822 the Unicode character scheme when presented with Unicode data--or
2823 instead uses a traditional byte scheme when presented with byte
2824 data.
2825
2826 Sadly, a wide octal escape does not cause a switch, and in a string with
2827 no other multibyte characters, these octal escapes are treated as before.
2828 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2829 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2830 Unicode string.
2831
2832 I have not perpetrated such confusion in PCRE. Up till now, it took just
2833 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2834 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2835 translate to the appropriate multibyte character.
2836
2837 29. Applied some refactoring to reduce the number of warnings from Microsoft
2838 and Borland compilers. This has included removing the fudge introduced
2839 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2840 a warning about an unused variable.
2841
2842 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2843 characters since release 4.0, because Perl (from release 5.004) does not.
2844 [Or at least, is documented not to: some releases seem to be in conflict
2845 with the documentation.] However, when a pattern was studied with
2846 pcre_study() and all its branches started with \s, PCRE still included VT
2847 as a possible starting character. Of course, this did no harm; it just
2848 caused an unnecessary match attempt.
2849
2850 22. Removed a now-redundant internal flag bit that recorded the fact that case
2851 dependency changed within the pattern. This was once needed for "required
2852 byte" processing, but is no longer used. This recovers a now-scarce options
2853 bit. Also moved the least significant internal flag bit to the most-
2854 significant bit of the word, which was not previously used (hangover from
2855 the days when it was an int rather than a uint) to free up another bit for
2856 the future.
2857
2858 23. Added support for CRLF line endings as well as CR and LF. As well as the
2859 default being selectable at build time, it can now be changed at runtime
2860 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2861 specify that it is scanning data with non-default line endings.
2862
2863 24. Changed the definition of CXXLINK to make it agree with the definition of
2864 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2865
2866 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2867 recursions. This makes a big different to stack usage for some patterns.
2868
2869 26. If a subpattern containing a named recursion or subroutine reference such
2870 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2871 the space required for the compiled pattern went wrong and gave too small a
2872 value. Depending on the environment, this could lead to "Failed: internal
2873 error: code overflow at offset 49" or "glibc detected double free or
2874 corruption" errors.
2875
2876 27. Applied patches from Google (a) to support the new newline modes and (b) to
2877 advance over multibyte UTF-8 characters in GlobalReplace.
2878
2879 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2880 difference for some implementation of PCRE in some Windows version.
2881
2882 29. Added some extra testing facilities to pcretest:
2883
2884 \q<number> in a data line sets the "match limit" value
2885 \Q<number> in a data line sets the "match recursion limt" value
2886 -S <number> sets the stack size, where <number> is in megabytes
2887
2888 The -S option isn't available for Windows.
2889
2890
2891 Version 6.6 06-Feb-06
2892 ---------------------
2893
2894 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2895 in pcreposix.h. I have copied the definition from pcre.h.
2896
2897 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2898 because pcre.h is no longer a built file.
2899
2900 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2901 not normally included in the compiled code.
2902
2903
2904 Version 6.5 01-Feb-06
2905 ---------------------
2906
2907 1. When using the partial match feature with pcre_dfa_exec(), it was not
2908 anchoring the second and subsequent partial matches at the new starting
2909 point. This could lead to incorrect results. For example, with the pattern
2910 /1234/, partially matching against "123" and then "a4" gave a match.
2911
2912 2. Changes to pcregrep:
2913
2914 (a) All non-match returns from pcre_exec() were being treated as failures
2915 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2916 error message is output. Some extra information is given for the
2917 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2918 probably the only errors that are likely to be caused by users (by
2919 specifying a regex that has nested indefinite repeats, for instance).
2920 If there are more than 20 of these errors, pcregrep is abandoned.
2921
2922 (b) A binary zero was treated as data while matching, but terminated the
2923 output line if it was written out. This has been fixed: binary zeroes
2924 are now no different to any other data bytes.
2925
2926 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2927 used to set a locale for matching. The --locale=xxxx long option has
2928 been added (no short equivalent) to specify a locale explicitly on the
2929 pcregrep command, overriding the environment variables.
2930
2931 (d) When -B was used with -n, some line numbers in the output were one less
2932 than they should have been.
2933
2934 (e) Added the -o (--only-matching) option.
2935
2936 (f) If -A or -C was used with -c (count only), some lines of context were
2937 accidentally printed for the final match.
2938
2939 (g) Added the -H (--with-filename) option.
2940
2941 (h) The combination of options -rh failed to suppress file names for files
2942 that were found from directory arguments.
2943
2944 (i) Added the -D (--devices) and -d (--directories) options.
2945
2946 (j) Added the -F (--fixed-strings) option.
2947
2948 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2949
2950 (l) Added the --colo(u)r option.
2951
2952 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2953 is not present by default.
2954
2955 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2956 items such as (?R) or (?1), when the recursion could match a number of
2957 alternatives. If it matched one of the alternatives, but subsequently,
2958 outside the recursion, there was a failure, the code tried to back up into
2959 the recursion. However, because of the way PCRE is implemented, this is not
2960 possible, and the result was an incorrect result from the match.
2961
2962 In order to prevent this happening, the specification of recursion has
2963 been changed so that all such subpatterns are automatically treated as
2964 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2965
2966 4. I had overlooked the fact that, in some locales, there are characters for
2967 which isalpha() is true but neither isupper() nor islower() are true. In
2968 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2969 and ordfeminine) are like this. This affected the treatment of \w and \W
2970 when they appeared in character classes, but not when they appeared outside
2971 a character class. The bit map for "word" characters is now created
2972 separately from the results of isalnum() instead of just taking it from the
2973 upper, lower, and digit maps. (Plus the underscore character, of course.)
2974
2975 5. The above bug also affected the handling of POSIX character classes such as
2976 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2977 permanent tables. Instead, the bit maps for such a class were previously
2978 created as the appropriate unions of the upper, lower, and digit bitmaps.
2979 Now they are created by subtraction from the [[:word:]] class, which has
2980 its own bitmap.
2981
2982 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2983 It is created by subtracting the vertical space characters (\x09, \x0a,
2984 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2985 subtraction was done in the overall bitmap for a character class, meaning
2986 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2987 be recognized. This bug has been fixed.
2988
2989 7. Patches from the folks at Google:
2990
2991 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2992 real life, but is still worth protecting against".
2993
2994 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2995 regular expressions".
2996
2997 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2998 have it.
2999
3000 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3001 "configure" and the latter not, in order to fix a problem somebody had
3002 with compiling the Arg class on HP-UX.
3003
3004 (e) Improve the error-handling of the C++ wrapper a little bit.
3005
3006 (f) New tests for checking recursion limiting.
3007
3008 8. The pcre_memmove() function, which is used only if the environment does not
3009 have a standard memmove() function (and is therefore rarely compiled),
3010 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3011 returning a result (though PCRE never actually uses the result).
3012
3013 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3014 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3015 returned instead of calling malloc() with an overflowing number that would
3016 most likely cause subsequent chaos.
3017
3018 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3019
3020 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3021 with this option is matched, the nmatch and pmatch options of regexec() are
3022 ignored.
3023
3024 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3025 provided in case anyone wants to the the POSIX interface with UTF-8
3026 strings.
3027
3028 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3029 C++ linking (needed for some HP-UX environments).
3030
3031 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3032 (unused parameter) and in the pcre_printint() function (omitted "default"
3033 switch label when the default is to do nothing).
3034
3035 15. Added some code to make it possible, when PCRE is compiled as a C++
3036 library, to replace subject pointers for pcre_exec() with a smart pointer
3037 class, thus making it possible to process discontinuous strings.
3038
3039 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3040 much the same function. They were added by different people who were trying
3041 to make PCRE easy to compile on non-Unix systems. It has been suggested
3042 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3043 for compiling on Windows systems. I have therefore replaced it with
3044 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3045 defaults to "extern" for C or "extern C" for C++, which works fine on
3046 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3047 SCOPE with something explicit in config.h. In addition:
3048
3049 (a) pcreposix.h still had just "extern" instead of either of these macros;
3050 I have replaced it with PCRE_DATA_SCOPE.
3051
3052 (b) Functions such as _pcre_xclass(), which are internal to the library,
3053 but external in the C sense, all had PCRE_EXPORT in their definitions.
3054 This is apparently wrong for the Windows case, so I have removed it.
3055 (It makes no difference on Unix-like systems.)
3056
3057 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3058 of recursive calls to match(). This is different to MATCH_LIMIT because
3059 that limits the total number of calls to match(), not all of which increase
3060 the depth of recursion. Limiting the recursion depth limits the amount of
3061 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3062 when PCRE is compiled, and changed at run time. A patch from Google adds
3063 this functionality to the C++ interface.
3064
3065 18. Changes to the handling of Unicode character properties:
3066
3067 (a) Updated the table to Unicode 4.1.0.
3068
3069 (b) Recognize characters that are not in the table as "Cn" (undefined).
3070
3071 (c) I revised the way the table is implemented to a much improved format
3072 which includes recognition of ranges. It now supports the ranges that
3073 are defined in UnicodeData.txt, and it also amalgamates other
3074 characters into ranges. This has reduced the number of entries in the
3075 table from around 16,000 to around 3,000, thus reducing its size
3076 considerably. I realized I did not need to use a tree structure after
3077 all - a binary chop search is just as efficient. Having reduced the
3078 number of entries, I extended their size from 6 bytes to 8 bytes to
3079 allow for more data.
3080
3081 (d) Added support for Unicode script names via properties such as \p{Han}.
3082
3083 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3084 matching that character.
3085
3086 20. When matching a repeated Unicode property with a minimum greater than zero,
3087 (for example \pL{2,}), PCRE could look past the end of the subject if it
3088 reached it while seeking the minimum number of characters. This could
3089 happen only if some of the characters were more than one byte long, because
3090 there is a check for at least the minimum number of bytes.
3091
3092 21. Refactored the implementation of \p and \P so as to be more general, to
3093 allow for more different types of property in future. This has changed the
3094 compiled form incompatibly. Anybody with saved compiled patterns that use
3095 \p or \P will have to recompile them.
3096
3097 22. Added "Any" and "L&" to the supported property types.
3098
3099 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3100 but give a compile time error if the value is greater than 0xff.
3101
3102 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3103 accidentally not being installed or uninstalled.
3104
3105 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3106 made were to insert the current release number. This seemed silly, because
3107 it made things harder for people building PCRE on systems that don't run
3108 "configure". I have turned pcre.h into a distributed file, no longer built
3109 by "configure", with the version identification directly included. There is
3110 no longer a pcre.h.in file.
3111
3112 However, this change necessitated a change to the pcre-config script as
3113 well. It is built from pcre-config.in, and one of the substitutions was the
3114 release number. I have updated configure.ac so that ./configure now finds
3115 the release number by grepping pcre.h.
3116
3117 26. Added the ability to run the tests under valgrind.
3118
3119
3120 Version 6.4 05-Sep-05
3121 ---------------------
3122
3123 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3124 "--" to be printed when multiple files were scanned, even when none of the
3125 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3126 consider it to be a bug, and have restored the previous behaviour.
3127
3128 2. A couple of code tidies to get rid of compiler warnings.
3129
3130 3. The pcretest program used to cheat by referring to symbols in the library
3131 whose names begin with _pcre_. These are internal symbols that are not
3132 really supposed to be visible externally, and in some environments it is
3133 possible to suppress them. The cheating is now confined to including
3134 certain files from the library's source, which is a bit cleaner.
3135
3136 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3137 file's purpose clearer.
3138
3139 5. Reorganized pcre_ucp_findchar().
3140
3141
3142 Version 6.3 15-Aug-05
3143 ---------------------
3144
3145 1. The file libpcre.pc.in did not have general read permission in the tarball.
3146
3147 2. There were some problems when building without C++ support:
3148
3149 (a) If C++ support was not built, "make install" and "make test" still
3150 tried to test it.
3151
3152 (b) There were problems when the value of CXX was explicitly set. Some
3153 changes have been made to try to fix these, and ...
3154
3155 (c) --disable-cpp can now be used to explicitly disable C++ support.
3156
3157 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3158 backslash in a target when C++ was disabled. This confuses some
3159 versions of "make", apparently. Using an intermediate variable solves
3160 this. (Same for CPP_LOBJ.)
3161
3162 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3163 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3164 necessary on certain architectures.
3165
3166 4. Added a setting of -export-symbols-regex to the link command to remove
3167 those symbols that are exported in the C sense, but actually are local
3168 within the library, and not documented. Their names all begin with
3169 "_pcre_". This is not a perfect job, because (a) we have to except some
3170 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3171 available (and never for static libraries). I have made a note to try to
3172 find a way round (a) in the future.
3173
3174
3175 Version 6.2 01-Aug-05
3176 ---------------------
3177
3178 1. There was no test for integer overflow of quantifier values. A construction
3179 such as {1111111111111111} would give undefined results. What is worse, if
3180 a minimum quantifier for a parenthesized subpattern overflowed and became
3181 negative, the calculation of the memory size went wrong. This could have
3182 led to memory overwriting.
3183
3184 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3185
3186 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3187 operating environments where this matters.
3188
3189 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3190 PCRE options from within the C++ wrapper.
3191
3192 5. Named capturing subpatterns were not being correctly counted when a pattern
3193 was compiled. This caused two problems: (a) If there were more than 100
3194 such subpatterns, the calculation of the memory needed for the whole
3195 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3196 back references of the form \12, where the number was greater than 9, were
3197 not recognized as back references, even though there were sufficient
3198 previous subpatterns.
3199
3200 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3201 versions of gcc, e.g. 2.95.4.
3202
3203
3204 Version 6.1 21-Jun-05
3205 ---------------------
3206
3207 1. There was one reference to the variable "posix" in pcretest.c that was not
3208 surrounded by "#if !defined NOPOSIX".
3209
3210 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3211 the cross-check on the old pcre_info() function, for the benefit of the
3212 cut-down version of PCRE that is currently imported into Exim.
3213
3214 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3215 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3216 patterns that start (?i) but I don't think that matters. The use of (?i) is
3217 just an example; this all applies to the other options as well.
3218
3219 4. Since libtool seems to echo the compile commands it is issuing, the output
3220 from "make" can be reduced a bit by putting "@" in front of each libtool
3221 compile command.
3222
3223 5. Patch from the folks at Google for configure.in to be a bit more thorough
3224 in checking for a suitable C++ installation before trying to compile the
3225 C++ stuff. This should fix a reported problem when a compiler was present,
3226 but no suitable headers.
3227
3228 6. The man pages all had just "PCRE" as their title. I have changed them to
3229 be the relevant file name. I have also arranged that these names are
3230 retained in the file doc/pcre.txt, which is a concatenation in text format
3231 of all the man pages except the little individual ones for each function.
3232
3233 7. The NON-UNIX-USE file had not been updated for the different set of source
3234 files that come with release 6. I also added a few comments about the C++
3235 wrapper.
3236
3237
3238 Version 6.0 07-Jun-05
3239 ---------------------
3240
3241 1. Some minor internal re-organization to help with my DFA experiments.
3242
3243 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3244 didn't matter for the library itself when fully configured, but did matter
3245 when compiling without UCP support, or within Exim, where the ucp files are
3246 not imported.
3247
3248 3. Refactoring of the library code to split up the various functions into
3249 different source modules. The addition of the new DFA matching code (see
3250 below) to a single monolithic source would have made it really too
3251 unwieldy, quite apart from causing all the code to be include in a
3252 statically linked application, when only some functions are used. This is
3253 relevant even without the DFA addition now that patterns can be compiled in
3254 one application and matched in another.
3255
3256 The downside of splitting up is that there have to be some external
3257 functions and data tables that are used internally in different modules of
3258 the library but which are not part of the API. These have all had their
3259 names changed to start with "_pcre_" so that they are unlikely to clash
3260 with other external names.
3261
3262 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3263 a different (DFA) algorithm. Although it is slower than the original
3264 function, it does have some advantages for certain types of matching
3265 problem.
3266
3267 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3268 including restarting after a partial match.
3269
3270 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3271 defined when compiling for Windows was sent to me. I have put it into the
3272 code, though I have no means of testing or verifying it.
3273
3274 7. Added the pcre_refcount() auxiliary function.
3275
3276 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3277 match before or at the first newline in the subject string. In pcretest,
3278 the /f option on a pattern can be used to set this.
3279
3280 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3281 would behave wrongly. This has been present in PCRE since release 4.0.
3282
3283 10. A number of changes to the pcregrep command:
3284
3285 (a) Refactored how -x works; insert ^(...)$ instead of setting
3286 PCRE_ANCHORED and checking the length, in preparation for adding
3287 something similar for -w.
3288
3289 (b) Added the -w (match as a word) option.
3290
3291 (c) Refactored the way lines are read and buffered so as to have more
3292 than one at a time available.
3293
3294 (d) Implemented a pcregrep test script.
3295
3296 (e) Added the -M (multiline match) option. This allows patterns to match
3297 over several lines of the subject. The buffering ensures that at least
3298 8K, or the rest of the document (whichever is the shorter) is available
3299 for matching (and similarly the previous 8K for lookbehind assertions).
3300
3301 (f) Changed the --help output so that it now says
3302
3303 -w, --word-regex(p)
3304
3305 instead of two lines, one with "regex" and the other with "regexp"
3306 because that confused at least one person since the short forms are the
3307 same. (This required a bit of code, as the output is generated
3308 automatically from a table. It wasn't just a text change.)
3309
3310 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3311 option but starts with a hyphen. Could be a pattern or a path name
3312 starting with a hyphen, for instance.
3313
3314 (h) "-" can be given as a file name to represent stdin.
3315
3316 (i) When file names are being printed, "(standard input)" is used for
3317 the standard input, for compatibility with GNU grep. Previously
3318 "<stdin>" was used.
3319
3320 (j) The option --label=xxx can be used to supply a name to be used for
3321 stdin when file names are being printed. There is no short form.
3322
3323 (k) Re-factored the options decoding logic because we are going to add
3324 two more options that take data. Such options can now be given in four
3325 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3326
3327 (l) Added the -A, -B, and -C options for requesting that lines of context
3328 around matches be printed.
3329
3330 (m) Added the -L option to print the names of files that do not contain
3331 any matching lines, that is, the complement of -l.
3332
3333 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3334 continue to scan other files.
3335
3336 (o) The -s option was incorrectly implemented. For compatibility with other
3337 greps, it now suppresses the error message for a non-existent or non-
3338 accessible file (but not the return code). There is a new option called
3339 -q that suppresses the output of matching lines, which was what -s was
3340 previously doing.
3341
3342 (p) Added --include and --exclude options to specify files for inclusion
3343 and exclusion when recursing.
3344
3345 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3346 Hopefully, it now does.
3347
3348 12. Missing cast in pcre_study().
3349
3350 13. Added an "uninstall" target to the makefile.
3351
3352 14. Replaced "extern" in the function prototypes in Makefile.in with
3353 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3354 world, but is set differently for Windows.
3355
3356 15. Added a second compiling function called pcre_compile2(). The only
3357 difference is that it has an extra argument, which is a pointer to an
3358 integer error code. When there is a compile-time failure, this is set
3359 non-zero, in addition to the error test pointer being set to point to an
3360 error message. The new argument may be NULL if no error number is required
3361 (but then you may as well call pcre_compile(), which is now just a
3362 wrapper). This facility is provided because some applications need a
3363 numeric error indication, but it has also enabled me to tidy up the way
3364 compile-time errors are handled in the POSIX wrapper.
3365
3366 16. Added VPATH=.libs to the makefile; this should help when building with one
3367 prefix path and installing with another. (Or so I'm told by someone who
3368 knows more about this stuff than I do.)
3369
3370 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3371 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3372 match everything, including newlines. This is not POSIX-compatible, but
3373 somebody wanted the feature. From pcretest it can be activated by using
3374 both the P and the s flags.
3375
3376 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3377
3378 19. libpcre.pc was being incorrectly installed as executable.
3379
3380 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3381 it now also looks for '\r' so that it will work unmodified on Windows.
3382
3383 21. Added Google's contributed C++ wrapper to the distribution.
3384
3385 22. Added some untidy missing memory free() calls in pcretest, to keep
3386 Electric Fence happy when testing.
3387
3388
3389
3390 Version 5.0 13-Sep-04
3391 ---------------------
3392
3393 1. Internal change: literal characters are no longer packed up into items
3394 containing multiple characters in a single byte-string. Each character
3395 is now matched using a separate opcode. However, there may be more than one
3396 byte in the character in UTF-8 mode.
3397
3398 2. The pcre_callout_block structure has two new fields: pattern_position and
3399 next_item_length. These contain the offset in the pattern to the next match
3400 item, and its length, respectively.
3401
3402 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3403 insertion of callouts before each pattern item. Added the /C option to
3404 pcretest to make use of this.
3405
3406 4. On the advice of a Windows user, the lines
3407
3408 #if defined(_WIN32) || defined(WIN32)
3409 _setmode( _fileno( stdout ), 0x8000 );
3410 #endif /* defined(_WIN32) || defined(WIN32) */
3411
3412 have been added to the source of pcretest. This apparently does useful
3413 magic in relation to line terminators.
3414
3415 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3416 for the benefit of those environments where the "b" makes a difference.
3417
3418 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3419 to know about it. I have put a hack into configure.in that adds in code
3420 to set GCC=yes if CC=icc. This seems to end up at a point in the
3421 generated configure script that is early enough to affect the setting of
3422 compiler options, which is what is needed, but I have no means of testing
3423 whether it really works. (The user who reported this had patched the
3424 generated configure script, which of course I cannot do.)
3425
3426 LATER: After change 22 below (new libtool files), the configure script
3427 seems to know about icc (and also ecc). Therefore, I have commented out
3428 this hack in configure.in.
3429
3430 7. Added support for pkg-config (2 patches were sent in).
3431
3432 8. Negated POSIX character classes that used a combination of internal tables
3433 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3434 [[:^ascii]]. Typically, they would match almost any characters. The other
3435 POSIX classes were not broken in this way.
3436
3437 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3438 to find the match, as PCRE was deluded into thinking that the match had to
3439 start at the start point or following a newline. The same bug applied to
3440 patterns with negative forward assertions or any backward assertions
3441 preceding ".*" at the start, unless the pattern required a fixed first
3442 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3443
3444 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3445 starting at the last subject character, bytes beyond the end of the subject
3446 string were read.
3447
3448 11. Renamed the variable "class" as "classbits" to make life easier for C++
3449 users. (Previously there was a macro definition, but it apparently wasn't
3450 enough.)
3451
3452 12. Added the new field "tables" to the extra data so that tables can be passed
3453 in at exec time, or the internal tables can be re-selected. This allows
3454 a compiled regex to be saved and re-used at a later time by a different
3455 program that might have everything at different addresses.
3456
3457 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3458 -R library as well as a -L library.
3459
3460 14. The debugging options of pcretest (-d on the command line or D on a
3461 pattern) showed incorrect output for anything following an extended class
3462 that contained multibyte characters and which was followed by a quantifier.
3463
3464 15. Added optional support for general category Unicode character properties
3465 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3466 support. It adds about 90K to the size of the library. The meanings of the
3467 inbuilt class escapes such as \d and \s have NOT been changed.
3468
3469 16. Updated pcredemo.c to include calls to free() to release the memory for the
3470 compiled pattern.
3471
3472 17. The generated file chartables.c was being created in the source directory
3473 instead of in the building directory. This caused the build to fail if the
3474 source directory was different from the building directory, and was
3475 read-only.
3476
3477 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3478 file. No doubt somebody will tell me if they don't make sense... Also added
3479 Dan Mooney's comments about building on OpenVMS.
3480
3481 19. Added support for partial matching via the PCRE_PARTIAL option for
3482 pcre_exec() and the \P data escape in pcretest.
3483
3484 20. Extended pcretest with 3 new pattern features:
3485
3486 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3487 write the compiled pattern to the file whose name is "rest-of-line".
3488 This is a straight binary dump of the data, with the saved pointer to
3489 the character tables forced to be NULL. The study data, if any, is
3490 written too. After writing, pcretest reads a new pattern.
3491
3492 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3493 compiled pattern from the given file. There must not be any
3494 occurrences of "<" in the file name (pretty unlikely); if there are,
3495 pcretest will instead treat the initial "<" as a pattern delimiter.
3496 After reading in the pattern, pcretest goes on to read data lines as
3497 usual.
3498
3499 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3500 and 16-bit fields in a compiled pattern, to simulate a pattern that
3501 was compiled on a host of opposite endianness.
3502
3503 21. The pcre-exec() function can now cope with patterns that were compiled on
3504 hosts of opposite endianness, with this restriction:
3505
3506 As for any compiled expression that is saved and used later, the tables
3507 pointer field cannot be preserved; the extra_data field in the arguments
3508 to pcre_exec() should be used to pass in a tables address if a value
3509 other than the default internal tables were used at compile time.
3510
3511 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3512 now diagnosed as an error. Previously, most of the time, a negative number
3513 would have been treated as zero, but if in addition "ovector" was passed as
3514 NULL, a crash could occur.
3515
3516 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3517 new versions from the libtool 1.5 distribution (the last one is a copy of
3518 a file called libtool.m4). This seems to have fixed the need to patch
3519 "configure" to support Darwin 1.3 (which I used to do). However, I still
3520 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3521 workstation).
3522
3523 24. Changed the PCRE licence to be the more standard "BSD" licence.
3524
3525
3526 Version 4.5 01-Dec-03
3527 ---------------------
3528
3529 1. There has been some re-arrangement of the code for the match() function so
3530 that it can be compiled in a version that does not call itself recursively.
3531 Instead, it keeps those local variables that need separate instances for
3532 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3533 needs to "recurse". Keeping track of where control must go is done by means
3534 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3535 hide most of the details from the main code, and operates only if
3536 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3537 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3538 operating.
3539
3540 To make it easier for callers to provide specially tailored get/free
3541 functions for this usage, two new functions, pcre_stack_malloc, and
3542 pcre_stack_free, are used. They are always called in strict stacking order,
3543 and the size of block requested is always the same.
3544
3545 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3546 PCRE has been compiled to use the stack or the heap for recursion. The
3547 -C option of pcretest uses this to show which version is compiled.
3548
3549 A new data escape \S, is added to pcretest; it causes the amounts of store
3550 obtained and freed by both kinds of malloc/free at match time to be added
3551 to the output.
3552
3553 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3554 what's available on my current Linux desktop machine.
3555
3556 3. When matching a UTF-8 string, the test for a valid string at the start has
3557 been extended. If start_offset is not zero, PCRE now checks that it points
3558 to a byte that is the start of a UTF-8 character. If not, it returns
3559 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3560 this is necessary because there may be backward assertions in the pattern.
3561 When matching the same subject several times, it may save resources to use
3562 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3563
3564 4. The code for checking the validity of UTF-8 strings has been tightened so
3565 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3566 containing "overlong sequences".
3567
3568 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3569 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3570 should have been "&", but it just so happened that all the cases this let
3571 through by mistake were picked up later in the function.
3572
3573 6. I had used a variable called "isblank" - this is a C99 function, causing
3574 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3575
3576 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3577 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3578 so the version is shown; (c) stop "make test" echoing "./RunTest".
3579
3580 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3581
3582 9. The prototype for memmove() for systems that don't have it was using
3583 size_t, but the inclusion of the header that defines size_t was later. I've
3584 moved the #includes for the C headers earlier to avoid this.
3585
3586 10. Added some adjustments to the code to make it easier to compiler on certain
3587 special systems:
3588
3589 (a) Some "const" qualifiers were missing.
3590 (b) Added the macro EXPORT before all exported functions; by default this
3591 is defined to be empty.
3592 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3593 that it reads its output file name as an argument instead of writing
3594 to the standard output and assuming this can be redirected.
3595
3596 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3597 class containing characters with values greater than 255, PCRE compilation
3598 went into a loop.
3599
3600 12. A recursive reference to a subpattern that was within another subpattern
3601 that had a minimum quantifier of zero caused PCRE to crash. For example,
3602 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3603 recursion. If the recursively-called subpattern itself had a zero repeat,
3604 that was OK.
3605
3606 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3607 buffer into which it was copied (for escape processing) was still set at
3608 1024, so long lines caused crashes.
3609
3610 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3611 "internal error: code overflow...". This applied to any character class
3612 that was followed by a possessive quantifier.
3613
3614 15. Modified the Makefile to add libpcre.la as a prerequisite for
3615 libpcreposix.la because I was told this is needed for a parallel build to
3616 work.
3617
3618 16. If a pattern that contained .* following optional items at the start was
3619 studied, the wrong optimizing data was generated, leading to matching
3620 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3621 matching string must start with a or b or c. The correct conclusion for
3622 this pattern is that a match can start with any character.
3623
3624
3625 Version 4.4 13-Aug-03
3626 ---------------------
3627
3628 1. In UTF-8 mode, a character class containing characters with values between
3629 127 and 255 was not handled correctly if the compiled pattern was studied.
3630 In fixing this, I have also improved the studying algorithm for such
3631 classes (slightly).
3632
3633 2. Three internal functions had redundant arguments passed to them. Removal
3634 might give a very teeny performance improvement.
3635
3636 3. Documentation bug: the value of the capture_top field in a callout is *one
3637 more than* the number of the hightest numbered captured substring.
3638
3639 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3640 in incorrectly linking with a previously installed version. They now link
3641 explicitly with libpcre.la.
3642
3643 5. configure.in no longer needs to recognize Cygwin specially.
3644
3645 6. A problem in pcre.in for Windows platforms is fixed.
3646
3647 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3648 pcretest, it used to include the size of the study block as part of its
3649 output. Unfortunately, the structure contains a field that has a different
3650 size on different hardware architectures. This meant that the tests that
3651 showed this size failed. As the block is currently always of a fixed size,
3652 this information isn't actually particularly useful in pcretest output, so
3653 I have just removed it.
3654
3655 8. Three pre-processor statements accidentally did not start in column 1.
3656 Sadly, there are *still* compilers around that complain, even though
3657 standard C has not required this for well over a decade. Sigh.
3658
3659 9. In pcretest, the code for checking callouts passed small integers in the
3660 callout_data field, which is a void * field. However, some picky compilers
3661 complained about the casts involved for this on 64-bit systems. Now
3662 pcretest passes the address of the small integer instead, which should get
3663 rid of the warnings.
3664
3665 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3666 both compile and run time, and gives an error if an invalid UTF-8 sequence
3667 is found. There is a option for disabling this check in cases where the
3668 string is known to be correct and/or the maximum performance is wanted.
3669
3670 11. In response to a bug report, I changed one line in Makefile.in from
3671
3672 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3673 to
3674 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3675
3676 to look similar to other lines, but I have no way of telling whether this
3677 is the right thing to do, as I do not use Windows. No doubt I'll get told
3678 if it's wrong...
3679
3680
3681 Version 4.3 21-May-03
3682 ---------------------
3683
3684 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3685 Makefile.
3686
3687 2. Some refactoring to improve the quality of the code:
3688
3689 (i) The utf8_table... variables are now declared "const".
3690
3691 (ii) The code for \cx, which used the "case flipping" table to upper case
3692 lower case letters, now just substracts 32. This is ASCII-specific,
3693 but the whole concept of \cx is ASCII-specific, so it seems
3694 reasonable.
3695
3696 (iii) PCRE was using its character types table to recognize decimal and
3697 hexadecimal digits in the pattern. This is silly, because it handles
3698 only 0-9, a-f, and A-F, but the character types table is locale-
3699 specific, which means strange things might happen. A private
3700 table is now used for this - though it costs 256 bytes, a table is
3701 much faster than multiple explicit tests. Of course, the standard
3702 character types table is still used for matching digits in subject
3703 strings against \d.
3704
3705 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3706 ending in _t are). So I've renamed it as ESC_tee.
3707
3708 3. The first argument for regexec() in the POSIX wrapper should have been
3709 defined as "const".
3710
3711 4. Changed pcretest to use malloc() for its buffers so that they can be
3712 Electric Fenced for debugging.
3713
3714 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3715 to read one or more bytes before the start of the subject string. Often this
3716 had no effect on PCRE's behaviour, but in some circumstances it could
3717 provoke a segmentation fault.
3718
3719 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3720 to try to read one or more bytes before the start of the subject string.
3721
3722 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3723 UTF-8 support could misbehave in various ways if the subject string
3724 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3725 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3726 back over UTF-8 characters.)
3727
3728
3729 Version 4.2 14-Apr-03
3730 ---------------------
3731
3732 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3733
3734 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3735 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3736 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3737 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3738 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3739 and BUILD_EXEEXT
3740 Note: automatic setting of the BUILD variables is not yet working
3741 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3742 compile-time but not at link-time
3743 [LINK]: use for linking executables only
3744 make different versions for Windows and non-Windows
3745 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3746 libraries
3747 [LINK_FOR_BUILD]: new variable
3748 [OBJEXT]: use throughout
3749 [EXEEXT]: use throughout
3750 <winshared>: new target
3751 <wininstall>: new target
3752 <dftables.o>: use native compiler
3753 <dftables>: use native linker
3754 <install>: handle Windows platform correctly
3755 <clean>: ditto
3756 <check>: ditto
3757 copy DLL to top builddir before testing
3758
3759 As part of these changes, -no-undefined was removed again. This was reported
3760 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3761 in any case.
3762
3763 3. Some tidies to get rid of compiler warnings:
3764
3765 . In the match_data structure, match_limit was an unsigned long int, whereas
3766 match_call_count was an int. I've made them both unsigned long ints.
3767
3768 . In pcretest the fact that a const uschar * doesn't automatically cast to
3769 a void * provoked a warning.
3770
3771 . Turning on some more compiler warnings threw up some "shadow" variables
3772 and a few more missing casts.
3773
3774 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3775 option, a class that contained a single character with a value between 128
3776 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3777
3778 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3779 option, a class that contained several characters, but with at least one
3780 whose value was between 128 and 255 caused PCRE to crash.
3781
3782
3783 Version 4.1 12-Mar-03
3784 ---------------------
3785
3786 1. Compiling with gcc -pedantic found a couple of places where casts were
3787 needed, and a string in dftables.c that was longer than standard compilers are
3788 required to support.
3789
3790 2. Compiling with Sun's compiler found a few more places where the code could
3791 be tidied up in order to avoid warnings.
3792
3793 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3794 first of these names is deprecated in the latest Autoconf in favour of the name
3795 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3796 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3797 analogy I have changed it to CFLAGS_FOR_BUILD.
3798
3799 4. Added -no-undefined to the linking command in the Makefile, because this is
3800 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3801 linking step for the pcreposix library.
3802
3803 5. PCRE was failing to diagnose the case of two named groups with the same
3804 name.
3805
3806 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3807 literal character that is needed in the subject for a match, and scans along to
3808 ensure that it is present before embarking on the full matching process. This
3809 saves time in cases of nested unlimited repeats that are never going to match.
3810 Problem: the scan can take a lot of time if the subject is very long (e.g.
3811 megabytes), thus penalizing straightforward matches. It is now done only if the
3812 amount of subject to be scanned is less than 1000 bytes.
3813
3814 7. A lesser problem with the same optimization is that it was recording the
3815 first character of an anchored pattern as "needed", thus provoking a search
3816 right along the subject, even when the first match of the pattern was going to
3817 fail. The "needed" character is now not set for anchored patterns, unless it
3818 follows something in the pattern that is of non-fixed length. Thus, it still
3819 fulfils its original purpose of finding quick non-matches in cases of nested
3820 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3821
3822
3823 Version 4.0 17-Feb-03
3824 ---------------------
3825
3826 1. If a comment in an extended regex that started immediately after a meta-item
3827 extended to the end of string, PCRE compiled incorrect data. This could lead to
3828 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3829
3830 2. Moved to autoconf 2.53 and libtool 1.4.2.
3831
3832 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3833 the special perltest8 script is no longer needed - all the tests can be run
3834 from a single perltest script.
3835
3836 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3837 by \s. It has now been removed in PCRE. This means it isn't recognized as
3838 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3839 class [:space:] *does* include VT, thereby creating a mess.
3840
3841 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3842 space and tab.
3843
3844 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3845 its new features into the main test script, reducing the number of scripts.
3846
3847 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3848 were backward compatible, and made the (?i) apply to the whole pattern, as if
3849 /i were given. Now it behaves more logically, and applies the option setting
3850 only to what follows. PCRE has been changed to follow suit. However, if it
3851 finds options settings right at the start of the pattern, it extracts them into
3852 the global options, as before. Thus, they show up in the info data.
3853
3854 8. Added support for the \Q...\E escape sequence. Characters in between are
3855 treated as literals. This is slightly different from Perl in that $ and @ are
3856 also handled as literals inside the quotes. In Perl, they will cause variable
3857 interpolation. Note the following examples:
3858
3859 Pattern PCRE matches Perl matches
3860
3861 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3862 \Qabc\$xyz\E abc\$xyz abc\$xyz
3863 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3864
3865 For compatibility with Perl, \Q...\E sequences are recognized inside character
3866 classes as well as outside them.
3867
3868 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3869 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3870 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3871 signed/unsigned warnings.
3872
3873 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3874 option for pcretest, so I've replaced it by a simple function that does just
3875 that job.
3876
3877 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3878 "pcregrep -".
3879
3880 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3881 Java package. This provides some syntactic sugar for simple cases of what my
3882 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3883 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3884 item, you can use this simplified notation. Note that only makes sense with
3885 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3886 greediness, whatever the setting of the PCRE_UNGREEDY option.
3887
3888 13. A change of greediness default within a pattern was not taking effect at
3889 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3890 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3891 was abstracted outside.
3892
3893 14. PCRE now supports the \G assertion. It is true when the current matching
3894 position is at the start point of the match. This differs from \A when the
3895 starting offset is non-zero. Used with the /g option of pcretest (or similar
3896 code), it works in the same way as it does for Perl's /g option. If all
3897 alternatives of a regex begin with \G, the expression is anchored to the start
3898 match position, and the "anchored" flag is set in the compiled expression.
3899
3900 15. Some bugs concerning the handling of certain option changes within patterns
3901 have been fixed. These applied to options other than (?ims). For example,
3902 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3903 been the other way round. Some of this was related to change 7 above.
3904
3905 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3906 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3907 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3908 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3909
3910 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3911 mode. Unlike ".", it always matches newline, whatever the setting of
3912 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3913 assertions. Perl allows it, but it doesn't (in general) work because it can't
3914 calculate the length of the lookbehind. At least, that's the case for Perl
3915 5.8.0 - I've been told they are going to document that it doesn't work in
3916 future.
3917
3918 18. Added an error diagnosis for escapes that PCRE does not support: these are
3919 \L, \l, \N, \P, \p, \U, \u, and \X.
3920
3921 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3922 reading past the end of the pattern in cases such as /[abcd/.
3923
3924 20. PCRE was getting more memory than necessary for patterns with classes that
3925 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3926
3927 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3928 compiling PCRE for use with Virtual Pascal.
3929
3930 22. Small fix to the Makefile to make it work properly if the build is done
3931 outside the source tree.
3932
3933 23. Added a new extension: a condition to go with recursion. If a conditional
3934 subpattern starts with (?(R) the "true" branch is used if recursion has
3935 happened, whereas the "false" branch is used only at the top level.
3936
3937 24. When there was a very long string of literal characters (over 255 bytes
3938 without UTF support, over 250 bytes with UTF support), the computation of how
3939 much memory was required could be incorrect, leading to segfaults or other
3940 strange effects.
3941
3942 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3943 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3944 there was a subsequent back reference to those brackets. This meant that, for
3945 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3946 possible to check for precisely this case. All we can do is abandon the
3947 optimization if .* occurs inside capturing brackets when there are any back
3948 references whatsoever. (See below for a better fix that came later.)
3949
3950 26. The handling of the optimization for finding the first character of a
3951 non-anchored pattern, and for finding a character that is required later in the
3952 match were failing in some cases. This didn't break the matching; it just
3953 failed to optimize when it could. The way this is done has been re-implemented.
3954
3955 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3956
3957 28. Added a new feature that provides some of the functionality that Perl
3958 provides with (?{...}). The facility is termed a "callout". The way it is done
3959 in PCRE is for the caller to provide an optional function, by setting
3960 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3961 global variable. By default it is unset, which disables all calling out. To get
3962 the function called, the regex must include (?C) at appropriate points. This
3963 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3964 This provides a means of identifying different callout points. When PCRE
3965 reaches such a point in the regex, if pcre_callout has been set, the external
3966 function is called. It is provided with data in a structure called
3967 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3968 matching continues; if it returns a non-zero value, the match at the current
3969 point fails. However, backtracking will occur if possible. [This was changed
3970 later and other features added - see item 49 below.]
3971
3972 29. pcretest is upgraded to test the callout functionality. It provides a
3973 callout function that displays information. By default, it shows the start of
3974 the match and the current position in the text. There are some new data escapes
3975 to vary what happens:
3976
3977 \C+ in addition, show current contents of captured substrings
3978 \C- do not supply a callout function
3979 \C!n return 1 when callout number n is reached
3980 \C!n!m return 1 when callout number n is reached for the mth time
3981
3982 30. If pcregrep was called with the -l option and just a single file name, it
3983 output "<stdin>" if a match was found, instead of the file name.
3984
3985 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3986 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3987 pcre_exec(). This saves a malloc/free per call. The default value of
3988 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3989 when configuring.
3990
3991 32. The default maximum size of a compiled pattern is 64K. There have been a
3992 few cases of people hitting this limit. The code now uses macros to handle the
3993 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3994 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3995 configuring. Tests 2 and 5 work only with 2-byte links because they output
3996 debugging information about compiled patterns.
3997
3998 33. Internal code re-arrangements:
3999
4000 (a) Moved the debugging function for printing out a compiled regex into
4001 its own source file (printint.c) and used #include to pull it into
4002 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4003 separate copies.
4004
4005 (b) Defined the list of op-code names for debugging as a macro in
4006 internal.h so that it is next to the definition of the opcodes.
4007
4008 (c) Defined a table of op-code lengths for simpler skipping along compiled
4009 code. This is again a macro in internal.h so that it is next to the
4010 definition of the opcodes.
4011
4012 34. Added support for recursive calls to individual subpatterns, along the
4013 lines of Robin Houston's patch (but implemented somewhat differently).
4014
4015 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4016 allow it to read and process whole directories in Win32. This code was
4017 contributed by Lionel Fourquaux; it has not been tested by me.
4018
4019 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4020 used to name a group. Names consist of alphanumerics and underscores, and must
4021 be unique. Back references use the syntax (?P=name) and recursive calls use
4022 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4023 numbers. The function pcre_fullinfo() can be used after compilation to extract
4024 a name/number map. There are three relevant calls:
4025
4026 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4027 PCRE_INFO_NAMECOUNT yields the number of entries
4028 PCRE_INFO_NAMETABLE yields a pointer to the map.
4029
4030 The map is a vector of fixed-size entries. The size of each entry depends on
4031 the length of the longest name used. The first two bytes of each entry are the
4032 group number, most significant byte first. There follows the corresponding
4033 name, zero terminated. The names are in alphabetical order.
4034
4035 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4036 case instead of 255. Making it the same both with and without UTF-8 support
4037 means that the same test output works with both.
4038
4039 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4040 calling malloc() with a zero argument.
4041
4042 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4043 optimization. I've improved things by keeping a bitmap of backreferences with
4044 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4045 fact referenced, the optimization can be applied. It is unlikely that a
4046 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4047 the match to follow \n) will appear inside brackets with a number greater than
4048 31, but if it does, any back reference > 31 suppresses the optimization.
4049
4050 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4051 of disabling numbered capturing parentheses. Any opening parenthesis that is
4052 not followed by ? behaves as if it were followed by ?: but named parentheses
4053 can still be used for capturing (and they will acquire numbers in the usual
4054 way).
4055
4056 41. Redesigned the return codes from the match() function into yes/no/error so
4057 that errors can be passed back from deep inside the nested calls. A malloc
4058 failure while inside a recursive subpattern call now causes the
4059 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4060
4061 42. It is now possible to set a limit on the number of times the match()
4062 function is called in a call to pcre_exec(). This facility makes it possible to
4063 limit the amount of recursion and backtracking, though not in a directly
4064 obvious way, because the match() function is used in a number of different
4065 circumstances. The count starts from zero for each position in the subject
4066 string (for non-anchored patterns). The default limit is, for compatibility, a
4067 large number, namely 10 000 000. You can change this in two ways:
4068
4069 (a) When configuring PCRE before making, you can use --with-match-limit=n
4070 to set a default value for the compiled library.
4071
4072 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4073 a different value is set. See 45 below.
4074
4075 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4076
4077 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4078 of things that can be changed at compile time. The first argument specifies
4079 what is wanted and the second points to where the information is to be placed.
4080 The current list of available information is:
4081
4082 PCRE_CONFIG_UTF8
4083
4084 The output is an integer that is set to one if UTF-8 support is available;
4085 otherwise it is set to zero.
4086
4087 PCRE_CONFIG_NEWLINE
4088
4089 The output is an integer that it set to the value of the code that is used for
4090 newline. It is either LF (10) or CR (13).
4091
4092 PCRE_CONFIG_LINK_SIZE
4093
4094 The output is an integer that contains the number of bytes used for internal
4095 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4096
4097 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4098
4099 The output is an integer that contains the threshold above which the POSIX
4100 interface uses malloc() for output vectors. See item 31 above.
4101
4102 PCRE_CONFIG_MATCH_LIMIT
4103
4104 The output is an unsigned integer that contains the default limit of the number
4105 of match() calls in a pcre_exec() execution. See 42 above.
4106
4107 44. pcretest has been upgraded by the addition of the -C option. This causes it
4108 to extract all the available output from the new pcre_config() function, and to
4109 output it. The program then exits immediately.
4110
4111 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4112 order to support additional features. One way would have been to define
4113 pcre_exec2() (for example) with extra arguments, but this would not have been
4114 extensible, and would also have required all calls to the original function to
4115 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4116 is used for passing in "extra" data from pcre_study().
4117
4118 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4119 contains the following fields:
4120
4121 flags a bitmap indicating which of the following fields are set
4122 study_data opaque data from pcre_study()
4123 match_limit a way of specifying a limit on match() calls for a specific
4124 call to pcre_exec()
4125 callout_data data for callouts (see 49 below)
4126
4127 The flag bits are also defined in pcre.h, and are
4128
4129 PCRE_EXTRA_STUDY_DATA
4130 PCRE_EXTRA_MATCH_LIMIT
4131 PCRE_EXTRA_CALLOUT_DATA
4132
4133 The pcre_study() function now returns one of these new pcre_extra blocks, with
4134 the actual study data pointed to by the study_data field, and the
4135 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4136 before. That is, this change is entirely upwards-compatible and requires no
4137 change to existing code.
4138
4139 If you want to pass in additional data to pcre_exec(), you can either place it
4140 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4141 block.
4142
4143 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4144 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4145 times with different match limits, until it finds the minimum value needed for
4146 pcre_exec() to complete. The value is then output. This can be instructive; for
4147 most simple matches the number is quite small, but for pathological cases it
4148 gets very large very quickly.
4149
4150 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4151 returns the size of the data block pointed to by the study_data field in a
4152 pcre_extra block, that is, the value that was passed as the argument to
4153 pcre_malloc() when PCRE was getting memory in which to place the information
4154 created by pcre_study(). The fourth argument should point to a size_t variable.
4155 pcretest has been extended so that this information is shown after a successful
4156 pcre_study() call when information about the compiled regex is being displayed.
4157
4158 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4159 because what follows is always an absolute path. (Later: it turns out that this
4160 is more than cosmetic for MinGW, because it doesn't like empty path
4161 components.)
4162
4163 49. Some changes have been made to the callout feature (see 28 above):
4164
4165 (i) A callout function now has three choices for what it returns:
4166
4167 0 => success, carry on matching
4168 > 0 => failure at this point, but backtrack if possible
4169 < 0 => serious error, return this value from pcre_exec()
4170
4171 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4172 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4173 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4174 use by callout functions. It will never be used by PCRE itself.
4175
4176 (ii) The pcre_extra structure (see 45 above) has a void * field called
4177 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4178 pcre_callout_block structure has a field of the same name. The contents of
4179 the field passed in the pcre_extra structure are passed to the callout
4180 function in the corresponding field in the callout block. This makes it
4181 easier to use the same callout-containing regex from multiple threads. For
4182 testing, the pcretest program has a new data escape
4183
4184 \C*n pass the number n (may be negative) as callout_data
4185
4186 If the callout function in pcretest receives a non-zero value as
4187 callout_data, it returns that value.
4188
4189 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4190 there were some redundant $(CFLAGS) in commands that are now specified as
4191 $(LINK), which already includes $(CFLAGS).
4192
4193 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4194 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4195 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4196 one-byte characters throughout. Note that case-insensitive matching applies
4197 only to characters whose values are less than 256. PCRE doesn't support the
4198 notion of cases for higher-valued characters.
4199
4200 (i) A character class whose characters are all within 0-255 is handled as
4201 a bit map, and the map is inverted for negative classes. Previously, a
4202 character > 255 always failed to match such a class; however it should
4203 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4204
4205 (ii) A negated character class with a single character < 255 is coded as
4206 "not this character" (OP_NOT). This wasn't working properly when the test
4207 character was multibyte, either singly or repeated.
4208
4209 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4210 mode, for example: \x{100}{2,3}.
4211
4212 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4213 singly or repeated) now correctly test multibyte characters. However,
4214 PCRE doesn't recognize any characters with values greater than 255 as
4215 digits, spaces, or word characters. Such characters always match \D, \S,
4216 and \W, and never match \d, \s, or \w.
4217
4218 (v) Classes may now contain characters and character ranges with values
4219 greater than 255. For example: [ab\x{100}-\x{400}].
4220
4221 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4222 PCRE in UTF-8 mode.
4223
4224 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4225 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4226 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4227 value.)
4228
4229 53. The single man page has become too large. I have therefore split it up into
4230 a number of separate man pages. These also give rise to individual HTML pages;
4231 these are now put in a separate directory, and there is an index.html page that
4232 lists them all. Some hyperlinking between the pages has been installed.
4233
4234 54. Added convenience functions for handling named capturing parentheses.
4235
4236 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4237 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4238 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4239 are faulted.
4240
4241 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4242 calling configure. These values are used when compiling the dftables.c program
4243 which is run to generate the source of the default character tables. They
4244 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4245 you will need to set these values.
4246
4247 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4248
4249
4250 Version 3.9 02-Jan-02
4251 ---------------------
4252
4253 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4254
4255 2. If --disable-static was given, the building process failed when trying to
4256 build pcretest and pcregrep. (For some reason it was using libtool to compile
4257 them, which is not right, as they aren't part of the library.)
4258
4259
4260 Version 3.8 18-Dec-01
4261 ---------------------
4262
4263 1. The experimental UTF-8 code was completely screwed up. It was packing the
4264 bytes in the wrong order. How dumb can you get?
4265
4266
4267 Version 3.7 29-Oct-01
4268 ---------------------
4269
4270 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4271 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4272 this didn't happen under Solaris 8, where I normally test things.
4273
4274 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4275 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4276 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4277 this fix an hour or so after the initial 3.7 release.)
4278
4279
4280 Version 3.6 23-Oct-01
4281 ---------------------
4282
4283 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4284 offsets passed as NULL with zero offset count.
4285
4286 2. The config.guess and config.sub files had not been updated when I moved to
4287 the latest autoconf.
4288
4289
4290 Version 3.5 15-Aug-01
4291 ---------------------
4292
4293 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4294 had been forgotten.
4295
4296 2. By using declared but undefined structures, we can avoid using "void"
4297 definitions in pcre.h while keeping the internal definitions of the structures
4298 private.
4299
4300 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4301 user point of view, this means that both static and shared libraries are built
4302 by default, but this can be individually controlled. More of the work of
4303 handling this static/shared cases is now inside libtool instead of PCRE's make
4304 file.
4305
4306 4. The pcretest utility is now installed along with pcregrep because it is
4307 useful for users (to test regexs) and by doing this, it automatically gets
4308 relinked by libtool. The documentation has been turned into a man page, so
4309 there are now .1, .txt, and .html versions in /doc.
4310
4311 5. Upgrades to pcregrep:
4312 (i) Added long-form option names like gnu grep.
4313 (ii) Added --help to list all options with an explanatory phrase.
4314 (iii) Added -r, --recursive to recurse into sub-directories.
4315 (iv) Added -f, --file to read patterns from a file.
4316
4317 6. pcre_exec() was referring to its "code" argument before testing that
4318 argument for NULL (and giving an error if it was NULL).
4319
4320 7. Upgraded Makefile.in to allow for compiling in a different directory from
4321 the source directory.
4322
4323 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4324 options bits, the pointer it was passed was to an int instead of to an unsigned
4325 long int. This mattered only on 64-bit systems.
4326
4327 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4328 generated) instead of pcre.in, which it its source. Also made the same change
4329 in several of the .c files.
4330
4331 10. A new release of gcc defines printf() as a macro, which broke pcretest
4332 because it had an ifdef in the middle of a string argument for printf(). Fixed
4333 by using separate calls to printf().
4334
4335 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4336 script, to force use of CR or LF instead of \n in the source. On non-Unix
4337 systems, the value can be set in config.h.
4338
4339 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4340 absolute limit. Changed the text of the error message to make this clear, and
4341 likewise updated the man page.
4342
4343 13. The limit of 99 on the number of capturing subpatterns has been removed.
4344 The new limit is 65535, which I hope will not be a "real" limit.
4345
4346
4347 Version 3.4 22-Aug-00
4348 ---------------------
4349
4350 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4351
4352 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4353
4354
4355 Version 3.3 01-Aug-00
4356 ---------------------
4357
4358 1. If an octal character was given, but the value was greater than \377, it
4359 was not getting masked to the least significant bits, as documented. This could
4360 lead to crashes in some systems.
4361
4362 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4363 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4364
4365 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4366 These just pass their arguments on to (pcre_free)(), but they are provided
4367 because some uses of PCRE bind it to non-C systems that can call its functions,
4368 but cannot call free() or pcre_free() directly.
4369
4370 4. Add "make test" as a synonym for "make check". Corrected some comments in
4371 the Makefile.
4372
4373 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4374 Makefile.
4375
4376 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4377 command called pgrep for grepping around the active processes.
4378
4379 7. Added the beginnings of support for UTF-8 character strings.
4380
4381 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4382 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4383 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4384 out for the ar command.)
4385
4386
4387 Version 3.2 12-May-00
4388 ---------------------
4389
4390 This is purely a bug fixing release.
4391
4392 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4393 of ZA. This was just one example of several cases that could provoke this bug,
4394 which was introduced by change 9 of version 2.00. The code for breaking
4395 infinite loops after an iteration that matches an empty string was't working
4396 correctly.
4397
4398 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4399 when matched against abbab (for example). After matching an empty string, it
4400 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4401 caused it to match further down the string than it should.
4402
4403 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4404 was there because it doesn't seem to be needed, and it causes trouble on some
4405 systems, as it is not a Standard C header. It has been removed.
4406
4407 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4408 were reported on the Macintosh. The changes were from
4409
4410 while ((c = *(++ptr)) != 0 && c != '\n');
4411 to
4412 while ((c = *(++ptr)) != 0 && c != '\n') ;
4413
4414 Totally extraordinary, but if that's what it takes...
4415
4416 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4417 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4418 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4419 assumes the way PCRE uses memmove() (always moving upwards).
4420
4421 6. PCRE is being used in one environment where strchr() is not available. There
4422 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4423 faster code anyway.
4424
4425
4426 Version 3.1 09-Feb-00
4427 ---------------------
4428
4429 The only change in this release is the fixing of some bugs in Makefile.in for
4430 the "install" target:
4431
4432 (1) It was failing to install pcreposix.h.
4433
4434 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4435
4436
4437 Version 3.0 01-Feb-00
4438 ---------------------
4439
4440 1. Add support for the /+ modifier to perltest (to output $` like it does in
4441 pcretest).
4442
4443 2. Add support for the /g modifier to perltest.
4444
4445 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4446 matches null strings.
4447
4448 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4449 pattern. Perl treats empty patterns specially - it reuses the most recent
4450 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4451 effect.
4452
4453 5. The POSIX interface was broken in that it was just handing over the POSIX
4454 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4455 required a bigger vector, with some working space on the end. This means that
4456 the POSIX wrapper now has to get and free some memory, and copy the results.
4457
4458 6. Added some simple autoconf support, placing the test data and the
4459 documentation in separate directories, re-organizing some of the
4460 information files, and making it build pcre-config (a GNU standard). Also added
4461 libtool support for building PCRE as a shared library, which is now the
4462 default.
4463
4464 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4465 09 are not valid octal constants. Single digits will be used for minor values
4466 less than 10.
4467
4468 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4469 existing programs that set these in the POSIX interface can use PCRE without
4470 modification.
4471
4472 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4473 return all that pcre_info() returns, plus additional data. The pcre_info()
4474 function is retained for compatibility, but is considered to be obsolete.
4475
4476 10. Added experimental recursion feature (?R) to handle one common case that
4477 Perl 5.6 will be able to do with (?p{...}).
4478
4479 11. Added support for POSIX character classes like [:alpha:], which Perl is
4480 adopting.
4481
4482
4483 Version 2.08 31-Aug-99
4484 ----------------------
4485
4486 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4487 trying to match at the startoffset position, but instead was moving forward to
4488 the next newline as if a previous match had failed.
4489
4490 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4491 and could get into a loop if a null string was matched other than at the start
4492 of the subject.
4493
4494 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4495 be distinguished at compile time, and for completeness also added PCRE_DATE.
4496
4497 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4498 in GnuWin32 environments.
4499
4500
4501 Version 2.07 29-Jul-99
4502 ----------------------
4503
4504 1. The documentation is now supplied in plain text form and HTML as well as in
4505 the form of man page sources.
4506
4507 2. C++ compilers don't like assigning (void *) values to other pointer types.
4508 In particular this affects malloc(). Although there is no problem in Standard
4509 C, I've put in casts to keep C++ compilers happy.
4510
4511 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4512 should be (const char *).
4513
4514 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4515 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4516 However, I haven't made this a standard facility. The documentation doesn't
4517 mention it, and the Makefile doesn't support it.
4518
4519 5. The Makefile now contains an "install" target, with editable destinations at
4520 the top of the file. The pcretest program is not installed.
4521
4522 6. pgrep -V now gives the PCRE version number and date.
4523
4524 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4525 causing the entire string to be ignored, instead of just the last character.
4526
4527 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4528 non-matching string, it can take a very, very long time, even for strings of
4529 quite modest length, because of the nested recursion. PCRE now does better in
4530 some of these cases. It does this by remembering the last required literal
4531 character in the pattern, and pre-searching the subject to ensure it is present
4532 before running the real match. In other words, it applies a heuristic to detect
4533 some types of certain failure quickly, and in the above example, if presented
4534 with a string that has no trailing " it gives "no match" very quickly.
4535
4536 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4537 other alternatives are tried instead.
4538
4539
4540 Version 2.06 09-Jun-99
4541 ----------------------
4542
4543 1. Change pcretest's output for amount of store used to show just the code
4544 space, because the remainder (the data block) varies in size between 32-bit and
4545 64-bit systems.
4546
4547 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4548 start matching at. This allows lookbehinds to work when searching for multiple
4549 occurrences in a string.
4550
4551 3. Added additional options to pcretest for testing multiple occurrences:
4552
4553 /+ outputs the rest of the string that follows a match
4554 /g loops for multiple occurrences, using the new startoffset argument
4555 /G loops for multiple occurrences by passing an incremented pointer
4556
4557 4. PCRE wasn't doing the "first character" optimization for patterns starting
4558 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4559 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4560 the letter 'x'. On long subject strings, this gives a significant speed-up.
4561
4562
4563 Version 2.05 21-Apr-99
4564 ----------------------
4565
4566 1. Changed the type of magic_number from int to long int so that it works
4567 properly on 16-bit systems.
4568
4569 2. Fixed a bug which caused patterns starting with .* not to work correctly
4570 when the subject string contained newline characters. PCRE was assuming
4571 anchoring for such patterns in all cases, which is not correct because .* will
4572 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4573 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4574 must be retried after every newline in the subject.
4575
4576
4577 Version 2.04 18-Feb-99
4578 ----------------------
4579
4580 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4581 computation of the store needed to hold the pattern was incorrect (too large).
4582 If such patterns were nested a few deep, this could multiply and become a real
4583 problem.
4584
4585 2. Added /M option to pcretest to show the memory requirement of a specific
4586 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4587
4588 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4589 compiled in such a way that the backtracking after subsequent failure was
4590 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4591 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4592
4593
4594 Version 2.03 02-Feb-99
4595 ----------------------
4596
4597 1. Fixed typo and small mistake in man page.
4598
4599 2. Added 4th condition (GPL supersedes if conflict) and created separate
4600 LICENCE file containing the conditions.
4601
4602 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4603 Perl, that is the internal \ allows the delimiter to be included in the
4604 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4605 the final delimiter, add \ to the end of the pattern (to test the error).
4606
4607 4. Added the convenience functions for extracting substrings after a successful
4608 match. Updated pcretest to make it able to test these functions.
4609
4610
4611 Version 2.02 14-Jan-99
4612 ----------------------
4613
4614 1. Initialized the working variables associated with each extraction so that
4615 their saving and restoring doesn't refer to uninitialized store.
4616
4617 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4618 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4619 fix the problem.
4620
4621 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4622 calls, and wasn't printing the correct value for compiling calls. Increased the
4623 default value of LOOPREPEAT, and the number of significant figures in the
4624 times.
4625
4626 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4627
4628 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4629 a building problem on Windows NT with a FAT file system.
4630
4631
4632 Version 2.01 21-Oct-98
4633 ----------------------
4634
4635 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4636 to character tables built by pcre_maketables() in the current locale. If NULL
4637 is passed, the default tables are used.
4638
4639
4640 Version 2.00 24-Sep-98
4641 ----------------------
4642
4643 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4644 it any more.
4645
4646 2. Allow quantification of (?>) groups, and make it work correctly.
4647
4648 3. The first character computation wasn't working for (?>) groups.
4649
4650 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4651 end of the subject) and add 5.005's \z, which really does match only at the
4652 very end of the subject.
4653
4654 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4655
4656 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4657 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4658 localized options. All options to pcre_study() were also removed.
4659
4660 7. Add other new features from 5.005:
4661
4662 $(?<= positive lookbehind
4663 $(?<! negative lookbehind
4664 (?imsx-imsx) added the unsetting capability
4665 such a setting is global if at outer level; local otherwise
4666 (?imsx-imsx:) non-capturing groups with option setting
4667 (?(cond)re|re) conditional pattern matching
4668
4669 A backreference to itself in a repeated group matches the previous
4670 captured string.
4671
4672 8. General tidying up of studying (both automatic and via "study")
4673 consequential on the addition of new assertions.
4674
4675 9. As in 5.005, unlimited repeated groups that could match an empty substring
4676 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4677 runtime if any iteration does actually match an empty substring.
4678
4679 10. Include the RunTest script in the distribution.
4680
4681 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4682 discrepancies, some of which were old and were also with respect to 5.004. They
4683 have now been fixed.
4684
4685
4686 Version 1.09 28-Apr-98
4687 ----------------------
4688
4689 1. A negated single character class followed by a quantifier with a minimum
4690 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4691 program crashes, or just wrong answers. This did not apply to negated classes
4692 containing more than one character, or to minima other than one.
4693
4694
4695 Version 1.08 27-Mar-98
4696 ----------------------
4697
4698 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4699
4700 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4701 latter must appear before anything that relies on it in the pattern.
4702
4703
4704 Version 1.07 16-Feb-98
4705 ----------------------
4706
4707 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4708 repeat of a potentially empty string).
4709
4710
4711 Version 1.06 23-Jan-98
4712 ----------------------
4713
4714 1. Added Markus Oberhumer's little patches for C++.
4715
4716 2. Literal strings longer than 255 characters were broken.
4717
4718
4719 Version 1.05 23-Dec-97
4720 ----------------------
4721
4722 1. Negated character classes containing more than one character were failing if
4723 PCRE_CASELESS was set at run time.
4724
4725
4726 Version 1.04 19-Dec-97
4727 ----------------------
4728
4729 1. Corrected the man page, where some "const" qualifiers had been omitted.
4730
4731 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4732 input syntax.
4733
4734 3. Fixed memory leak which occurred when a regex with back references was
4735 matched with an offsets vector that wasn't big enough. The temporary memory
4736 that is used in this case wasn't being freed if the match failed.
4737
4738 4. Tidied pcretest to ensure it frees memory that it gets.
4739
4740 5. Temporary memory was being obtained in the case where the passed offsets
4741 vector was exactly big enough.
4742
4743 6. Corrected definition of offsetof() from change 5 below.
4744
4745 7. I had screwed up change 6 below and broken the rules for the use of
4746 setjmp(). Now fixed.
4747
4748
4749 Version 1.03 18-Dec-97
4750 ----------------------
4751
4752 1. A erroneous regex with a missing opening parenthesis was correctly
4753 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4754 on some systems.
4755
4756 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4757 it was reported that one broken compiler failed on the former because "code" is
4758 also an independent variable.
4759
4760 3. The erroneous regex a[]b caused an array overrun reference.
4761
4762 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4763 fail on data ending with that character. (It was going on too far, and checking
4764 the next character, typically a binary zero.) This was specific to the
4765 optimized code for single-character negative classes.
4766
4767 5. Added a contributed patch from the TIN world which does the following:
4768
4769 + Add an undef for memmove, in case the the system defines a macro for it.
4770
4771 + Add a definition of offsetof(), in case there isn't one. (I don't know
4772 the reason behind this - offsetof() is part of the ANSI standard - but
4773 it does no harm).
4774
4775 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4776 most of the places where whitespace preceded '#'. I have given up and
4777 allowed the remaining 2 cases to be at the margin.
4778
4779 + Rename some variables in pcre to eliminate shadowing. This seems very
4780 pedantic, but does no harm, of course.
4781
4782 6. Moved the call to setjmp() into its own function, to get rid of warnings
4783 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4784
4785 7. Constructs such as \d{8,} were compiling into the equivalent of
4786 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4787 outcome, but in this particular case used more store than had been allocated,
4788 which caused the bug to be discovered because it threw up an internal error.
4789
4790 8. The debugging code in both pcre and pcretest for outputting the compiled
4791 form of a regex was going wrong in the case of back references followed by
4792 curly-bracketed repeats.
4793
4794
4795 Version 1.02 12-Dec-97
4796 ----------------------
4797
4798 1. Typos in pcre.3 and comments in the source fixed.
4799
4800 2. Applied a contributed patch to get rid of places where it used to remove
4801 'const' from variables, and fixed some signed/unsigned and uninitialized
4802 variable warnings.
4803
4804 3. Added the "runtest" target to Makefile.
4805
4806 4. Set default compiler flag to -O2 rather than just -O.
4807
4808
4809 Version 1.01 19-Nov-97
4810 ----------------------
4811
4812 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4813 like /([ab]*)*/, that is, for classes with more than one character in them.
4814
4815 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4816 as /((?>a*))*/ (a PCRE_EXTRA facility).
4817
4818
4819 Version 1.00 18-Nov-97
4820 ----------------------
4821
4822 1. Added compile-time macros to support systems such as SunOS4 which don't have
4823 memmove() or strerror() but have other things that can be used instead.
4824
4825 2. Arranged that "make clean" removes the executables.
4826
4827
4828 Version 0.99 27-Oct-97
4829 ----------------------
4830
4831 1. Fixed bug in code for optimizing classes with only one character. It was
4832 initializing a 32-byte map regardless, which could cause it to run off the end
4833 of the memory it had got.
4834
4835 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4836
4837
4838 Version 0.98 22-Oct-97
4839 ----------------------
4840
4841 1. Fixed bug in code for handling temporary memory usage when there are more
4842 back references than supplied space in the ovector. This could cause segfaults.
4843
4844
4845 Version 0.97 21-Oct-97
4846 ----------------------
4847
4848 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4849
4850 2. Optimized negated single characters not to use a bit map.
4851
4852 3. Brought error texts together as macro definitions; clarified some of them;
4853 fixed one that was wrong - it said "range out of order" when it meant "invalid
4854 escape sequence".
4855
4856 4. Changed some char * arguments to const char *.
4857
4858 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4859
4860 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4861 pcretest.
4862
4863
4864 Version 0.96 16-Oct-97
4865 ----------------------
4866
4867 1. Added a simple "pgrep" utility to the distribution.
4868
4869 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4870 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4871 where "ddd" means "one or more decimal digits".
4872
4873 3. Fixed serious bug. If a pattern had a back reference, but the call to
4874 pcre_exec() didn't supply a large enough ovector to record the related
4875 identifying subpattern, the match always failed. PCRE now remembers the number
4876 of the largest back reference, and gets some temporary memory in which to save
4877 the offsets during matching if necessary, in order to ensure that
4878 backreferences always work.
4879
4880 4. Increased the compatibility with Perl in a number of ways:
4881
4882 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4883 to request this handling. The option can be set at compile or exec time.
4884
4885 (b) $ matches before a terminating newline by default; an option
4886 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4887 mode). The option can be set at compile or exec time.
4888
4889 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4890 the same as Perl's. If the decimal number it represents is less than 10
4891 or there aren't that many previous left capturing parentheses, an octal
4892 escape is read. Inside a character class, it's always an octal escape,
4893 even if it is a single digit.
4894
4895 (d) An escaped but undefined alphabetic character is taken as a literal,
4896 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4897 escapes.
4898
4899 (e) {0} is now permitted. (The previous item is removed from the compiled
4900 pattern).
4901
4902 5. Changed all the names of code files so that the basic parts are no longer
4903 than 10 characters, and abolished the teeny "globals.c" file.
4904
4905 6. Changed the handling of character classes; they are now done with a 32-byte
4906 bit map always.
4907
4908 7. Added the -d and /D options to pcretest to make it possible to look at the
4909 internals of compilation without having to recompile pcre.
4910
4911
4912 Version 0.95 23-Sep-97
4913 ----------------------
4914
4915 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4916 \x20 at the start of a run of normal characters. These were being treated as
4917 real characters, instead of the source characters being re-checked.
4918
4919
4920 Version 0.94 18-Sep-97
4921 ----------------------
4922
4923 1. The functions are now thread-safe, with the caveat that the global variables
4924 containing pointers to malloc() and free() or alternative functions are the
4925 same for all threads.
4926
4927 2. Get pcre_study() to generate a bitmap of initial characters for non-
4928 anchored patterns when this is possible, and use it if passed to pcre_exec().
4929
4930
4931 Version 0.93 15-Sep-97
4932 ----------------------
4933
4934 1. /(b)|(:+)/ was computing an incorrect first character.
4935
4936 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4937 but not actually doing anything yet.
4938
4939 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4940 as Perl does (e.g. [-az] or [az-]).
4941
4942 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4943 all possible positions.
4944
4945 5. Split up into different modules to avoid including unneeded functions in a
4946 compiled binary. However, compile and exec are still in one module. The "study"
4947 function is split off.
4948
4949 6. The character tables are now in a separate module whose source is generated
4950 by an auxiliary program - but can then be edited by hand if required. There are
4951 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4952 toupper() in the code.
4953
4954 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4955 make them global. Abolish the function for setting them, as the caller can now
4956 set them directly.
4957
4958
4959 Version 0.92 11-Sep-97
4960 ----------------------
4961
4962 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4963 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4964
4965 2. Caseless matching was not working in character classes if the characters in
4966 the pattern were in upper case.
4967
4968 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4969
4970 4. Make PCRE_ANCHORED public and accept as a compile option.
4971
4972 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4973 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4974 pass them.
4975
4976 6. Give an error if bad option bits passed at compile or run time.
4977
4978 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4979 pcretest to cause it to pass that flag.
4980
4981 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4982 options, and the first character, if set.
4983
4984 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4985
4986
4987 Version 0.91 10-Sep-97
4988 ----------------------
4989
4990 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4991 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4992
4993 2. PCRE was looping on encountering an indefinitely repeated back reference to
4994 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4995 Perl does - treats the match as successful.
4996
4997 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5